University of Southern California Dissertations and Theses

Forecasting Selected Statewide Recreation Requirements

(USC Thesis Other)

Forecasting Selected Statewide Recreation Requirements

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 72-17,483 McMAHON, Peter Joseph, 1931- FORECASTING SELECTED STATEWIDE RECREATION REQUIREMENTS. University of Southern California, Ph.D., 1972 Economics, general University Microfilms, A X ERO X Com pany, Ann Arbor, Michigan © Copyright by Peter Joseph McMahon 1972 FORECASTING SELECTED STATEWIDE RECREATION REQUIREMENTS by Peter Joseph McMahon A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Economics) January 1972 UNIVERSITY O F SO U T H E R N CALIFO RNIA TH E GRADUATE SC H O O L U N IV ER SITY PARK LOS A N G ELE S, C A L IF O R N IA 9 0 0 0 7 This dissertation, written by PETER JOSEPH McMAHON under the direction of h .l s „ . Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Gradu ate School, in partial fulfillment of require ments of the degree of D O C T O R O F P H I L O S O P H Y Dean D ate F e b r u a r x . . l972. DISSERTATION COMMITTEE PLEASE NOTE: Some pages may have indistinct print. Filmed as received. University Microfilms, A Xerox Education Company TABLE OP CONTENTS LIST OP TABLES..................................... iv LIST OF FIGURES................................... vi Chapter I. INTRODUCTION ................................. 1 The Problem ; Review of the Literature Definitions of Terms Used ! Methods of Research and Analysis j Organization of the Remainder of 1 the Dissertation I j II. PROBLEMS OF MEASURING AND PROJECTING | DEMAND FOR OUTDOOR RECREATION........... 24 i I Basic Concepts of Demand and Supply ■ Nature of the Outdoor Recreation | Market and Problems Affecting Demand Determination ! Use of Visitation Data and the Identification Problem j Conclusions ! III. ACTUAL METHODS USED IN PROJECTING ! OUTDOOR RECREATION DEMAND: THE "MARKET BENEFIT," "INDIRECT" AND "DIRECT" APPROACHES................. 59 Early Studies | Recreation Demand and Benefit | Evaluation Methods j Recreation Demand Studies Using the Indirect Approach Recreation Demand Studies Using the Direct Approach Conclusions ii Chapter IV. THE "SOCIOECONOMIC" AND OTHER METHODS USED IN PROJECTING OUTDOOR RECREATION DEMAND 118 Projecting Demand from Knowledge of Socioeconomic Patterns An Econometric Approach to Rec reation Demand and Supply Analysis User Response Model Other Methods of Demand Proj ection Conclusions V. FACTORS AFFECTING THE DEMAND FOR OUTDOOR RECREATION................. 199 Socioeconomic Factors Time and Distance Conclusions VI. AN APPROACH TO FORECASTING STATEWIDE OUTDOOR RECREA TION DEMAND AND SUPPLY Scope of the Model Underlying Theory of the Model Model Development and Data Sources Conclusions REQUIREMENTS 251 VII. SUMMARY AND CONCLUSIONS 369 Summary of Major Findings Conclusions APPENDIX 385 BIBLIOGRAPHY 421 LIST OF TABLES Table Page IV-1 Participation Days in 16 Major Summertime Outdoor Recrea tion Activities, 1960-2000 ............. 127 V-l Relationship Between Socioeconomic Factors and Participation in Outdoor Recreation ...................... 209 V-2 Relationship Between Socioeconomic Factors and Four Types of Recre ational Activity.......... 215 V-3 First Order Interrelationships — 2 Coefficients of Determination, R-y . . . 219 V-4 Factors Preventing Desired Outdoor Activity........................ 225 V-5 Reasons for Not Participating as Many Days as One Would Like — Summer Season.......................... 226 VI-1 Recreation Activity by Travel- Time Zones............................... 296 VI-2 Total Potential Demand for Eleven Outdoor Recreation Activities in Visitor Days, 1960 ............... . 302 VI-3 Total Potential Demand for Eleven Outdoor Recreation Activities in Visitor Days, 1970 .... o , ... . 303 VI-4 Total Potential Demand for Eleven Outdoor Recreation Activities in Visitor Days, 1980 304 VI-5 Total Potential Demand for Eleven Outdoor Recreation Activities in Visitor Days, 1990 305 iv Table Page VI- 6 Summary of Supply Variables........... 316 VI- 7 Summary of Visitation to Arizona Recreation Areas ............. 318 VI- 8 Correlation Matrix for 17 Supply Variables................. 328 VI- 9 Comparison Between Actual and Estimated Visitation............. 341 VI-10 Facility Combinations of Supply- Demand Equilibrium, 1960 ............. 351 VI-11 Facility Combinations of Supply- Demand Equilibrium, 1970............. 354 VI-12 Facility Combinations of Supply- Demand Equilibrium, 1980 ....... 355 VI-13 Facility Combinations of Supply- Demand Equilibrium, 1990 ............. 356 I t V LIST OF FIGURES Figure Page II-l Equilibrium of demand and supply .... 30 II-2 Effects of shifting demand and supply curves on equilibrium price................................... 34 II-3 Effects of correlated shifts in demand and supply curves ............. 36 II-4 Demand and supply at zero price........ 38 II-5 Effects of shifting demand and supply functions on recreation demand curve estimation ............... 54 III-l Travel cost approach................... 75 III-2 Consumer surplus........................ 77 V-l Von Thunen agricultural location model ........................ 239 V-2 Simple outdoor recreation location model ........................ 240 VI-1 Travel-time zones from the population center of gravity of the Phoenix-Tucson metro politan areas.......................... 294 VI-2 Aggregate potential demand: 1960-1990 .............................. 306 VI-3 Market demand curve for eleven park-oriented recreation activities: 1970 307 VI-4 Facility combinations of supply- demand equilibrium: 1960 359 vi Figure Page VI-5 Facility combinations of supply- demand equilibrium: 1970 360 VI-6 Facility combinations of supply- demand equilibrium: 1980 361 VI-7 Facility combinations of supply- demand equilibrium: 1990 362 vii CHAPTER I j i INTRODUCTION j In recent years, the demand for outdoor recrea- J tion has increased rapidly. This demand, created by an 1 I expanding urban population with greater leisure time, j income and mobility, is expected to grow even more rapidly in the future. Along with this growing demand has been increasing public concern as to the adequacy of recrea tion facilities and natural resources to meet future recreation demands. Efforts by state governments to provide outdoor j recreation facilities and allocate natural and financial j resources to meet these growing demands have intensified in recent years. For example, appropriations for state park systems increased 50 percent between 1950 and 1955, 60 percent between 1955 and 1960, and 75 percent between 1962 and 1967.1 Expenditures by state governments for construction of new park and recreation facilities are now running 40 percent above their 1969 level.2 lu.S. Bureau of the Census, Statistical Abstract of the United States: 1970 (91st ed.; Washington, D.C.: Government Printing Office, 1970), p. 200. ^Dan M. Bechter, "Outdoor Recreation," Monthly Review (Federal Reserve Bank of Kansas City, 1970), p. 16. 1 These efforts, however, have not kept pace with the demand. In the United States, where resources for outdoor recreation had seemed virtually unlimited, it is now a common experience to be turned away from parks or ] find inadequate or undeveloped facilities. Mounting en- j vironmental awareness regarding the effect of increasing j demand on the quality of resources has recently resulted ! in a system of "rationing" to help preserve wilderness I areas; special permits are now being issued to limit i *5 I recreation activities in some of these areas.3 ; i The Problem Historically, state government has had an impor- j tant responsibility in providing outdoor recreation opportunities for its residents. Even greater emphasis on this role was stressed by the Outdoor Recreation Re- j sources Review Commission (ORRRC) as a result of its j j nationwide survey of outdoor recreation. For states to J fulfill their responsibilities, however, the ORRRC also | stressed the need for long-range, comprehensive statewide I outdoor recreation planning for the development of rec reation areas to help meet these growing demands. As ------------------------ ! J"Wilderness-Use Curbs 'Just Around the Corner,'" j Los Angeles Times, September 27, 1971, Part I, pp. 3, 25. noted by the ORRRC: State governments must clearly intensify their current activities if they are to fulfill their responsibilities as major suppliers of outdoor recreation services. A first requirement is a comprehensive long-range plan for the develop ment of outdoor recreation o p p o r t u n i t i e s . 4 With the prospects of even larger increases in demand and | I greater pressures on resources in the future, meaningful I i forecasts of recreation demand and the supply of facil- j ities necessary to meet this demand are urgently needed j to guide these planning e f f o r t s . ^ j In general, however, state outdoor recreation | planning is still in its early stages. The data and j techniques to make the necessary forecasts and enable effective planning are in many instances still unavailable; | or costly to develop. j Moreover, the highly complex nature of the rec reation market makes it difficult to apply the tools of j microeconomic demand-supply analysis to making these j forecasts. While there has been a growing interest on ^Outdoor Recreation Resources Review Commission, | Outdoor Recreation for America (Washington, D.C.: Gov ernment Printing Office, 1962), p. 139. 5 J a c k L. Knetsch, "Assessing the Demand for Out door Recreation," Journal of Leisure Research, Vol. I, No. 1 (Winter 1969), p. 85. 4 the part of economists in developing approaches to enable such projections, there is no generally accepted tech nique. The lack of sufficient data regarding recreation demand, the "zero price" and "identification" problems and the "jointness" of demand and supply are but a few of the complications, as will be discussed later in this study, that make it difficult to apply traditional eco nomic constructs to estimating future demand and supply requirements. Although a variety of approaches have been developed in attempts to overcome these difficulties, their application to statewide recreation planning is extremely limited. Statement of the Problem There are three critical questions that must be resolved by the recreation planner to enable preparation of a comprehensive recreation plan: 1. How can a true measure of future recreation demand be determined, given current data limitations, the complex nature of the recreation market, and the time and financial constraints typically faced by the planner?; 2. What effect does the existing supply of facil ities have on recreation behavior?; and 5 3. What are the amounts, types and optimum com binations of facilities that must be supplied, and where should they be located to best satisfy demand, given the effect of existing supply on recreation behavior? The approaches developed to provide solutions to the first question have been hampered by the lack of data. Unfortunately, there are still insufficient data regarding recreation demand, particularly at the state level, to provide reasonably accurate parameters to de rive a true demand function for outdoor recreation. Moreover, the time and costs to collect reliable data and conduct the analysis are still prohibitive in almost all cases. However, at least some approximation of future demand must be made, given these constraints. On the j other hand, the approximation must be as accurate as pos- | sible to be useful for recreation planning, programming j I and budgeting. [ The demand estimating procedures commonly used have been deficient in a number of areas. For example, most studies usually consider only the demands of the population in a certain age group, rather than the demands of the total population. Also, while consideration is sometimes given to the socioeconomic factors affecting demand, there is little or no proper account taken on the effects of travel time and distance on demand. Al- i though some approaches attempt to incorporate the effects of travel time and distance, they usually exclude the j socioeconomic characteristics of the population and im- j properly use past visitation (or consumption) data to | determine potential demand. On the other hand, when j socioeconomic and distance factors have been included in some econometric models to estimate demand and provide j solutions to the "identification" problem, reduced form i ! market clearing equations, which explicitly include I supply, are used to project demand. Because of the naturei os such models, a true measure of demand, or for that matter, a reasonable approximation of potential demand, j is not obtained and the identification problem is not : solved. Moreover, when estimates of potential demand are j made (albeit inadequate), there is little or no account taken of the proportion of participation that does not ■ occur at public recreation areas, such as parks, although a park and its related facilities are the primary type of supply component planned for and provided by public agencies. These approaches to estimating future demand can either grossly overestimate or underestimate potential 7 demand and thus have limited value for recreation planning. Solutions to the second question regarding the effect of existing supply on recreation behavior have also been ignored in previous studies. Consideration of this effect is crucial to recreation planning as the planner must know the influence or attraction that the amount, type and location of existing supply has on rec reation behavior, as new facilities must be provided within this framework. Otherwise, optimum use of exist ing resources is not possible. Answers to the third question concerning optimum supply requirements are closely related to the second issue noted above, but again, previous studies have been deficient in this aspect. Usually, determination of supply requirements is made by comparing the amount of facilities demanded (typically derived from a recreation standard) with the existing level of supply, with the re sult being a deficiency (undersupply) or a surplus (over supply) . This technique does not specify the optimum combination of the amounts and types of facilities, or their proper locations, to best satisfy demand, given the influence of existing facilities and the behavior of rec 8 reationists. Consequently, this approach is also inade quate for recreation planning. Importance of the Problem Outdoor recreation has become one of the fastest growing activities in the United States. It is a major use of natural and financial resources, involving some 300 million acres of land and water areas,® and direct government expenditures of over $1 billion annually.7 It is also a major leisure time activity in which over 90 percent of the population participates.® It is also a generator of billions of dollars of consumer spending. Consumer expenditures on recreational goods and services totaled $36 billion in 1969.^ In the future, the demand for outdoor recreation is expected to increase dramatically. According to re cent projections, recreation demand by the year 2000 is Outdoor Recreation for America, p.223. 7Statistical Abstract of U.S., pp. 199-201. ®Eva Mueller and Gerald Gurin, Participation in Outdoor Recreation: Factors Affecting Demand Among Amer ican Adults. ORRRC Study Report 19 (Washington, D.C.: Government Printing Office, 1962), p. 21. ^Bechter, op. cit.. p. 15. 9 expected to reach 16.8 billion occasions,1® an increase of 10.4 billion occasions over the level in 1965. This represents a total increase of 160 percent as compared with an expected population increase of 76 percent over the period.11 The rapidly increasing demand for recreation has been such that states are faced with the problem of con tinuing adjustments in resource allocations to meet this demand. As outdoor recreation increases in magnitude and importance, more land, water and financial resources will be needed. However, the same factors which created this demand — increased population, urbanization, leisure time, income and mobility — are contributing to other demands which compete for these same limited public resources. For state government to meet its responsibility in providing outdoor recreation opportunities, meaningful forecasts of future recreation demands and the supply ^Recreation occasions are the number of separate days or portions of a day a person participates in out door recreation activities. 11U.S. Department of the Interior, Bureau of Out door Recreation, Outdoor Recreation Trends (Washington, D.C.: Government Printing Office, April, 1967), pp. 20- 10 requirements to meet these demands are essential. Such projections are necessary for a number of reasons. For example, rational public planning and budgetary alloca tion decisions require knowledge of the future magnitude and spatial distribution of demand for specific outdoor recreation activities and facilities in order to effi ciently allocate land, water and financial resources. These forecasts can serve as guidelines to determine the quantity of resources and facilities that must be sup plied as well as the need for additional recreation areas in the future. Based upon these forecasts, land and water resources could be set aside now while they are still available to prevent their preemption by other uses. Also, these projections can help determine the types of facilities preferred by people, the quantities necessary, and the proper locations so as to best satisfy these demands. Such forecasts are thus essential to en able preparation of long-range comprehensive recreation planning programs. The question then becomes not whether such projections are essential, but what approach should be used to make them. 11 Objectives of the Study The major objective of this study is to develop an approach useful for recreation planners at the state level to project potential outdoor recreation demand and the optimum combinations of facilities that must be sup plied to meet this demand. An empirical model to make these forecasts will be formulated based on theoretical microeconomic constructs as well as other theories re garding recreation that have been developed in recent years. In recognition of the usual data and financial limitations faced by recreation planners, the model uti lizes information that is currently available or that can be obtained easily to minimize the costs of data collec tion. In addition, it is designed to help overcome the deficiencies noted in previous approaches and provide answers to the critical questions currently faced by recreation planners. Other objectives of this study are to: examine the complex nature of the recreation market and problems involved in applying traditional economic constructs to measuring and projecting demand and supply requirements; critically review the actual methods used in projecting 12 demand in terms of their applicability to projections of outdoor recreation demand in general and to statewide forecasts in particular; and, analyze and identify which socioeconomic characteristics of the population, as well as other factors are most relevant in projecting future demand and which may be disregarded. To focus more clearly on these objectives and to provide empirical data for development and application of the model, a case study was made for the state of Arizona. This state contains as diverse a collection of natural recreation resources as can be found anywhere in the na tion. Its sharp contrasts of terrain, climate and varied scenery offer a wide variety of year-around recreation opportunities. As such, it provides the framework for development of a model that should be of assistance to other states in preparing comprehensive plans for outdoor recreation. Limitations and Scope of the Study This study is concerned only with the demand and supply requirements for public outdoor recreation facil ities to be furnished by public agencies. More specif ically, the study is limited to determining the demand 13 for park-oriented activities and the supply of park facil ities that must be provided to meet this demand, since parks are the typical recreation product supplied by public agencies with public funds. It is recognized that the private sector also provides recreation opportunities, but as will be shown later in this study, these facilities are extremely limited. In addition, this study deals only with the rec reation demand and supply requirements of Arizona resi dents. Nonresidents or out-of-state visitors are excluded as the demand for outdoor recreation facilities generated by these visitors was considered negligible. Moreover, comparisons to other states or to the nation as a whole are made only when appropriate to the overall objectives of this study. Further, the study deals with demand forecasts and supply requirements for the state as a whole, or an area that is considered representative of the entire state, and not for individual cities, communities or specific sites. A broad view is necessary and even advantageous to assure consideration of a comprehensive outdoor recrea tion program. Also, the study is concerned with outdoor recreation in non-urban areas, since outdoor recreation 14 is typically undertaken in a non-urban environment and excludes urban-type facilities such as neighborhood parks, playgrounds and the like. Finally, this study specifically excludes consid eration of the economic benefits and costs derivable from outdoor recreation. The approach developed in this study, however, will provide data useful for these analyses. Review of the Literature Although a large body of literature on outdoor recreation has been developed, most of it relates to methods of placing a value on user benefits for use in cost-benefit analyses to justify projects used for recre ation purposes. It has only been within the past few years that economists have become actively engaged in ap plying the theoretical constructs of economic theory to quantitative analyses to estimate and project recreation demand. The recent approaches have been geared toward statistical estimation of recreation demand functions and solutions to the "identification problem." However, be cause of the nature of these models, estimates of poten tial demand in its true economic sense cannot be obtained and the end result is yet another aspect of the identifi- 15 cation problem. A comprehensive review and critique of the actual methods that have been developed to estimate and project recreation demand, and their applicability to statewide forecasts of recreation demand and supply requirements, are contained in Chapter III of this report. Definitions of Terms Used The following definitions are basic to this study. Outdoor Recreation The term, "outdoor recreation," although in com mon use for many years, generally has an ambiguous mean ing due to the lack of a universal, precise definition. The concept of outdoor recreation employed in this study pertains to any leisure time recreational activity vol untarily undertaken in a non-urban environment character ized by a natural, outdoor setting.12 Potential Demand for Outdoor Recreation The term, "potential demand," used in this study is based upon the hypothesis that people have certain desires for outdoor recreation activities regardless of 1 2 Mueller and Gfurin, op. cit. , p. 1. 16 the supply of recreation facilities available to them. Potential demand refers to the desire and ability of people, both physical and financial, to participate in recreation activities. Potential demand differs from past visitation in that the latter refers to the past use of existing facilities or past consumption. The distinction between potential demand and con sumption is crucial to studies in recreation economics, particularly those dealing with demand projections. Rec reation economists have typically relied on consumption or past visitation data to determine recreation demand. The ex post assumption, however, is seldom made explicit with the result that the reader of such studies is fre quently unaware of the historical bias implicit in the study conclusions. Other difficulties such as the "identification problem" have also arisen from use of such visitation data. In this study, potential demand will be used, as noted above, to help overcome these problems. Recreation Activities The types of outdoor activities pursued by people are many and varied. Again, however, the lack of uni versal definitions for the various outdoor recreation 17 activities have been a source of misunderstanding and confusion in recreation economics and in dealing with outdoor recreation data. Because terms such as boating, nature study, etc., hold a different meaning for differ ent people, the following definitions, adapted from the National Recreation Survey,13 outline the dimensions of those park-oriented outdoor recreation activities consid ered in this study. Boating (other than sailing or canoeing) The recreational use of any boat other than sailboats, canoes or houseboats is covered by this definition. It encom passes the use of rowboats, outboard or inboard motor boats, rafts and floats for fishing, water skiing, etc. Camping.— By camping is meant living out-of-doors using a bedroll, sleeping bag, trailer, tent or a hut open on one or more sides for shelter, provided the per son takes his bedding, cooking equipment or food with him. Excluded from this category are formal camps such as Boy Scout camps, as well as cabins, lodges, church youth camps and other group camping activities. 1 Outdoor Recreation Resources Review Commission, National Recreation Survey, ORRRC Study Report 19 (Wash ington, D.C.: Government Printing Office, 1962), pp. 108- 109. 18 Driving for pleasure.— Included in this category are both riding and driving for pleasure. If the driving were mixed, the determining factor is whether it was primarily for pleasure. Activities such as racing are excluded. Hiking (on trails with pack).— The limitation "on trails with pack" excludes casual walking and nature walks. Horseback riding.— This includes only recreation riding. Riding to or from work or school, or riding as part of a job is excluded. Nature walks.— Included here are walks for the purpose of observing plants, birds or animals, as well as the collection of specimens, photographing natural subjects, etc. Picnicking Any activity away from home which involves the preparation or eating of a meal out-of-doors is included in this activity. Sightseeing.— This activity consists of inten tionally looking at something of interest out-of-doors. The "intentional" limitation excludes such things as casually looking from the car window during a trip. If a person took a particular route or went out of his way to see a particular sight, it is called sightseeing. 19 Swimming.— Swimming as well as playing in the surf or water, skin or scuba diving, etc., are included in this category. Walking for pleasure.— Any walking not included under hiking or nature walks, from early morning "consti tutionals" to all day walks without a pack, falls within this category. Water skiing.— This includes any of the various sports whereby a person is towed behind a boat while on an aquaplane, water skis, or other apparatus of this type. Recreation Facilities Recreation facilities is a general term used in this study that refers to land and water areas specif ically set aside for particular outdoor recreation activities (e.g., picnic areas, tent camps, etc.) as well as the recreation or man-made units for these activities (e.g., picnic tables, tent spaces, etc.). Methods of Research and Analysis The base data used in this study were drawn almost exclusively from primary data sources obtained through personal correspondence. Requests for information con 20 cerning outdoor recreation in general and the subject of this study in particular were sent to the Western Region and National headquarters offices of the Bureau of Outdoor Recreation in California and Washington, D.C., respec tively. Bibliographical materials as well as referrals to various state agencies that had undertaken or were in the process of undertaking outdoor recreation plans were received. Requests for available information concerning these plans were then sent directly to these state agencies and copies of various state outdoor recreation plans were received. Among the state agencies contacted was the Arizona Outdoor Recreation Coordinating Commis sion (AORCC) , located in Phoenix, Arizona. The AORCC is the legally designated representative for preparing the state's outdoor recreation plans and development of fed eral, state and local agencies within the state. In ad dition to the materials received from the AORCC, referrals to other sources were furnished. Information was also obtained through personal and telephone interviews but use of this technique was limited. Interviews were held with BOR Western Region representatives as well as AORCC officials to obtain clarification of various data as well as further insights 21 into the subject matter of the study. Library investigations were made to determine other primary as well as secondary sources of data avail able for this study. The chief primary sources included the pertinent publications of the Outdoor Recreation Re sources Review Commission, the Bureau of Outdoor Recrea tion and the U.S. Bureau of the Census. Library investi gations were also made to obtain published literature (books, articles, periodicals and reports) concerning the various economic theories and methods relating to the determination of recreation demand. A variety of statistical and mathematical analyses were made of the pertinent empirical base data. Trend analyses were necessary to determine and proj ect the var ious factors affecting recreation demand and make the demand projections. Correlation studies were undertaken to determine the degree of relationship between certain supply variables, and least squares regression as well as algebraic techniques were used to formulate a predicting equation to forecast future facility requirements to meet potential demand. Analytical results were derived in large part with the aid of a computer. 22 Organization of the Remainder of the Dissertation The remainder of this study is organized in six additional chapters. Chapters II, III, IV and V form the background for developing an approach for forecasting statewide outdoor recreation demand and supply require ments. In Chapter II, the various problems of measuring and projecting demand for outdoor recreation are examined. First, the basic concepts of price theory related to demand-supply analysis and market price determination are reviewed. Then, the complex nature of the recreation market is described and the difficulties in applying tra ditional economic constructs to measuring and projecting recreation demand are analyzed. Chapters III and IV critically reviews the actual methods used in projecting recreation demand and evaluates these techniques in terms of their applicability to pro jections of outdoor recreation demand in general and to statewide forecasts of demand and supply requirements in particular. Chapter V analyzes the patterns of outdoor rec reation demand and identifies which socioeconomic char acteristics of the population and other factors are most 23 highly correlated with and best explain recreation be havior. The most relevant factors affecting outdoor recreation demand are then determined. Based upon the findings of the previous chapters, an approach to forecasting statewide outdoor recreation demand and supply requirements is developed in Chapter VI. In developing this approach, an empirical model is formu lated to project outdoor recreation demand and the optimum combinations of facilities that must be supplied to meet this demand. The outputs of this model provide answers to the critical questions noted earlier that must be re solved by the recreation planner. As such, the approach developed can be a valuable tool for recreation planning and useful in preparing comprehensive outdoor recreation plans in other states. In Chapter VII, the major findings of the study are presented along with recommendations for further study. Various statistical tables referred to but not included in the body of the study are contained in the Appendix. CHAPTER II PROBLEMS OF MEASURING AND PROJECTING DEMAND FOR OUTDOOR RECREATION The usual analytical approach in measuring the demand for a commodity is to construct a statistical de mand function or curve based upon a series of observed corresponding relations between quantities which were bought or sold at various prices in a market over a cer tain period of time. In estimating or projecting future demand, a theoretical demand curve is constructed based upon expected quantities demanded at alternative prices. To estimate or project the quantity that will be bought or sold, however, a theoretical supply curve must be con structed based upon anticipated quantities supplied at alternative prices; where the demand and supply curves intersect, the market price and quantity that will be bought or sold are established. Since demand relations can be highly subjective and complex, it is often convenient to approximate them by using statistical methods. A demand function whose coefficients have been estimated empirically using econo metric methods can show the structural relationship that 24 25 exists between the amount of a good or service that would be purchased during a given time period and a number of independent variables. Econometrics introduces the prob abilistic elements of the real world, rather than spec ifying the exact functional relationships among variables as is done in economic theory. As Goldberger noted: A main task of econometric theory, indeed, is to provide a bridge between the exact rela tionship of economic theory and the disturbed relationships of economic reality.1 Quantification of specific demand function, how ever, requires a sound foundation in economic theory so that a functional relationship between an observed de pendent variable, observed independent variables and un observed disturbances can be specified for statistical testing. Due to the nature of the outdoor recreation market, however, it is difficult to apply the tools of traditional demand-supply analysis to measure and project the demand for outdoor recreation and determine the supply of resources and facilities needed to meet this demand. To gain a better understanding of this unique situation, this chapter will first review some of the ^A. S. Goldberger, Econometric Theory (New York: John Wiley & Sons, Inc., 1964), p. 2. 26 basic concepts of price theory related to demand-supply analysis and market price determination. Given these basic concepts, it will then review the nature of the rec reation market and the problems involved in applying tra ditional economic constructs to projecting recreation demand. Basic Concepts of Demand and Supply In economics, demand for consumer goods and serv ices is regarded primarily as a function of price. That is, demand for a particular commodity is defined as the various amounts consumers desire or are willing to pur chase in a given market in a given period of time at all possible alternative prices per unit, other things re maining the same (price of the good, consumer income and tastes, prices of substitute and complementary goods, range of products available and price level).^ Other things being equal, the desired rate of purchase per unit of time varies inversely with price, i.e., the higher the price, the less consumers will take. ^R. H. Leftwich, The Price System and Resource Allocation (rev. ed.; New York: Holt, Rinehart & Winston, 1961), p. 27. 27 Demand refers to an entire demand schedule or demand curve showing possible price-quantity combinations per unit of time, ceteris paribus, which typically has a negative slope because of the decreasing marginal utility to the consumer of an additional unit of the commodity.3 The demand curve is a maximum concept indicating the max imum quantities that will be taken and the maximum prices that will be paid per unit of time. At given prices, a consumer will be willing to take smaller amounts, if smaller amounts were all he could obtain, but would not take larger amounts; he will pay no more for a particular quantity but would pay less. When changes occur in the factors held constant (or ceteris paribus conditions), the result is a change in demand or a shifting of the demand curve.4 The quantity of a particular commodity that will be supplied is also a function of price. Supply of a good is defined as the various quantities sellers are willing to place on a given market in a given period of time at all possible alternative prices, other things 3Donald S. Watson, Price Theory and Its Uses (Boston: Houghton Mifflin Company, 1963), pp. 50-55. ^Leftwich, op. cit., pp. 28-30. being held constant (technology, supplies or price of inputs, taxes and subsidies, features of nature and C period of adjustment). It is the relationship between prices and quantities per unit of time which sellers are willing to sell. Other things being equal, the amount supplied will be greater the higher the price. Similar to a demand curve, the supply curve shows the possible price-quantity combinations per unit of time representing the attitudes of sellers, ceteris paribus. Usually, it is upward sloping because of the increasing marginal cost to the supplier of producing an additional unit of the commodity.^ The supply curve also shows the maximum quantities supplied on the market. At a given price, suppliers will be willing to supply less, but will not supply more. A supply curve will also shift if there is a change in the factors held constant.7 Price is therefore common to both the demand and supply sides of the market and the effect of a price ^Richard A. Bilas, Intermediate Microeconomic Theory (preliminary ed.; New York: McGraw-Hill Book Company, 1965), p. II-4. ^Watson, op. cit., p. 229. 7Leftwich, op. cit., pp. 30-31. 29 change on demand is the opposite of its effect on supply. The usual market transaction involves the meeting of de mand and supply at a price where buyers are willing to purchase a certain amount and sellers are willing to offer or sell the same amount. At this market price, or equilibrium point where the demand curve and supply curve intersect, the market is cleared. A lowering of price may bring forth demand for a larger quantity but this de mand cannot be translated into purchases unless supplies increase. A graphic illustration of the equilibrium of de mand and supply in a simple price-quantity analysis is shown in Figure II-l. At point A, where the demand curve and supply curve intersect, the market is in equilibrium and the price OP-^ is the equilibrium price. That is, at price OPi, consumers are willing to take quantity OQ^ per unit of time and suppliers are willing to supply this same amount. At this market price, demands are equal to supplies. At a higher price, such as OP2, demand is OQ2, whereas supply is OQ3. The excess of supply between BC (or the quantity between OQ2 and OQ3) forces the price down. This higher price can prevail only briefly because sellers attempt to sell more than buyers are willing to 30 P R I C E 2 3 0 Q Q Q Q U A N T I T Y Figure II-l — Equilibrium of demand and supply 31 buy at the same price. Likewise, a price lower than the equilibrium price, such as P3, can exist only briefly because the excess demand EF (or the quantity between OQ^ and OQ5) pushes the price up; buyers try to purchase more than sellers want to sell. When the demand and sup ply conditions are such that any disturbances of the equilibrium price OP^ will set in motion forces to bring it back to that level, the equilibrium is said to be stable.® This important concept, that all responses or transactions in a market are assumed to be at points of equilibrium (or points of intersection between the demand curve and supply curve) and that each observed point is one point on both the demand curve and supply curve, is one of the basic theoretical building blocks of price g theory. in empirical analysis, however, this model has severe limitations. It is static in that it assumes the tastes of consumers and technologies available to sup pliers, which mutually determine the quantities produced, Q A. W. Stonier and D. C. Hague, A Textbook of Economic Theory (2d ed.; New York: John Wiley & Sons, Inc., 1961), pp. 31-32. 9Watson, op. cit., pp. 220-222. 32 are held constant? it does not allow for variations in demand, supply and price over time and as such is un realistic. In a real-world market situation, the market is dynamic. Demand curves and supply curves are con stantly in motion and new equilibrium prices are con stantly being reached. The equilibrium price is actually one point on both the demand curve and supply curve at an instant in time. Given the vagaries of the market and the forces operating to bring about changes in prices, demand and supply, it is difficult, if not impossible, to establish an equilibrium price or show that the position would be stable.Consequently, it is difficult at best to determine whether the observed fluctuations in equi librium prices were due to demand responses or supply responses, and therefore difficult to construct a demand curve. In fact, whether it is at all possible to con struct a demand curve from data which pertain only to the points of intersection of a theoretical unknown demand curve with a theoretical unknown supply curve at different I points in time, is highly questionable.^ This difficulty ^°Kenneth E. Boulding, Economic Analysis, Volume I; Microeconomics (4th ed.; New York: Harper & Row, Pub lishers, 1966), pp. 50-51. •^H. H. Liebhafsky, The Nature of Price Theory (Homewood: The Dorsey Press, Inc., 1963), pp. 107-108. 33 is the "identification problem" long familiar to econo mists in attempting to construct demand curves and deter mining whether the observed responses or fluctuations in equilibrium prices were caused by changes in demand or supply. This problem may be illustrated by the diagrams in Figures II-2 and II-3 which were derived from the analyses made in 1927 by Working.^ Figure II-2 (A) , for example, shows the demand curve shifting from to D2 and a supply curve shifting from Si to S2. In this situation, the shifting of both curves is shown as being approximately equal. Under these conditions, a series of prices will result which are graphically represented by the dots in Figure II-2 (B). It is from such data that a demand curve would have to be constructed, but it is evident that no satisfactory curve could be fitted; a line of one slope would give 1 substantially as good a fit as any other. J If, however, demand is relatively constant and the supply curve shifts more than the demand curve, as 1 7 E. J. Working, "What Do Statistical 'Demand Curves' Show?" in Quarterly Journal of Economics, Vol. XLI (1927), pp. 212-235, as reprinted in K. Boulding and S. Stigler, eds., Readings in Price Theory (Chicago: Richard D. Irwin, Inc., 1952), pp. 97-115. ^Ibid. , pp. 102-103. P R I C E ( A ) Q U A N T I T Y P R I C E ( ° ) Q U A N T I T Y PR IC E Q U A N T I T Y P R I C E ( B ) 34 • • • Q U A N T I T Y ( D ) P R I C E Q U A N T I T Y P R I C E w* Q U A N T I T Y Figure II-2 — Effects of shifting demand and supply curves on equilibrium price 35 shown in Figure II-2 (C), a different series of prices would result; this is depicted in Figure II-2 (D). Here, a true demand curve could be fitted to these prices. However, a supply curve of any slope could be fitted to the data and intersect the demand curve at a number of points. Similarly, if supply is relatively constant and the demand curve shifts more than the supply curve, as depicted in Figure II-2 (E), the situation shown in Fig ure II-2 (F) could result. Here, it may be noted that only a supply curve could be fitted to these prices.1^ In the above noted illustrations, it was assumed that demand and supply curves shift independently to one another and at random. This may not, however, be the situation; it is possible that the shifts may be corre lated, i.e., that a shift of the demand curve to the right may be accompanied by a shift of the supply curve to the left and vice versa. This situation is shown in Figure II-3 (A) . Here it may be noted that a series of equilibrium prices would result from the intersection of Di with Si, and so forth. If a statistical demand curve (D) was fitted to these points, it would not conform to l^ibid., pp. 103-104. ^5Ibid., pp. 104-105. P R C E P R I C E Q U A N T I T Y F i g u r e I I - 3 — E f f e c t s o f c o r r e l a t e d s h i f t s i n d e m a n d a n d s u p p l y c u r v e s the theoretical demand curve (D^, D2 or D3) ; it would be steeper (or more inelastic). If, on the other hand, a shift of the demand curve to the right is accompanied by a shift in the supply curve to the right, the resulting statistical demand curve would be as shown in Figure II-3 (B). This demand curve (D) again fails to conform to the theoretical demand curve (D]_, D2 or D3) but in 16 this instance is more elastic. Without carrying the illustrations any further, it is evident that determination of equilibrium points and construction of a demand curve, given these fluctua tions, is difficult. It also points out, however, that supply as well as demand factors must be used in attempt ing to analyze conditions of equilibrium. At this point, it is also appropriate to note that demand and supply do not necessarily result in ex change at a price. As pictured in Figure II-4 (A), there is a demand curve and supply curve but no price. The de mand is such that buyers are not willing to pay prices as high as the sellers want to receive for the smallest quantities they are willing to sell. As shown in Figure II-4 (B), the demand is such that buyers will take •^Ibid. , pp. 107-109. 38 P R I C E ( A ) ( B ) Q U A N T I T Y P R I C E 0 Q Q Q U A N T I T Y Figure II-4 — Demand and supply at zero price I 39 quantity OQi, but no more, at a zero price. Sellers, however, are willing to offer quantity OQ2 at no cost to the buyer. Thus, the commodity can command no price in this supply-demand situation and is a "free good.1,17 Given the basic demand and supply concepts dis cussed above, the next section of this chapter deals with the nature of the outdoor recreation market and the dif ficulties in applying theoretical constructs to measuring and projecting outdoor recreation demand. Nature of the Outdoor Recreation Market and Problems Affecting Demand Determination As noted in the previous section, price is common to both the demand and supply sides of the market and the equilibrium price determines the quantities bought or sold in the market. In outdoor recreation, however, there is no conventional market pricing. As Wennergren noted: . . . Many recreational uses of our natural resources are not regulated by conventional market pricing and therefore are not subject to formal prices and fees. Consequently, usage appears to be free, at least in the sense normally ascribed to other marketed commodities. ■^Donald S. Watson, Price Theory and Its Uses (Boston: Houghton Mifflin Company, 1963), p. 222. 1®E. Boyd Wenngergren, "Valuing Non-Market Priced Recreational Resources," Land Economics, Vol. XL, No. 3 (August 1964), p. 303. 40 For example, public outdoor recreation is not marketed per se. It is supplied at zero or close to zero prices as a matter of public policy and as such, price is not determined by the market mechanism. That is, the costs of supplying and maintaining recreation sites and facilities are borne for the most part by governmental units. Further, there are no charges or only nominal fees for use of the sites or facilities. Recreation, however, is not a "free good" as described previously. It does entail some costs to the consumer (e.g., costs of equipment and transportation), but consideration of these costs is especially difficult due to the problems in col lecting accurate information on the variety of small ex penses incurred for different items on many different recreation occasions.^ Moreover, outdoor recreation does not fall neatly into the classification of a public or "collective good."^O Collective goods refer to those goods which ■^Eva Mueller and Gerald Gurin, The Demand for Outdoor Recreation (Ann Arbor: University of Michigan, 1961) , p. 3. on For elaboration of this concept, see two papers by Paul A. Samuelson, "The Pure Theory of Public Expendi tures," Review of Economics and Statistics, Vol. XXXVI, No. 4 (November 1954), pp. 387-389? and "Diagrammatic Ex position of a Pure Theory of Public Expenditures," Review of Economics and Statistics, Vol. XXXVII, No. 4 (November 41 society or segments of society consumes as a group. Sim ply stated, consumption by one individual does not "use up" the good or affect the satisfaction associated with consumption by other members of society. Classic examples of public goods are national defense, police and fire protection, education, etc. However, consumption by a recreationist can "use up" the good and affect the con sumption of other recreationists. Examples here are the crowding of parks and situations where recreationists are turned away due to full park utilization. Also, while outdoor recreation appears to possess many of the char acteristics of a public good, the behavior of the consumer (or recreationist) and factors affecting his consumption (or amount of participation) are much like those for any ordinary economic good. Stated differently, it is now widely held by economists that the consumer does not view outdoor recreation opportunities differently in terms of their contribution to his satisfactions any differently than he views the large number of other goods among which he must choose, given the constraint of a limited budget. 1955), pp. 350-356. See also, William J. Baumol, Welfare Economics and the Theory of the State (2d ed.; Cambridge, Mass.: Harvard University Press, 1967), 171 pp. 42 Thus, there exists for outdoor recreation, just as for other goods, a demand function reflecting the quantities the consumer is willing to purchase at alternative prices. In addition, under free market conditions, the consumer will include recreation within the collection of goods among which he allocates expenditures in maximizing sat isfactions. 2^ While not a public good, the absence of market pricing for outdoor recreation has been attributed to its "public goods" aspects. For example, Davidson, Adams and Seneca^ attribute it to public goods externalities and state "... there is no need to ration public goods be tween individuals and no set of market prices for public goods is useful for individual production and/or consump tion decisions. The authors cite three interrelated factors re sulting from public externalities which lead to "failure" O '! xIvan M. Lee, "Economic Analysis Bearing Upon Outdoor Recreation" in ORRRC Study Report 24, Economic Studies of Outdoor Recreation (Washington, D.C.: Govern ment Printing Office, 1962), pp. 5-6. 22Paul Davidson, Gerald F. Adams, and Joseph Seneca, "The Social Value of Water Recreational Facilities Resulting from an Improvement in Water Quality: The Dela ware Estuary," in Allen V. Kneese and Stephen C. Smith, eds., Water Research (Baltimore: The Johns Hopkins Press, 1966). 23Ibid., p. 177. 43 in the market for outdoor recreation facilities. First, a public good externality exists which is due to "off- peak" demand resulting from the time variability of demand for the good, and the non-storability of the good over time so that excess supply quantities cannot be car ried over to periods of peak demand. Second, there exists an "option demand" which reflects a value of the resource to those who wish to maintain the option to con sume in the future (even though they are presently not participating in the enjoyment of the resource) which leads to a public good externality. Third, an "oppor tunity demand" exists which reflects a future value that possibly arises from the use by those who may learn to enjoy a facility or service which they are presently not using; this "opportunity demand" of "learning-by-doing" leads to further public externalities.24 The absence of market pricing for outdoor recrea tion may also be traced to institutional and technical conditions. As Wennergren stated: First, there is a notion in the United States that certain services should be provided with out charge through public ownership of the re sources that produce the service. Outdoor recreation is one activity that has been sub- 24Ibid., pp. 182-187. stantially relegated to this position. . . . Second, even when private ownership is in volved, the nature of the product [^crea tion] is often subject to technical condi tions which make collection of charges from beneficiaries impractical, if not impossible. . . . As a result, non-fee or non-market priced usage is commonly associated with most forms of outdoor recreation.25 It must be noted that it is not the purpose of this study to argue whether zero or only nominal charges for use of public outdoor recreation facilities are de sirable or undesirable. Nor is it the purpose to discuss the economic theories of externalities, public goods and collective action as they relate to the economic rationale and social policy for the public provision of outdoor rec reation. These topics have been well covered in the economic literature.^6 The major point here is that the absence of conventional market pricing makes it difficult ^^Wennergren, op. cit., p. 304. 26 For example, see Robert L. Bish, The Public Economy of Metropolitan Areas, Markham Series in Public Policy Analysis (Chicago: Markham Publishing Company, 1971); Warren C. Robinson, "The Simple Economics of Pub lic Outdoor Recreation," Land Economics, Vol. XLIII, No. 1 (February 1967), pp. 71-83? Richard A. Musgrave, The Theory of Public Finance (New York: McGraw-Hill Book Company, 1959); Paul A. Samuelson, "Aspects of Public Expenditure Theory," Review of Economics and Statistics, Vol. XL, No. 4 (November 1958), pp. 329-338? and F. M. Bator, "The Anatomy of Market Failure," Quarterly Journal of Economics, Vol. LXXII, No. 3 (August 1958), pp. 351- 379. 45 to price recreation and analyze the demand for it in the usual economic sense. Another major complication in attempting to meas ure and project recreation demand is the absence of well- defined units of measurement. The economic concept of demand implies a relationship between the number of units of a commodity purchased and price per unit. For most ordinary commodities, demand is usually measured in units of dimension, weight or time with a dollar value per unit of measurement. However, no such convenient units are available or easily applicable to measuring outdoor rec reation demand.^7 As a result, the number of visits to a recreation site is commonly used as a measure of OQ demand. ° The use of visitation data in most recreation demand analyses, however, has created other problems which will be discussed later in this chapter. In addition, the nature of the recreation supply itself is such that it presents problems to traditional economic analysis. For example, a park typically con- ^Lee, op. cit. , p. 6. ^®Outdoor Recreation Resources Review Commission, Prospective Demand for Outdoor Recreation, ORRRC Study Report 26 (Washington, D.C.: Government Printing Office, 1962), p. 1. 46 tains facilities for accommodating several activities (picnic tables, campsites, etc.). Recreation sites are thus "jointly produced." In addition, a recreation outing to a site usually consists of participation in a collec tion of complementary activities (hiking, sightseeing, etc.) and in effect represents a "joint demand." These supply-demand conditions make it difficult to separate and analyze the effects of each component of supply and demand.^ Another problem relating to the character of the recreation product is that it must be consumed at the site where it is produced. Moreover, supply is not evenly distributed in relation to consumers (as in the case of retail outlets, for example) and its location in many instances depends on the availability of natural re sources.^ Ordinary goods are transported to the con sumer and the price of transportation included in the price of the good (or in the marginal cost to the sup plier) . However, the costs of getting to a recreation site are borne by the recreationists. The problems of 29Lee, op. cit., pp. 36-37. 30Marion Clawson, "The Crisis in Outdoor Recrea tion, " American Forests (March-April, 1959). 47 obtaining such costs were noted previously. The critical importance of location as a factor affecting outdoor rec reation demand will be discussed in detail in Chapter V. Another characteristic of recreation supply is that it is reusable.31 It can be used over and over again unlike most consumer goods. Consequently, it is difficult to determine the amount consumed. In addition, the qual ity of the supply effects its use, although determining these effects is difficult.32 Another, and perhaps the major obstacle in meas uring and projecting recreation demand, is the lack of quantitative data regarding demand. As noted by Knetch, "the data on outdoor recreation are either nonexistent or terrible."33 For instance, most of the data that are ^Recreation supply is reusable only to the ex tent that the resources are properly preserved and main tained. Natural resources can deteriorate to the extent that they are unusable for recreation or other purposes. An example is the growing number of polluted lakes and streams where fish and other marine life have been de stroyed and the waters no longer safe for swimming (e.g., Lake Erie) . 09 Marion Clawson and Jack L. Knetsch, Economics of Outdoor Recreation. Resources for the Future (Balti more: The Johns Hopkins Press, 1966), p. 145. •^Jack L. Knetsch, "Some Topics of Interest in Outdoor Recreation Economics from the Proceedings of a Seminar, April 5-6, 1965, Outdoor Recreation Research (Texas: Texas A & M University), p. 9. available relate to visits to specific areas rather than to recreation demand for new areas. Even records of j these data, however, are fragmentary or piecemeal.34 jn addition, the data are for visitation to sites which typically offer a collection of facilities for a variety | of recreation activities; visitation information on each type of activity is typically not gathered and it is dif- j ficult to relate these data to demand for new facilities j or to participation in particular types of recreation activities.35 Thus, the lack of adequate data presents a i I major problem in recreation demand analysis. As Clawson and Knetsch observed: In some ways, the study of demand for outdoor i recreation today is about where the study of j demand for agricultural commodities was 40 years or more ago. . . . The data problem for | outdoor recreation is now especially serious, but the use of the best presently available J data may aid in pointing the way to collec- j tion of much more suitable data in the f u t u r e .36 3^0RRRC Study Report 26, p. 1. 35warren C. Robinson, "Economic Evaluation of Outdoor Recreation Benefits," in Outdoor Recreation Re sources Review Commission, Economic Studies of Outdoor Recreation. ORRRC Study Report 24 (Washington, D.C.: Government Printing Office, 1962), pp. 7-8. 3®Clawson and Knetsch, op. cit., p. 63. 49 Use of Visitation Data and the Identification Problem As seen from the above, the nature of the recrea tion market is such that it poses a number of problems that make it difficult to analyze the demand for recrea tion. As a result, most recreation demand studies regard "demand" and visitation to an existing recreation site as being identical.3^ Projections of past trends in visita tion or use of a site observed over a period of time are •30 typically considered measures of future demand. In the usual economic sense, however, demand and supply functions each imply a schedule of alternatives; visitation, on the other hand, is actually past use or consumption, and refers to a single point, i.e., the in tersection of both the demand curve and supply curve at a given point in time. Visitation data, therefore, do not separate the effects of demand and supply conceptually or statistically and as such, do not indicate whether 37 For an excellent discussion of the distinction between demand in the technical sense and demand in the popular sense, see Raleigh Barlowe, Land Resource Eco nomics (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1958), pp. 18 ff. 38 Wennergren, op. cit., p. 304. 50 observed differences in visitation (or quantities con sumed) reflect demand responses or supply r e s p o n s e s .39 Use of visitation data to estimate recreation demand thus gives rise to the misidentification of the function being investigated or the "identification problem" discussed in the preceding section of this chapter. The identification problem in the study of rec reation demand results when visitation data are used for empirical estimation. Traditionally, the identification problem has been linked with the use of time series data, since both demand and supply functions can shift over t i m e .40 What is observed in a particular market at a given time is, therefore, the simultaneous solution of two multivariate relations — the supply and the demand functions. If each of these functions was dependent upon different explanatory variables, then regressing the quantity actually exchanged upon one of the sets of ex planatory variables would give an estimate of the asso- on S. V. Ciricacy-Wantrup, "Conceptual Problems in Projecting the Demand for Land and Water," in Land Economics Institute, Modern Land Policy (Urbana: Uni versity of Illinois Press, 1960), pp. 41-68. ^Working, op. cit. , pp. 212-235. 51 ciated relationship. If the two sets of independent var iables contain the same or closely related variables, however, conditions of collinearity or near collinearity may be encountered. Collinearity occurs when two or more independent variables are highly interdependent (which can easily be the case, given the peculiar nature of the outdoor recreation market) so that the net effect of each on the dependent variable is difficult to measure and it may be impossible to tell which, if either, of the rela tionships are being investigated. This problem is further complicated by the absence of a price for recreation, which in most studies of consumer demand is taken as an explanatory variable in both the demand and supply rela tionships. 41 Even if this is not the case, fluctuations in both the demand and supply functions over time will make estimation of the demand function difficult when using visitation data. Although it is widely assumed that price is not an explanatory variable in the supply func tion for most recreation services, and that such a func tion is vertical in nature for any given time period ^Robert J. Kalter and Lois E. Gossee, "Recreation Demand Functions and the Identification Problem," Journal of Leisure Research. Vol. II, No. 1 (Winter 1970) , p. 47. 52 because of its public provision, the relative shifts in the supply and demand curves over time are unknown. Thus, the essential elements required to trace the demand function are missing. In the long run, both the demand and supply func tions can be expected to vary and supply and demand shifts cannot be assumed to be independent. This latter point is of special importance for outdoor recreation since changes in supply depend on the actions of public decisionmakers. However, fitting a curve to the series of points exhibited when either or both of these factors are present cannot be expected to produce the true demand function. As Working pointed out long ago: The methods used . . . may, under some condi tions, yield a demand curve, under others, a supply curve, and under still different con ditions, no satisfactory result may be ob tained. 42 The result depends on both the relative shifts in the de mand and supply functions and the degree of correlation between these functions. This can be true whether or not price is a variable contained in both functions.43 42Working, op. cit., pp. 212-235 43Kalter and Gosse, op. cit., p. 48. 53 Figure II-5 (A) illustrates the problem. As shown here, shifting of the supply curve for recreation is represented by the vertical lines to S4, while the demand curves for the corresponding time periods are de noted by to D4. It is clear that fitting a statistical demand curve through the intersections of the correspond ing curves (or through the points of equilibrium), does not represent the theoretical demand curve. In fact, a fitted demand curve through these points would approxi mate a supply curve. What is needed is a situation where the demand curve exhibits little or no variation while the supply curve shifts over a wide range so that the points of in tersection more closely approximate the true demand curve. The desired situation is shown in Figure II-5 (B) . Thus, the use of visitation data gives rise to the identification problem and further complicates demand estimation. The implications for recreation planners with respect to supplying facilities, however, have not been fully realized. As Morrison noted: If demand statements are in any way to serve as a guide for investments, that is, specif ically to help plan the supply side of the ledger, the problem of having to separate demand from supply becomes a major concern. Improper accounting for demand and supply ( A ) P R I C E Q U A N T I T Y 54 ( B ) P R I C E Q U A N T I T Y Figure I1-5 — Effects of shifting demand and supply functions on recreation demand curve estimation 55 — the "identification problem" as economists call it — could lead to projections of pres ent participation which would perpetuate present imbalances in supply.44 For example, visitation data do not consider the con straint of supply on demand. As indicated previously, one of the characteristics of the outdoor recreation market is the "learning-by-doing" phenomenon. This means that a person who participates in water skiing and knows how to water ski has a demand for this activity; the learning process and thus demand for the activity therefore re quires a supply to exist and be a c c e s s i b l e .45 Inasmuch as changes in the supply of recreation facilities depends on the decisions of governmental bodies and not on the usual price-quantity supplied relationship, it is most likely that visitation figures understate demand. As Daane points out: . . . It is impossible for the public to register its demand for recreation facilities not currently available. . . . It is conceivable that demand would be greater if the supply were increased.46 C. C. Morrison, "A National Survey of Outdoor Recreation Participation and Preferences," speech de livered at the Annual Meeting of the Association of Amer ican Geologists (Toronto, Canada, August 30, 1966). 45uavidson, et al., op. cit., pp. 182-187. ^^Kenneth E. Daane, The Economic Implications of the Regional Park System in Maricopa County (Tempe, Arizona: Bureau of Business Services, College of Business Administration, Arizona State University, 1964), p. 12. 56 By way of example, the observed differences in recreation participation exhibited by populations in dif ferent regions of the country which have similar climatic conditions were found due largely to availability of fa cilities. As Clawson and Knetsch noted: . . . It should not surprise us . . . that people in Knoxville, Tennessee, located in the midst of a half dozen large reservoirs with adequate access and public facilities, water ski in greater numbers than people in Washington, D.C., which is deficient in suitable water. These differences do not by themselves indicate differences in demand for water skiing any more than gross visita tion figures to parks represent statements of demand alone. The figures in both cases are the result of the interaction between demand and supply factors and are the measurement of consequent consumption, or quantities taken by recreationists, given these supplies and demands.47 There is no doubt that visitation figures do, in an ex post sense, reflect demand. However, they also reflect the availability or lack of supply. Visitation is actu ally a measure of consumption at the existing supply con ditions, and not a measure of potential d e m a n d . 48 Thus, the use of past visitation data to project demand can create increasingly more difficult problems in ^ C l a w s o n and Knetsch, op. cit. , p. 115. 48Knetsch, op. cit., p. 11. 57 recreation planning. For example, if demand forecasts for statewide outdoor recreation plans are based on par ticipation rates without taking proper account of supply considerations, it could lead to the assumption that peo ple will demand increasing quantities of what they now have and can perpetuate present imbalances in supply. That is, if some areas show far greater participation rates on the part of the population for swimming, this could, if taken as a statement of demand without consid eration of available supply opportunities, lead to plan ning decisions to build even more facilities in areas already adequately served and where the rates are highest, rather than providing them in deficient areas.^9 Conclusions As seen from the above, the nature of the outdoor recreation is highly complex. The lack of sufficient data regarding potential demand, the "zero" price and "identification" problems, and the "jointness" of demand and supply are but a few of the complications that make it difficult to analyze the demand for outdoor recreation ^Jack L. Knetsch, "Assessing the Demand for Out door Recreation," Journal of Leisure Research, Vol. I, No. 1 (Winter 1969), p. 86. 58 using traditional economic constructs. In fact, many economists have argued that it cannot be done.50 As a result, there is no generally accepted method for such measurements and projections.51 A review of both the early studies that have been undertaken and the more significant of the approaches that have been developed in attempts to overcome these problems is made in the following chapters. 5®Marion Clawson, Methods for Measuring the Demand for and Value of Outdoor Recreation, Reprint No. 10, Resources fc— the Future, Inc. (Washington, D.C., February 1959), p. 9. 51oRRRC Study Report 26, p. 1. CHAPTER III ACTUAL METHODS USED IN PROJECTING OUTDOOR RECREATION DEMAND: THE "MARKET BENEFIT," "INDIRECT" AND "DIRECT" APPROACHES Due to the accelerating participation in outdoor recreation in recent years and the tremendous increase expected in the future, there has been a growing interest among economists to develop approaches to project recre ation demand. The purposes of this chapter and Chapter IV are to critically review and evaluate these techniques in terms of their applicability to projections of outdoor recreation demand in general and to statewide forecasts in particular. Because of the relatively new interest on the part of economists in recreation research, the rapidity with which a large number have entered the field, and the variety of demand estimation procedures that have been developed, there is some duplication and overlapping of approaches. Consequently, it is nearly impossible to neatly classify the methods by precise type or in a chronological order. To provide a reasonable degree of organization, this chapter first reviews the early rec- 60 reation studies made prior to 1947. The approaches de veloped since that time, organized by generalized classi fications of the types of formulations that evolved, i.e., the "market benefit," "indirect" and "direct" ap proaches, are then analyzed. Other methods of demand projection are discussed in Chapter IV. Early Studies Prior to 1927, very little published material dealing with the economics of outdoor recreation appeared in the various economic journals. One of the main reasons for this lack of academic interest by economists was the general lack of participation in outdoor recreation ac tivities. Before that time, the long working hours and heavy work loads left little time for outdoor recreation and participation was limited to opportunities in or near the home. However, the advent of automobile, improve ments in highways, the shortened work week, and increased affluence which came with industrialization in the early part of the twentieth century all contributed to an in terest in and need for recreation of a type different from that traditionally available. With it developed an interest on the part of economists in the economics of outdoor recreation. 61 Interest in the economics of recreation developed first at the University of Wisconsin. George Wehrwein and his collaborators authored a number of early articles for publication in the University's Journal of Land and Public Utility Economics. Additionally, the New England states, because of the importance of recreation to their individual economies, also became interested. In all, however, only five journal articles and a handful of popular items apparently comprise the whole of the acces sible literature on the economics of recreation during the 1927-1946 period.1 An article published by Wehrwein^ in 1927 was probably the first article dealing with recreation to be written by an economist for other economists. This article described many aspects of recreation problems which are still central to current issues. For example, Wehrwein noted the emerging problems of urban congestion, the public-private conflict, and the need for preserva tion of natural wonders before they were spoiled by population encroachment. ■'■Paul W. Barkley, "The Development of Research in the Economics of Recreation" in Cooperative Regional Re search Technical Committee for Project No. WM-59, An Eco nomic Study of the Demand for Outdoor Recreation, Report No. 1 (San Francisco, 1968), p. 3. ^G. S. Wehrwein, "Some Problems of Recreation 62 In 1933, Wehrwein and Spilman-^ reported the re sults of a study conducted in the recreation-oriented portions of the Great Lakes region. During the final years of the predepression era, large blocks of land in the Lakes region were subdivided for recreational devel opment and the establishment of summer homes. This speculative venture, however, was accompanied by fraud, faulty descriptions and misleading advertising. With the depression that ensued, much of this land reverted to the government for back taxes and many local commu nities were faced with the double burden of lowered taxes and increased costs of servicing platted land. The key points of the article were the problems of developing equitable assessment and taxation techniques and the severe difficulties of land administration when it is used for recreational purposes. An article by Chidester^ in 1934 emphasized land use in its title but dealt primarily with the economic Land," Land Economics, Vol. Ill, No. 2 (May 1927). G. S. Wehrwein and R. G. Spilman, "Development and Taxation of Private Recreational Land," Land Eco nomics, Vol. IX, No. 4 (November 1933). ^L. W. Chidester, "The Importance of Recreation as a Land Use in New England," Land Economics, Vol. X, No. 2 (May 1934). 63 | impact of recreation. Although the study was not based j i on new methodologies or investigative devices, it was a I specialized analysis of the recreation industry, and j i stressed the importance of recreation in creating em- j ! ployment, income and land value. Unfortunately, j i Chidester was not able to benefit from the economic base | j work made famous by Homer Hoyt and others later in the j i 1930s. Chidester obtained information on employment, in- I come and place of residence of recreationists vacationing | i in the New England area and only slightly more data and slightly different manipulations would have been needed to develop an economic base study. | In 1942, Wehrwein and Johnson^ collaborated on an article again dealing with recreation in the Great Lakes region. The article was purely descriptive and similar to the 1927 study by Wehrwein. However, the new feature was the focusing of the research on the individ- j ual tourist. Individual interviews were used to deter mine various factors and characteristics of tourists and some of the conclusions were that: older people stay in recreation areas for longer periods of time; and, tour- ists buy negligible amounts of produce from local farmers. ^G. S. Wehrwein and H. A. Johnson, "Zoning Land for Recreation," Land Economics, Vol. XVIII, No. 1 (February 1942). 64 | : ! While these and other items are noted, the major points were that many externalities associated with recreation j were severe enough to warrant social control and that ! zoning was a particularly useful tool in controlling them. | 6 ^ Another Wehrwein-Johnson article in 1943, also | 1 dealing with the Great Lakes region, took on the char- | acteristics of a demand study. In this article, attempts were made to determine the qualities and features which i prospective tourists look for in making decisions re garding recreation. Further, although degrees and direc- ; tions of causality were not clearly isolated or sub stantiated, it was asserted that after ten years, a major j I dam in the region gave rise to the installation of a num- 1 ber of all-year, summer and fishing resorts, as well as J many secondary businesses. Thus, the economic literature concerning outdoor recreation in this period followed an evolutionary pat tern. The early writers (1920s) did little more than I isolate problems of allocation, impact and institutions associated with outdoor recreation in a market-oriented ^G. S. Wehrwein and H. A. Johnson, "A Recreation Livelihood Area," Land Economics. Vol. XIX, No. 2 (May 1943), pp. 193-206. economy. Later writers in the 1930s and early 1940s began to provide empiric content to their efforts but were restricted to description. Recommendations were j notably absent and the use of theory was extremely i limited. Recreation Demand and Benefit | Evaluation Methods j Starting in the late 1940s, various approaches ! I to estimating recreation demand started to appear in the economic literature, and there was considerable growth in researcher's sophistication. The difficulties of using traditional constructs were noted and expedients were devised to circumvent these problems. Most of the ! approaches, however, were aimed toward finding solutions I "for the relatively esoteric questions of whether the j benefits to society from outdoor recreation can or cannot j be measured."7 They were developed in response to the need felt by authorities charged with the provision and administration of recreation facilities for some quanti tative means of appraising their actions.8 ^Warren C. Robinson, "The Simple Economics of Public Recreation," Land Economics, Vol. XLIII, No. 1 (February 1967), p. 71. ^David W. Seckler, "On the Uses and Abuses of Economic Science in Evaluating Public Outdoor Recrea- 66 | For example, following World War II, with the re lease of gasoline rationing and other travel restrictions, participation in outdoor recreation increased and visits ; to national forests, national and state parks and other | public recreation areas rose rapidly. Although available j statistics indicate that attendance at major types of public recreation areas had been increasing steadily : since 1910, the greatest increase occurred since the end of that war.^ With the increased demand pressures on major pub lic recreation areas, there was increased concern, partic ularly among the various federal agencies responsible for ! providing and maintaining recreation and natural resource I areas (e.g., the National Park Service, the Forest Serv- I j ice, Bureau of Land Management, etc.). As outdoor recre- ! ation gained in importance relative to other uses of resources, a measure was needed to establish priorities ! among projects or the various uses of a given resource. In economic terms, for example, a park for the purpose i | ------------------------ t tion," Land Economics. Vol. XLII, No. 4 (November 1966), p. 485. ^Marion Clawson and Jack L. Knetsch, Economics of Outdoor Recreation, Resources for the Future (Baltimore: Johns Hopkins Press, 1966), pp. 43-44. 67 of recreation would be justified if the incidental bene fits from recreation which resulted from this use made the difference between total benefits and costs higher than net benefits from the next most desirable project. Thus, various approaches were developed to determine the demand and benefit conditions of outdoor recreation. These techniques, however, were aimed primarily at placing a value on the benefits derived from outdoor recreation for use in benefit-cost analyses to establish priorities 1 1 among projects or resource uses.11 Some of the most widely used methods may be clas sified into two general categories — the market-benefit and demand schedule estimation methods. Market-Benefit The market-benefit methods attempt to measure recreation activities in terms of the dollar value of -^For excellent reviews of these various methods | and their applications, see L. J. Lerner, "Quantitative Indices of Recreational Value," Proceedings of the Com- j mittee on the Economics of Water Resources Development, ; Western Agricultural Economics Research Council and West- i ern Farm Economics Association (Reno, Nevada, 1962), pp. 55-80; and Omer L. Carey, "The Economics of Recreation: i Progress and Problems," Western Economic Journal, Vol. | III, No. 1 (Spring 1965), pp. 172-181. ■^Warren C. Robinson, "Economic Evaluation of Out door Recreation Benefits," in Outdoor Recreation Resources Review Commission, Economic Studies of Outdoor Recreation, ORRRC Study Report 24 (Washington, D.C.: Government 68 goods and services generated by recreationists and use these values as proxy indicators of recreation benefits. For example, the dollar sum of market transactions such as expenditures on outdoor recreation equipment, the I business generated by tourist facilities located near recreation sites, the value added by special local indus tries catering to outdoor recreationists and increases in i I land values around the recreation site comprise the meas- ! ure of benefits. I I Implicitly, these approaches maintain that the I i price-utility relationship of the market transactions j approximate the benefits of outdoor recreation to the i users. In general, however, they add little to an under standing of the demand and benefit conditions of outdoor recreation. These approaches attempt to move directly from dollar market sums to recreation benefits. No atten tion is given to the key issue, namely, that it is the recreation activity per se and not the associated factors that result in the utility to the recreationists. An other shortcoming of the market-benefits measures is that a systematic relationship between the amount consumed and various causal factors is not considered. It is this Printing Office, 1962), p. 49. 69 vital issue, an analysis of a demand relationship and all that it entails, that is the basic tool of economic in- i vestigation in determining benefit conditions. In fact, | | the market-benefit measures offer no analysis to generate ! ; j a recreation demand structure, to quantify demand or to j ! 1 9 ' j specify the dynamic properties of the demand function.^ i Before a benefit measure for recreation can be j obtained, a demand relationship must be established, re- j | ! lating quantity of recreation (in user days or visits) to ; I i the important causal variables of recreation activxty. j This requirement prompted serious research by economists : to develop an approach that would provide a sound theo- , retical basis for estimating and predicting recreation ! demand conditions. As a result, attempts were made to j develop demand schedules for outdoor recreation. I Demand Schedule Estimation : The development of simulated market demand sched ules for outdoor recreation is associated with the "travel-cost" method devised by Hotelling and applied by by Clawson. It is based on a gravity analysis model | which uses physical distance as the key variable in •^Ibid. , pp. 61-65. 70 determining recreation activity. In social science re search, this technique has been used to establish em pirical relations between economic masses (e.g., city population and income aggregates) and the frequency of economic movement (e.g., transportation volume, income flows, volume of communication messages). Hotelling and Clawson extended this concept to outdoor recreation, as did other economists. Most valuation procedures based on this method begin with the construction of a simulated market-demand schedule for outdoor recreation depicting the functional relation between quantity demanded (use of a certain site in terms of number of visitors) and market price or cost (the expenditures required to enter the site in terms of entrance fees or travel costs). The concept is that with a typical demand curve having a negative slope, the num ber of visitors will decrease as the price or cost is increased. A certain recreation site would be priced at zero and the price raised by small increments to the point where no one would visit the site (with Y inter cept) . The quantities demanded at these varying prices would then be recorded and the demand curve formulated. 71 Due to the lack of conventional market pricing noted previously, the development of demand curves for outdoor recreation has been hampered and the manner in which they are derived has been a matter of debate. Either price and use variations over a small range are extrapolated to the appropriate intercepts, or price variations imputed through estimates of travel costs, recreation equipment costs, or both. As a result, a num ber of alternative approaches have been developed in at tempts to derive a simulated market-demand schedule for outdoor recreation. In general, however, they may be classified into two major groups: the indirect approach, which has been used most frequently; and, the direct approach. Recreation Demand Studies Using the Indirect Approach These efforts revolve on the basic relationship between distance and demand and the Hotelling-Clawson approach is the accepted generalization and interpreta tion of the problem. There have been a series of addi tions and refinements to the original work, yet the main idea remains that it is the functional demand schedule for recreation that provides the basis to predict demand 72 levels at given recreation sites and subsequently to establish a measure of benefit. I j Hotelling Approach | The approach developed by Hotelling^^ in 1947 is | the forerunner of today's recreation resource valuation procedures. It was the first approach that set forth a methodology to formulate a demand curve for a recreation I area which would then be used to place a value on the j recreation benefits derivable from the site. Hotelling suggested that a demand curve for a recreation site be derived by using the travel costs of recreationists to get to it as a measure of price. He | assumed that the people using the area would receive i I benefits which were roughly proportional to the expendi- ! j tures of the most distant users of the area (marginal users' travel costs thus equal average benefits). For those living closer to the area, however, travel costs would be less while the benefits received would be the | same. Thus, a measure of surplus would be e n j o y e d . !3h . Hotelling, letter to Director of the United States National Parks Services (1947) as cited in Robin son, "Economic Evaluation," pp. 45-69. ^ I b i d . , p . 6 6 . 73 In Hotelling's method, concentric zones are de fined around each recreation site so that travel costs to the site from all points within each zone are approxi mately constant. The number of persons or a representa tive sample of persons entering the recreation site i during a year are listed according to their zone of ori- j gin. It is assumed that the pleasure derived from j visiting the site is at least equal to or worth the cost j I of the trip (otherwise the trip would not be made), and that the cost can be estimated fairly accurately. i Assuming that the benefits are the same irrespec- j tive of the distance, there is a consumers surplus con- | sisting of the differences in transportation costs for those living near the site. Then, by comparing the costs of visiting the recreation site from each zone with the number of people coming from the zone, together with the total population of the zone, a statistical demand func tion for the site can be derived. This function may be integrated to derive consumer's surplus. Area attendance variations due to entrance fee impositions may also be estimated from the demand function. Hotelling also stated that the problem of demand between different parks could be treated by questioning the people sampled as to 74 which other recreation areas they attended during the sampling period. This would give a set of demand func- | tions in place of a demand c u r v e .15 One of the clearest expositions of the applica tion of the Hotelling method was given by Scott.I6 The j ; j | diagram he used is reproduced in Figure III-l. ! 1 When the number of visitors per thousand popula- | • ! tion from each zone is plotted against the travel costs i ! from each zone, as shown in Figure III-l, a curve like j | A^H is formed (assuming it is linear). Hotelling suggests j this curve be treated as the demand curve for visitors j from zone 0 because it shows the number of visitors will- ; ing to pay each cost. Thus, if people in zone 0 had to j | pay a toll equal to the travel costs actually incurred I by visitors from zone 2 (AC), their visits per thousand I | population would fall from AA^ to CC^, assuming income and tastes among zones were equal.^ 15Ibid. ^Anthony Scott, "The Valuation of Game Resources: Some Theoretical Aspects," in Canadian Fisheries Reports, No. 4 (Ottawa: Department of Fisheries, Queens Printer, 1965), p. 28. 75 TRAVEL COST/MILE 7 6 5 4 3 2 0 Z O N E S (M I L E S ) NUMBER OF VISITS PER 1,000 POPULATION Figure III-l — Travel cost approach 76 Scott gives the equation for the Hotelling demand curve as: vi - f(C±) = 0 = ^ I ; where: ! | Cj_ = average cost per visit of traveling from j zone i j = visits per year from zone i j ! i Such a curve might be derived by regressing visits per j ' IQ I j zone on travel costs. j At this point, it is important to note that the ; concept of a consumer surplus is applied in a variety of economic approaches used to derive a demand curve for and j | to evaluate recreational benefits of a particular site. | Consumer's surplus, as identified by Alfred Marshall, is | | the difference between what a consumer actually pays for i | a commodity and the amount he is willing to pay for it. Individual consumers surpluses are added together by as suming that the utility of income of all individuals is identical. ! As shown in Figure III-2, aggregate consumer's ! ! surplus is the area P^PX under the demand curve when the price is OP^; it is P2PY when the price is OP2. The ^®Ibid. , pp. 28-31. PRICE P CONSUMER SURPLUS P. P 2 0 Q, Qg QUANTITY DEMANDED Figure III-2 — Consumer surplus 78 value to the consumer of the quantity purchased OQi is | OPXQi. This valuation rests on the theory that when the | price is OP-^ and the purchases are OQi, the value of the | marginal unit is OP-^, While the value of the first unit ! ! j consumed is greater than this amount or OP. Thus, the j j total value is OPXQi. The cost of this quantity OQ, how- I i ever, is only the area OPiXQi* The difference between j ; the area OPXQi and OP1XQ1 is the consumer*s surplus or ! PlPX on the quantity OQi.^-^ This concept is used in j i j Hotelling-type models used to evaluate the demand for and j benefits of recreation sites. ! i The approach developed by Hotelling did little i to overcome the previously noted major problems associated ! with measuring and predicting outdoor recreation demand. Among other deficiencies, it is based upon consumption, rather than demand, assumes identical income and tastes j 7 j of recreationists, and does not consider supply. Never theless, it was the first methodology that attempted to derive a demand formulation within the constructs of con ventional demand analysis and as such, represents a major contribution to the field of recreation economics. 19 Donald S. Watson, Price Theory and Its Uses (Boston: Houghton Mifflin Company, 1963), pp. 58-59. 79 Trice-Wood Approach Trice and Wood^O utilized Hotelling's technique in a study to determine recreation demand and benefits from proposed reservoirs in California. They derived a j demand curbe based on distance travelled to a particular site and assumed that persons living close to a facility enjoyed a surplus benefit in that they get use of it at a lower cost than do the more distant users. Their ap- ! proach, while similar to Hotelling's model and suffering from the same deficiencies, was unique in that it was the first practical application of the travel-cost method, 1 actually placing a value on a recreational resource, and | "still soundly rooted in the conceptual framework of i | economics."21 j Through a series of interviews, Trice and Wood i | obtained estimates of the number of visits to certain ! recreation areas made by recreationists within various i travel time zones from the areas. They then computed an average cost of travel per visitor day and plotted a de- ; mand curve for each recreation area. By establishing a I 20 Andrew H. Trice and Samuel E. Wood, "Measurement of Recreation Benefits," Land Economics, Vol. XXXIV, No. 3 (August 1958), pp. 195-207. 21 Barkley, op. cit., p. 9. 80 maximum or "bulk line" market value per area at the 90th percentile, they next assessed the value per visitor day of recreation at that area, which was equivalent to the observed unit of travel cost below the 90th percentile. An approximation to a consumer's surplus value or "free benefit" per visitor day was calculated by subtracting the median travel cost per visitor day from the 90th per- I oo i centile "bulk line" travel cost per visitor day.^ ; | The Trice-Wood consumer's surplus may thus be j expressed as: I | n 5 (cn-Ci) [f(c±)] ! C ! | where represents the lowest unit travel expense, Cn the unit travel expense of the 90th percentile, and the unit expense of the intervening percentiles. This method has been criticized for selection of the 90th percentile as the cutoff point. In fact, selec tion of a cutoff point is the main drawback of all the travel-cost models for recreation benefit evaluation. Trice and Wood give no rationale for using it and their reference to Hotelling's approach does not help to iden tify this point. Further, by using the 90th percentile, 2^Trice and Wood, op. cit., pp. 204-206. they cut off the high travel cost per visitor day, i.e., I the top portion of their demand curve; a few visitors j from remote zones could greatly alter the consumer sur- j I i I plus attributable to a site. Studies using the consumer ! j surplus concept as a measure of benefits are confronted I by the necessity to ensure that all /isitors, particularly the few from remote zones, are included in the estimation. In addition, as noted by L e s s i n g e r 2 3 in a critique! i j of the Trice-Wood approach, the validity of their con sumer's surplus concept is questionable. He reasons that j when people choose to live near a recreation area they | i often do so at the sacrifice of living closer to another j ] place (e.g., place of employment) and thus an opportunity ! cost, must be assessed to the recreation area because of j j its location. Consequently, an economic substitution effect between home and recreation area locations may exist with the outcome being an absorption of consumer's surplus by market action.^4 Hines^S also criticized the Trice-Wood model be- i i ^Jack Lessinger, "Measurement of Recreation Bene fits — A Reply," Land Economics, Vol. XXXIV, No. 4 (November 1958), pp. 369-370. ^ L a w r e n c e g. Hines, "Measurement of Recreation Benefits," Land Economics, Vol. XXXIV, No. 4 (November 1958), pp. 365-367. 82 cause the "costs of travel" index they developed does not i ! remove differences in users' incomes. Further, the as- | sumptions of identical consumer preferences and the con- | stant marginal utility of money among visitors to a j j recreation area are unrealistic. However, he feels that ; j i a solution would be to ". . . appraise benefits in a way j to discriminate between various users and to sum up the j different individual evaluations. . . ."2^ Clawson Approach The methodology developed by C l a w s o n ^ " ? in 1959, ; ! although still a form of the basic travel-cost model, is ; i i often considered one of the most important contributions i made to the field of recreation demand analysis. As a i result, his "demand curve" approach has been utilized in j many subsequent studies undertaken since that time. 1 Clawson reasoned that to estimate a demand sched- I | ule or curve for a given recreation area, consideration cannot only be given to the total attendance for a single year and to the average cost for all visits to the area | 26Ibid.. p. 367. j j 27jyiarion Clawson, "Methods for Measuring the De- j mand for and Value of Outdoor Recreation," Reprint No. j 10, Resources for the Future, Inc. (Washington, D.C., j 1959). (as is done in Hotelling's travel-cost approach); this I ! gives only a single line in the tabular data and a single | point on the curve. What is needed is a range in the ob- | servations in order to trace out a significant portion of i the curve. For this purpose, he suggests primary reliance j ; i | on geographic analysis, using differences in number of j visits and in cost per visit from different zones to estimate the basic relationship. In so doing, he accepts j the existing patterns of population distribution, income of the population, location of recreation and transporta tion facilities and other f a c t o r s . Clawson thus rejects the major assumptions in : i i Hotelling's travel-cost approach because it assumed what i ' | I he termed an unplausible relationship, i.e., that all ! people have identical preferences with respect to visit- I | ing a given recreation area. Further, all visitors and potential visitors are assumed to value the area identi cally and this was taken to be equal to the travel costs of the most distant visitor. Clawson thus feels his for- ! mulation allows the more realistic view that individuals value recreation areas d i f f e r e n t l y . 29 ^®Clawson and Knetsch, op. cit., p. 64. 29Ibid. 84 In developing his methodology, Clawson suggests that the estimation of a demand curve for a recreation area proceed in two stages. The first stage involves the I derivation of a demand curve for the "total recreation experience" which is estimated from actual experience of individuals engaged in outdoor recreation; the second j stage is the derivation of a demand curve for the "recre- | ation opportunity per se" from the first demand curve.^0 j i The construction of a "Clawson demand curve," in outline, begins with the delineation of concentric equi- j I distance zones around a given recreation area. An ! estimate is then made of the cost and time required to j get to a given recreational facility from each concentric | tributary zone. These times and costs are then related to the proportion of the population in each of the trib utary zones which actually visits the given facility. Prom this analysis, a total monetary cost of visiting the recreational facility can be estimated for each tributary zone. A demand curve is then constructed by plotting average cost per visit against the number of actual per capita visita. In effect, this gives the quantity of ! recreational experience demanded at various prices. ■^Clawson, op. cit. , p. 13. Clawson claims that this demand curve represents an | "approximation to the demand curve of theory for the rec- i j reation experience as a whole. 1 1 ! To derive the demand curve for the given recrea- I tional facility from the demand curve for the total j ! i j recreational experience, Clawson makes two assumptions. : | The first is that the users of the recreational facility j would view an increase in entrance fees in the same way j ! I as an equal increase in the total travel cost of a visit, i ; The second is that the visitors from one zone would be- i j i have similarly to people in other zones, if costs in time j | and money were equal. Using these assumptions, the effect j I of an increase in user fees can be estimated by postu- I | lating increments in total cost and reading off the per ! capita rate of visits which could be expected from each tributary zone. These new per capita rates for each zone, multiplied by the populations of the zones, would yield I an estimate of the total number of visits. From similar calculations of the estimated number of visits at each | | level of increased fees, a new demand curve can be | plotted. Clawson contends that this curve approximates the true demand curve for the recreation opportunity ^Ibid. , p. 18. 86 itself. Further, that from such a demand curve, it would be possible to predict the number of visits to an existing facility which would result from either a reduction in travel cost or travel time (e.g., from building better access roads) or a change in entrance f e e s .32 Clawson's method is simple but beset with many deficiencies. One is the difficulty of ensuring that the frequency of visits from remote zones is taken into con sideration. Another is the need to include substitutes or, what is more pertinent, the price proxy of substitute sites. Charging the full cost of travel of the trip to the site implies that no other site was visited. A trip may be undertaken to visit many sites but such trip diver sions have not been accounted for in Clawson's evalua tion.33 Aside from the complication of visits to more than the site being evaluated, the assumptions that the pattern of tastes and income among recreationists do not change is open to question. Further, this method is not applicable to the problem of predicting demand for new recreational facilities with which the public has no 33Ibid., pp. 23-30. 33scott, op. cit., pp. 28-31. 87 familiarity. The result of one such demand analysis j would not necessarily be valid for another recreational facility, nor would it remain valid for any given facil- ! ity for very long. In a few years, the basic factors I underlying the analysis might change dramatically. For j ; these reasons, Knetsch-3^ suggested the inclusion of addi- j ' tional variables to the basic travel-cost model. Such j ! variables might be income, some measure of the avail- I ability of substitute areas, and congestion. A formula | incorporating all of these variables would be more useful I in predicting demand for new recreational facilities than j I I . i Clawson's model. Knetsch's approach is reviewed later 1 ! in this chapter. i ' I j i | As noted earlier, Clawson's method had often been i ! considered one of the most important contributions to recreation demand analysis. In reality, however, after adjusting for different concentric travel time zones, I | Clawson's derivation for the whole recreation experience ; is the same as Hotelling's. In fact, a number of authors j | refer to it as the Hotelling-Clawson model. Further, as ! | noted by Barkley: ■^Jack L. Knetsch, "Outdoor Recreation Demands and Benefits," Land Economics, Vol. XXXIX, No. 4 (November 1963), pp. 386-396. 88 The Clawson document remains somewhat perplexing. It is honest. It is relatively complete. It uses secondary data to make a useful point. But it does not make a genuine original contribution to either the application of economic theory to recreation problems or to research methods in resource economics. Nor can it be claimed that the Clawson document added to our abilities to make decisions and policy. The major contribu tion of the work seems to be the name of the au thor and the likely clientele of his efforts. At the time he was writing . . . Marion Clawson was an internationally known resource economist of impeccable repute. The fact that he lent his name to the subject matter meant wide readership and subsequently increased interest in the topic. . . . It is quite likely that the literally hun dreds of papers, articles, monographs and re search projects which have arisen in the decade of the 1960s can trace a good deal of their origin and philosophic roots to statements made by Clawson in 1959.35 Other Approaches Using the I Hotelling-Clawson Method | There were a number of other empirical studies | made in the middle and late 1960s that attempted to de- j velop approaches to recreation demand estimation utilizing | Clawson's method. Although these approaches basically suffer from the same deficiencies as the Clawson model in terms of demand prediction, there was a noticeable ! trend toward increasing efforts to incorporate socio- ! economic demand estimates and determine the effects of I ^^Barkley, op. cit., pp. 9-10. these variables on recreation demand. This trend was probably due to the results of the outdoor recreation studies undertaken by the Outdoor Recreation Resources Review Commission (ORRRC) which began being published in 1962. The ORRRC works are reviewed in the following chapter. Of the approaches undertaken in the Clawson tra dition in the middle and late 1960s, only a few stand out as being of lasting interest. These studies, listed by author, are briefly discussed below. Lerner.— An example of an approach modeled after Hotelling-Clawson formulation, which became the vogue in the early 1960s, is that developed by Lernerin 1962. Lerner extended this method to derive a demand curve for a recreation site from each concentric distance zone around the site and calculated a consumer's surplus as a measure of benefits. Lerner1s approach differs from the Hotelling-Clawson method only in that visits and costs, the dependent and independent variables, respectively, -^Lerner, op. cit. 90 are expressed in terms of visitor days rather than the number of visits or cost per visit. With Lerner1s approach, linear regression analyses were used to formulate the demand curve for a site. In outline form, travel cost, C-j, and visitation rates (—) j P ! (number of visits, V, per 100,000 population, P) are 1 observed for various concentric travel time zones around ! the site and the results presented in the regression | equation: j Ci = a + b (—). for zones i = 1 ... n p i I or (|). = (Cj--a) P 1 E i To predict the visitation rate from any zone where a I i fee, f, is levied, total costs, Ct, are broken down j S into two components which may be expressed as: I n - n . - - f - (ci + f-a) | Ct - C1 - f V i - b----- l ! P . or ^ = _i (C-i + f-a) | 1 b 1 I i j or, in functional form, = V (Pj_, C^, f) j j j The number of visits from any zone i is then a function of population, travel costs and fee. However, since 91 population and travel costs are known for a given zone, the demand curve for that zone becomes Vi - V(f). 3^ Individual zone benefits, as assessed by Lerner, are the integral of the estimated demand function for the zone between the fee charged and the estimated price | (fee) at which demand from that zone would equal zero. | , i That is, benefits from zone i minus f are the ceiling fee ' i I where demand would equal zero. The consumer's surplus j i i for the site is obtained by repeating this calculation j for each zone. The benefits measures by this method are, therefore, the willingness to pay hypothetically collect ible fees by a discriminating monopolist.38 | Aside from having the basic deficiencies noted i I for the Clawson approach, the accuracy of Lerner1s method is also lessened because of the indirect way of measuring i the discriminating monopoly revenue which must be used ! in the absence of a market price for recreation. Thus, | consumer surplus will not be the same as the revenue ob- | I tained by a discriminating monopolist. Also, due to the lack of homogeneity between the distance zones, it is I 1 difficult to distinguish between the income effect and S^Lerner, op. cit., pp. 7-9. 3®Ibid, p. 9. 92 other effects generated by differences between distance zones. However, Lerner feels that the income effect is probably small and the reliability of this method may be i improved through the advent of refined methods.39 j i | Knetsch.— The approach developed by Knetsch^O in | 1 9 6 3 was also modeled after Clawson's method. Knetsch1s i main contribution was the specification of the functional i j relations of the variables affecting recreation demand j i and the addition of other variables to the basic travel j | cost model. I Knetsch derives the Clawson demand curve through i the use of a hypothetical example. The principle under- j lying this method is the assumption that observed pat terns of visits to parks by residents of different cities can be explained in terms of the travel cost of reaching the parks. Visits per unit of the total population de creases with increasing travel costs. A demand curve can be constructed from number of visits by origin and travel cost information in the following way. With an initial entrance fee of zero, the number of visits represents the zero-cost point on the demand curve. By imposing an 39Ibid., p p . 6 9 - 7 5 . ^Knetsch, op. cit. 93 additional price in the form of an entrance fee, the num ber of visits from a particular city would decline as if the travel cost were increased by the amount of the en trance fee (assuming homogenous preference functions). I By postulating increments in travel cost and calculating i I i the number of visits that would result, a demand curve j | can be traced out and plotted on a graph. Clawson's de- j mand curves may thus be expressed as V = f (C) where V, ! the rate of visits to an area, is a function of C, the i j I costs of such v i s i t s . j However, due to the previously noted limitations j of the Clawson approach, Knetsch suggests the addition of i further variables to the basic travel-cost model. These i I variables include: Y, the income of population groups; S, substitute recreation areas that might be relevant | for any group; and, G, some measure of congestion at the ! recreation area. The addition of these variables would i then extend Clawson's equation to V = f (C, Y, S, G) . This equation, Knetsch asserts, would mitigate the weak ness of the original formulation which assumed that the | demand relationship would remain the same in each area. 41Ibid., pp. 388-390. 94 Further, he claims it would be useful for predicting de- ! 42 | mand for new recreation resources. ! Knetsch also notes that the effects of fee col- ] lection can also be predicted from a demand curve of the I ; I type prescribed, since it measures the consumers' will- | i i | ingness to pay for recreation. To the extent that the j i ! consumer regards an admission fee as different from a | 1 travel cost, however, the model is weakened.43 j j Another flaw in the travel-cost method pointed | . I j i i out by Knetsch arises from the fact that it does not take j i ! into consideration the time required to travel to a site, j I Travel time is closely correlated with travel cost. How- j ever, the existence of this double factor results in a systematic underestimation of true demand in the process of postulating increases in travel cost and calculating the corresponding expected number of visits. Although | travel time will not change, travel cost will be increased | through the imposition of a user f e e .44 ^ Ibid. , pp. 390-392. 4-^ibid. , p. 394. ^Ibid. , pp. 394-395. 95 Wennergren.— W e n n e r g r e n 4 ^ presented a model in the Clawson tradition to derive a demand curve for out door recreation using pleasure boating for expository purposes. His main argument was that even in the absence of market pricing, use of water for pleasure boating in- j volves user costs and that when such costs are properly j | delineated, they can serve as proxy prices even for non- j ! market priced recreational commodities. He suggests that j | in addition to the travel costs to and from the site as j used in the Clawson approach, expenditures incurred at | the site should also be included. These total costs I then, not just the travel costs, constitute the relevant j expenditures and are the marginal costs of the boating experience. As such, they simulate a price for boating and are the principal determinants of the quantity that will be taken.4* > Wennergren bases his hypothesis on the following assumptions: First, the boater spends his income and other resources in such a way as to maximize his total derived utility or satisfaction. Second, E. Boyd Wennergren, "Valuing Non-Market Priced Recreational Resources," Land Economics, Vol. XL, No. 3 (August 1964), pp. 303-314. 46Ibid., p. 305. 96 the boater has perfect knowledge regarding the various costs of boating and the utility or satisfaction that he receives from the differ ent quantities that may be taken. Third, the boating experience generates a total utility function which at some point encounters dimin ishing marginal utilities. . . . Fourth, the units of utility and cost are equivalent and a net utility can be derived. Fifth, major decisions pertaining to individual boating trips are made prior to departure, and the boating activity is the causal agent in the individual's decision to undertake the outdoor experience.47 The logic of his argument is that a prospective boater first must make the decision to join the boating force and will weigh the cost of purchasing a boat and the re lated equipment against the expected benefits or utility he will receive from the activity. A decision to join the boating force suggests that the expected utility from joining is at least equal to the cost. Subsequent simi lar decisions must be made each year following the initial purchase. However, the cost considerations differ. The original purchase costs are now fixed and considered by Wennergren as irrelevant to the decision. The important costs are now the annual fixed costs, such as license fees, insurance, and so forth, which now constitute the relevant cost considerations to join the boating force. 47Ibid. 97 A positive decision to join thus suggests that the ex pected addition to total utility will at least equal the added costs. Even these costs, according to Wennergren, are irrelevant to the boat owner's decision as to how many times he will use his boat during a season. The relevant costs are the variable expenditures involved in travel to and from the site plus those incurred while at the site. These variable costs incurred on a per trip basis are the marginal costs and the boater, if he makes the trip, ex pects the additional or marginal utility to be derived to be equal to or greater than the marginal costs. Therefore, given the marginal cost and a diminishing mar ginal utility function for the boating experience, the boater will take the total number of boating trips during the season such that the marginal cost will equal the marginal utility. It is at this quantity of boating that his satisfaction from the total boating experience will be maximized. Further, whether a boater decides to go boating during a given year depends on the assumption he knows whether these criteria can be satisfied.4® ^®Ibid., pp. 305-307. 98 Wennergren then suggests that statistical demand curves be constructed by following Clawson's approach using the marginal costs as the price variable and the number of boating trips made by the boater during a given j i i period as the quantity variable. Aggregate demand curves I i I would be formulated from individual boater demand sched- I i ules considering the origin of boaters and the location | of specific boating sites.49 ! I This approach still suffers from the same weak nesses as the Clawson model. In addition, the exclusion j of fixed recreation expenditures is questionable. Accord-i ing to McNeely, such expenditures, including depreciation ! on recreational equipment are all relevant to the outdoor i | recreation experience and should "be included in an esti- i | mate of the 'proxy' price."50 in a later article, how- | ever, Wennergren^l attempts to support his exclusion of | such costs by demonstrating the small effect total equip- 49Ibid., pp. 307-313. 5°John G. McNeely, "Estimation of Demand for Water Based Outdoor Recreation Activities," An Economic Study of the Demand for Outdoor Recreation, Report No. 1, Coopera tive Regional Research Technical Committee for Project No. WM-59 (1958), p. 49. 51 E. Boyd Wennergren, "Surrogate Pricing of Out door Recreation," Land Economics. Vol. XLIII, No. 1 (February 1967), pp. 15-116. ment costs and total annual equipment for a boating season | using a questionnaire and record of costs. The number of I I | individual boating trips taken to all sites during the season were then regressed against travel and on-site j costs per trip, total equipment costs, and total annual costs. The only variable found to be statistically sig- i : | 1 nificant was travel and on-site costs per trip, i.e., | go ; variable costs. | A more basic criticism of Wennergren's approach, : however, is the assumption that the marginal utility j curve is identical with the demand curve. This assumption | has been attached by both S e c k l e r 5 3 and P e a r s e . 5 4 | Seckler, for example, does: I . . . Not believe that statistical demand curves measure the utility function of recreational fa- | cilities [or of anything else for that matter]. I Statistical demand curves are simply a conven- i ient way of summarizing a set of empirical ob servations in a functional statement. Any attempt to squeeze connotations regarding util ity or welfare out of such data is a dubious practice at best. . . . Statistical demand curves 52Ibid. 53seckler, op. cit. S^Peter H. Pearse, "A New Approach to the Evalua tion of Non-Priced Recreational Resources," Land Eco nomics. Vol. XLIV, No. 1 (February 1968). 100 do not measure the diminishing marginal utility of recreational facilities nor of any other com modity, rather they reflect the diminishing marginal utility of income. The slope and posi tion of statistical demand curves is largely a function of income distribution.55 Gray and Anderson.— In 1964, Gray and Anderson,30 | i in studying the economics of recreation in south-central j ! i New Mexico, applied Clawson's method to determine the ; value of a recreational resource. In an area with mul- j tiple recreation uses, participants were identified as to type of recreation activity in which they planned to par- ! ticipate (general, race track or fishing) and their origin j (local or non-local).^7 Demand curves were generated from data obtained through interviews of recreationists who estimated the actual and maximum amounts they would spend in a given activity. The value of the resource product was deter- i mined to be the sum of the areas under the demand curves generated by local and non-local visitors. A similarity S^seckler, op. cit., p. 487. 5^James R. Gray and L. Wayne Anderson, Recreation Economics in South-Central New Mexico, Bulletin 488, Agricultural Experiment Station, University Park (New Mexico: New Mexico State University, 1964). 5^Ibid., p. 9. 101 was found to exist between the actual amount spent by non-local visitors and the maximum amount that local visitors indicated they would spend. This similarity was indicative of the actual amount that recreationists would i spend to obtain the activity. The difference between the j two values generated was the consumer's surplus of the j ! resource accruing to the local population or the value of ! the resource. i Brown. Singh and Castle.— A prominent demand j 1 study utilizing Clawson's method was undertaken by Brown, j Singh and Castle^ to determine the net economic value of j a sport fishery. In their approach, the authors used an elaborate sampling and interviewing technique to deter mine the average expenditures made by fishermen for dur able equipment (fixed costs) as well as individuals (or variable) costs incurred during a particular outing. In cluded in the individual trip or variable costs was a transportation item which, when coupled with information ^Ibid. , pp. 15-20. ^william G. Brown, Ajmer Singh, and Emery N. Castle, An Economic Evaluation of the Oregon Salmon and Steelhead Sport Fishery, Technical Bulletin 78, Agricul tural Experiment Station (Corvallis, Oregon: Oregon State University, 1964). regarding the residence of the individual fisherman, made possible the development of a demand curve as outlined i | by Clawson. i ! The authors go a large step forward, however, in j ; i I using the demand curve as a device for calculating the ! j j value of the natural resources which provide the fishing i opportunity. They made the distinction between fixed and ; : i i i variable costs since the marginal cost curve was assumed | ! I ! i to be influenced only by variable costs. Assuming the sport fisherman tries to equate marginal costs with mar- j | ginal utility, his "variable costs would indicate equality! between marginal costs and marginal utilities.1 1 Maximum net economic value then, if the resource owner (the state) I 7 I j behaved as a monopolist in a free market-type economy, I I would be at the point of unitary elasticity on the demand i ! curve.^ The authors recognized the many deficiencies in j | using this procedure. Further, due to data limitations, ! they used an alternate single equation model for esti- ! i mating demand (days of activity per unit of population) I by subzone based on average variable cost per day, ^°Ibid., p. 11. ^ Ibid. , pp. 28-29. 103 average family income and average miles per trip for the particular subzone. This method yielded an estimate of net economic value comparable to that derived by the Clawson method.^2 Boyet and Tolley.— The approach utilized by Brown, Singh and Castle noted above includes some reference to the effects of population density and income. It does not, however, treat these determinants of demand as com prehensively as the method used by Boyet and T o l l e y 6 3 in projecting demand for recreation facilities in North Carolina. This latter research traces to the Clawson- type demand analysis and uses distance and visitation data as a source of demand curves. The work, however, is advanced and refined in a number of ways. First, visitors to recreation areas were classi fied by reference to state of origin rather than Clawson's concentric zone method. This enabled the researchers to avoid the delineation of zones and, more importantly, made available scores of data series available for states 62Ibid., pp. 37-41. ^Wayne e. Boyet and George S. Tolley, "Recreation Projection Based on Demand Analysis," Journal of Farm Economics. Vol. XLVIII, No. 4 (November 1966). 104 | that would be meaningless in a concentric zone context. I Visits to various national parks from each state were i | estimated based on distance from the state, population i and per capita income. Originally, median age, median education, percentage of population in urban areas and j percentage of population that is white were included in | the analysis, but it was found that these variables added little to the usefulness of the model.®4 ! Boyet and Tolley also abandoned the visits per I capita style of research in favor of a multiple relation- I ship which includes an area's total population as one ! independent variable. Further, and perhaps the most sig- i i nificant innovation of the Boyet-Tolley approach, was the j I inclusion of demand shifters based on income distribution. I Here, the authors used an exponential function with the I power of the independent variable representing its elas ticity with respect to the dependent variable. Projec- i tions were made by substituting distance, projected pop- | ulation and projected income into the derived equations. I | Unitary elasticity for population meant that demand would i | not change proportionately based on population. However, j I | an income elasticity greater than unity inferred that a 64Ibid., pp. 984-987. 105 I greater than proportional increase in demand could be i attributed to increased per capita income. 3 | Estimates of cross-elasticities between recrea- i tion areas were used to determine if recreation facilities I were competitive in attracting visitors. The basic model I to estimate direct elasticities of the various independ- j i ! i ent variables was expanded to include distances from | I I respective states to the recreation facility for which a j I cross-elasticity estimate was desired. The cross-elas ticities arrived at in this manner may not be as accurate j as possible due to their exponential origin? however, j i | j their derivation does suggest a way in which they might j j be obtained and indicated need for further research.66 I j It should be noted that unlike many studies of | this era, the Boyet-Tolley effort was geared toward a projection of demand rather than an estimation of value. I The general conclusion is that demand projections made | in earlier studies and reports (particularly the ORRRC reports which will be discussed later) may be grossly j understated because of the omission of income distribution i | as a demand shifter. ! i 65Ibid., pp. 990-991. 66Ibid., p. 992. 106 It is also interesting to note that M y l e s ,67 based upon the suggestions of Boyet and Tolley noted above, investigated the effect of population, distance and alternative sites on summer group visits to recrea tion sites. An analysis of the various effects on rec reation site visitation were investigated through regres sion analysis and it was found that: Regression equations indicate that the number of visits to western Nevada lakes was inversely re lated to distance; however, parameters were not consistent for different lakes and for different home areas. The addition of a variable such as distance to another lake often increased the correlation significantly and sometimes makes a significant change in b-values of distance to the site. The wide variability in regression coefficients relating distance and visits per capita to lakes where a large percentage of the use is local suggests that each lake has a dif ferent distance or psuedo-demand f u n c t i o n . 6 8 Johnston and Pankey.— A current study of this same genera is the one by Johnston and P a n k e y . 69 working ^George A. Myles, "Competitive Aspects of Rec reational Areas," Proceedings 1967, Western Farm Economics Association (Las Cruces, N ew Mexico, 1967). 68Ibid_. , p. 251. 69W. E. Johnston and V. S. Pankey, "Use Prediction Models for Corps of Engineers Reservoirs in California," An Economic Study of the Demand for Outdoor Recreation, Report No. 1, Cooperative Regional Research Technical Committee for Project WM-59 (1958). 107 with data regarding recreational use at reservoirs in I I j California, the authors examined the importance of in come, age, education, urbanization and population as ! demand shifters. A number of multiple regression models | i j I ! were used to test relationships among these and other j I i | variables. Additionally, these authors added to the j ; i sophistication of recreation research by generalizing the Clawson model, using distance from the site as a proxy ; for price, and including several site-specific variables aimed at introducing recreation site quality as an impor- j tant determinant in projected demand. j ! In their approach, the authors attempted to de- j | scribe use at various reservoirs using a wide variety of | socioeconomic characteristics as potential demand shifters ; and the distance of the population in various counties | from the reservoirs. All socioeconomic characteristics ' | i except population were found insignificant in a prelim- I inary analysis of data. The significant variables repre- | senting reservoir characteristics were land area and | water surface area.^ i It was found that per capita use as a dependent I j variable may be an erroneous choice in analyzing sub- j ^Qlbid.5 pp. 23-24. 108 aggregate use since per capita use may not remain con stant with increases in population. Distance from the reservoir was found to affect day use differently than night use. Precision in estimating total use could, it was felt, be increased by predicting subaggregate use and then using the sum as the prediction of total use or demand. It was also concluded that using county data in recreation analysis offers many advantages in that it is easy to assemble and allowed specification of density and recreation alternative variables. Further, that models using county data may be more powerful than models using di stance zones.71 Travel Cost Savings.— An interesting variation on the travel cost model is the "travel cost savings" ap proach. This method was utilized in the work on the Meramec Basin Research Project72 to estimate the attend ance and recreational benefits of a proposed reservoir near St. Louis. Attendance data were first developed for existing reservoirs by regressing the distance (in air 71lbid., pp. 30-35. 7 2 Meramec Basin Research Project, "Recreation," The Meramec Basin. Vol. Ill (Water Needs and Problems) St. Louis: Washington University, 1961). 109 miles) from these sites against per capita annual attend- i ance (visitor days). The data were stratified by high i i J and low per capita annual attendance expectations, based ■ ] on population and reservoir characteristics. Prom these | regressions, the estimated demand schedule was computed : f o r t h e p r o p o s e d i m p o u n d m e n t . 7 3 i A consumer's surplus was then assessed on the ; basis of the travel cost savings of visitors who would ! have visited another more distant site or been willing to I travel farther if the proposed site had not been con structed closer to St. Louis. To determine the value of alternative sites, various entrance fee levels were as sumed. The maximum revenue obtainable was at the point of unitary elasticity on the demand curve, where marginal i I revenue was zero.^ ! Merewitz,^ in an expansion of the travel cost i savings approach, also formulated a demand curve and con- | sumers1 surplus for a reservoir. He expressed demand in ! terms of visits to the lake from a particular county as a i i 1 - ■ _ i r " ' - - - 1 ' ’ 1 ' 1 ' 1 1 - n I ^ I b i d . , p p . 5 - 1 5 . 74ibid., pp. 24-27. 75Leonard Merewitz, "Recreational Benefits of Water Resource Development," Water Resources Research, Vol. II, No. 4 (1966). j function of air distance from the population center to J the lake, total county population and population density. | j Since population and population density were assumed constant within any county, they were transformed into a i constant coefficient. Thus, visits then became a func- 7 I ! i | tion of distance a l o n e . ^6 | Aside from possessing the inherent difficulties ! I of the travel cost models, a demand curve or benefits j ; I computed from the travel cost saved at an alternative j ! site is by itself an unreliable indicator. The hypothesis j that a closer site saves people in that zone the cost of j traveling to a more distant zone and thus results in savings, does not consider that the attitude of a visitor I 1 to a site depends not only on location but also on qual- i ity. Further, it does not consider that the recreation tastes and behavior of consumers might be different. Recreation Demand Studies Using I the Direct Approach I j ) The direct approach is essentially a bidding pro- i I i ! cedure in which users are asked how their use of an area i ] will be affected as their costs of using the area in- | crease. The increments of cost are raised until the ^ Ibid. , pp. 627-629. Ill respondent indicates he would not use the area. The fundamental idea of the direct approach was suggested by Ciriacy-Wantrup in 1952. He contended that: J The marginal rate of substitution between market goods and extra-market goods may be obtained ob- j | jectively through the observation of either j j actual or hypothetical situations of choosing. j And, if the market good is money, the marginal j rates of substitution can be interpreted as an I individual demand schedule for the extra- j market good.77 : i To obtain a demand schedule for recreation sites, | i ! I ! he suggested direct interrogation of the users of the site j ; | : to ascertain "how much money they are willing to pay for j : i ! successive additional quantities of a collective extra- j ! j market good."78 Questioning of recreationists would be | by either personal interviews at the site, or by asking ! the recreationist to complete a questionnaire. Various ; methods were noted to remove bias from answers and ques- i ! tions and the results ". . . would correspond to a market | demand schedule."79 77S. V. Ciriacy-Wantrup, Resource Conservation, ; Economics and Policies (Berkeley: University of California | Press, 1952), p. 240. 78Ibid., p. 240 79Ibid., pp. 240-241. 112 | By way of example, Davis8® used the direct tech- | nique to derive a demand curve for hunting at a particular i j area. The purpose of his study was to determine the value i i of benefits from big game hunting in a private forest. j His procedure entailed: interviews with hunters in se- | I i I i lected camp grounds to determine their socioeconomic and j I participation characteristics? and, use of a bidding pro- ! cedure to ascertain how the respondent would react to j increased costs of visiting the area. The information j would, Davis felt, be valid if, ". . . consumers are act- j ! ing rationally on good information and that all benefits j from consumption are specific to the consumers."8^ Further, through the bidding procedure, demand j curves could be derived: I i j . . . by arraying the responses of willingness to pay per household unit and cumulating down ward. That is, the number of buyers who would pay as much or more than a given price is cal culated for each price . . . and the demand curve for an individual would thus show that | he uses the area by a constant amount below a certain price [his willingness to pay] and he j uses the area not at all above this p r i c e . 8 2 | 8®Robert K. Davis, "The Value of Big Game Hunting ! in a Private Forest," Transactions of the Twenty-Ninth ! North American Wildlife and Natural Resources Conference (1964) . 81Ibid., pp. 394-395. S2lbid., 395. 113 By pooling all interviews and using multiple re gression techniques, Davis observed that willingness to pay per visit could best be explained by the hunter's household income, length of visit and years of acquaint ance with the area visited. He then took samples of the hunter population, classified them according to these characteristics and computed the willingness to pay from a logarithmic estimating equation. The demand curve was then formulated by cumulating the distribution of visits at successively lower prices.®3 On advantage to the direct approach is that it approximates the market by allowing users (or consumers) to indicate their willingness to pay for a particular recreation activity or site. However, this technique has some basic, inherent problems. The most obvious is that of obtaining a rational, consistent measurement of price. A demand schedule thus derived "is a hypothetical rather than actual schedule since charges for the recrea tional service will not actually be made."84 Further, as pointed out by Crutchfield in an article concerning the valuation of a sport fishery, that "surveys of 'intention 83Ibid., p p . 3 9 7 - 4 0 0 . 84Lerner, op. cit., p. 6. to buy1 are subject to biases which frequently would lead ! to overestimates of demand. It is one thing to say that one's fishing would not be affected by a $10 license fee, and quite another if it becomes necessary to pay the $10."85 Despite this and other difficulties, he suggests "... these problems could be overcome in a properly de signed survey, carried out by highly competent enumer ators. "88 Conclusions The writers in the 1920s did little more than isolate problems of allocation, impact and institutions associated with outdoor recreation. Later works in the 1930s and early 1940s began to provide empiric content to their efforts but the use of economic theory was extremely limited. Since the late 1940s, the work has become in creasingly more sophisticated and useful to recreation planners and policy makers. With the demand estimation models of the late 1940s and extending into the late QC James A. Crutchfield, "Valuation of a Fishery Resource," Land Economics, Vol. XXXVIII, No. 2 (May 1962), pp. 153-154. 86ibid., p. 154. i 1950s, researchers widened their scope to include not | | only estimates of value, but also estimates of demand. | i Since the early 1960s, economists have shown an increas ing interest in outdoor recreation and the research has deepened to include many variables affecting recreation demand. Use of these approaches for forecasting state wide outdoor recreation demand, however, is extremely limited. | For example, the market benefit methods attempt to translate measures of recreational activity into dollar values of goods and services generated by recrea tionists, and use these values as proxy indicators of recreation benefits. These methods, however, offer no i analysis to generate a structural demand function, to I quantity demand, or to specify the dynamic properties of j 1 the demand function. i The "indirect approach" or "travel-cost" approach j assumes an inverse relationship between recreation demand and distance to a given recreation site and attempts to i assign an explicit dollar value to distance as a proxy | for price. The underlying logic claims that travel dis- | tance is the major determinant of recreation demand and ! its "price" or "cost" (in terms of time and/or expenses 116 | incurred) is thus the correct market price proxy (or in- j j dicator of utility) of the recreation "good." Demand j | schedules are then estimated on the basis of this i distance-use relation and a demand curve constructed. Demand projections are then made based on projections of population and distance. i I Although extended and refined from its original * formulation, the main idea remains that a demand schedule i ; | ; for recreation at any existing site can be constructed j and a subsequent analysis made to measure benefits. j Moreover, this technique still possesses many deficiencies J for statewide outdoor recreation planning. These diffi- j , j ; culties include: the sole dependence on travel distance ! as a causal variable; the difficulties in collecting ade- i quate "cost" data; the implied relevance of travel costs as the pertinent measure of benefits (utility); the neg- I lect (empirically) of integrating other relevant demand variables (e.g., socioeconomic factors such as income) j 7 i I into the analysis; incorrect use of past visitation (con- j sumption) to an existing site as a measure of potential | demand; and, ability only for predicting future visits to existing sites and not to projecting demand for new sites for which the public has no familiarity. As will be 117 shown later, however, the travel cost approach can be useful with certain modifications. The "direct" approach involves questionnaire techniques designed to elicit "price" information from participating recreationists based upon their willingness to pay for a particular recreation activity or site. This technique is also inadequate for statewide outdoor recreation planning due to the basic difficulty in ob taining a rational, consistent measurement of price from the interviews. The "socioeconomic methods" and other commonly used approaches to forecasting recreation demand and supply requirements are reviewed in the following chapter. CHAPTER IV j THE "SOCIOECONOMIC" AND OTHER METHODS USED I | IN PROJECTING OUTDOOR RECREATION DEMAND ! i Starting in the 1960s, the "socioeconomic" meth- j I i ods of demand projection gained prominence primarily as j : j I a result of the recreation studies undertaken by the | Outdoor Recreation Resources Review Commission. Many ! j j statewide outdoor recreation plans prepared since then : j have utilized the results of the ORRRC works as the basis j j • i I for their forecasts of recreation demand and supply re- j i t quirements. This chapter describes the socioeconomic as j well as other methods used to project demand and supply ! requirements. It also reviews two selected statewide i I j outdoor plans prepared using the socioeconomic approach. | Projecting Demand from Knowledge I of Socioeconomic Patterns The socioeconomic methods of projecting recrea tion demand involve the assumption that current relation- ! I ships between recreational participation and socioeconomic characteristics of the population can be applied to the expected future socioeconomic structure of society. A brief review of the most important studies undertaken to 118 119 date follows. ; ORRRC Projections of I Recreation Demand I i ! The most complete and in many ways the most so- I ! j phisticated approach to both measuring and forecasting ! | j recreation demand as well as determining the various | I ! ! factors affecting demand for various outdoor recreation i I | I activities, was that undertaken by the Outdoor Recreation ! Resources Review Commission.* Moving beyond earlier j i methods, the Outdoor Recreation Resources Review Commis- ! ; | sion (hereinafter referred to as ORRRC) took an aggregate ‘ and empirical approach to demand determination which was j j far advanced as compared to the earlier approaches dis- i ! cussed so far in this section. That is, the approach was i ! not one of considering past, present and future visita tion to or usage of specific sites, but rather, was re- ! lated to relations between recreation activity and vari ous socioeconomic characteristics of people, j The approach developed by the ORRRC stems from several factors. One is that the Act of Congress which ^Marion Clawson and Jack L. Knetsch, Economics of Outdoor Recreation, Resources for the Future (Baltimore: Johns Hopkins Press, 1966), p. 128. 120 created the ORRRC (Public Law 85-470) charged it with | the responsibility, among other things, to ". . . deter- J ! mine the amount, kind, quality and location of such 1 outdoor recreation resources and opportunities as will i | be required by the year 1976 and the year 2000. . . ."2 j i To make this determination, the ORRRC undertook the task I I ! of analyzing and forecasting the demand for outdoor rec- j | reation to these target years. i In reviewing the existing data that had been col lected regarding outdoor recreation and the various j methods used to project recreation demand, the ORRRC j recognized this work was not suited to its objectives and . that indirect methods would have to be employed. As j noted earlier, for example, most of the approaches were I | oriented toward attempts to provide economic justifies- I : tion for establishing priorities among projects involving recreation use of resources. Further, much of the work on recreation demand analysis had sought to develop I | standard methods for projecting and evaluating recreation i demand at particular sites. The ORRRC assignment, how- | I ever, required forecasts of outdoor recreation demand on Outdoor Recreation Resources Review Commission, Outdoor Recreation for America (Washington, D.C.: Gov ernment Printing Office, 1962), p. 192. 1 121 a much broader scale and not related to specific sites. The ORRRC, therefore, developed its approach to determine recreation demand for a variety of recreation activities which could be projected into the future for the nation and large census regions.3 The method used by the ORRRC consists of two main parts. First, two major studies were undertaken to determine the effect of underlying factors associated with participation in outdoor recreation: the survey con ducted by the Survey Research Center of the University of Michigan;^ and, the National Recreation S u r v e y . 5 Both studies were based upon nationwide samples of the population taken in the 1959-1961 period which provided considerable data as to the preferences and behavior of people throughout the United States with respect to out door recreation. These studies are reviewed in detail in 3 Outdoor Recreation Resources Review Commission, Prospective Demand for Outdoor Recreation, ORRRC Study Report No. 26 (Washington, D.C.: Government Printing Office, 1962), pp. 1-2. ^Eva Mueller and Gerald Gurin, Participation in Outdoor Recreation: Factors Affecting Demand Among Amer ican Adults. ORRRC Study Report 20 (Washington, DTcTT Government Printing Office, 1962). ^Outdoor Recreation Resources Review Commission, National Recreation Survey, ORRRC Study Report 19 (Washington, D.C.: Government Printing Office, 1962). 122 the following chapter of this report. Concurrent with these studies, additional studies were undertaken to develop projections to the years 1976 and 2000 of the population and socioeconomic factors anticipated as im- j | portantly associated with variations in rates of partici- j I pation. The second part of the approach was the synthe- j sis of these data to forecasts of recreation demand to ; the target years. Due to the somewhat restricted data in ! the Survey Research Center study, the National Recreation | Survey was used as the basic point of departure for the i ORRRC projections. | Methodology.— The National Recreation Survey re vealed a high degree of correlation between certain socio economic characteristics of the nation's population 12 years of age and older and participation rates in various outdoor activities. Through multiple regression analysis of cross section data, rather than time series data, the independent effect of five socioeconomic factors (family j income, education, occupation, place of residence and \ 60utdoor Recreation Resources Review Commission | Staff, National Planning Association, Bureau of Labor Statistics, U.S. Department of Labor and A. J. Goldenthal, Projections to the Years 1976 and 2000: Economic Growth, Population. Labor Force and Leisure and Transportation, ORRRC Study Report No. 23 (Washington, D.C.: Government Printing Office, 1962). 123 i age-sex) on participation in 17 different outdoor recrea- ! tion activities as established and participation rates (for 1960) in these activities were determined. This was done with more accuracy than in any previous study. Using the projected value of the various socio- I I economic factors to the years 1976 and 2000, as determined i ! in a separate study, the gross effects of these factors j on participation rates from 1960 to the target years were j j | then estimated. This was done by reweighting the 1960 rates according to the projected distributions of the j ; i population by each of the five socioeconomic factors. j i i As in common with cross section analyses, the basic as- | sumption was that the relationships between participation i rates and the socioeconomic factors observed in 1960 | would continue in the future. The gross effects were i | then reduced to a net basis using adjustments developed i through multi-variable analysis and the separate effects of each factor in explaining changes in per capita par- i ticipation rates were determined. For two additional i | factors, leisure and per capita opportunity to partici- j i pate, the net effects were estimated separately. It is interesting to note that the effect of dif ferences in per capita opportunity to participate, i.e., the availability of facilities, was incorporated into 124 the analysis, as: The type, quality and extent of facilities for outdoor recreation which are easily accessible to persons' homes very probably influence per sons' propensity to engage in outdoor recrea tion. Clearly, no person can engage in a par ticular activity if he has no opportunity to do so. 7 The measure of recreation opportunity was essen tially derived by obtaining "opportunity" ratings for 16 | i outdoor recreation activities in each of the 66 primary j sampling units (PSU's), across the country used in the j Survey Research Center survey. Opportunity was considered| to be a function of three elements: the quantity and i quality of physical resources; the accessibility of and j development of these resources; and, the relative degree of use made of them. The ratings were made by various experts in the field of outdoor recreation (e.g., the National Park Service, U.S. Forest Service, etc.) and a rating questionnaire and scale were utilized to evaluate and score the various areas in terms of these elements. This was done for each activity and an opportunity index developed for each PSU and region across the nation.® 7ORRRC Report 26, p. 39. ®Ibid., pp. 39-48. 125 The relation between average opportunity scores and per capita participation rates for each of the various activ ities were than plotted. It was then judgmentally assumed that changes in opportunity would result in movement of the nationwide average along this line in the direction of the best region or the region providing the most rec reation opportunities for a particular activity. In effect, the availability of recreation opportunities in various regions was normalized against the best region. The separate effects of all seven socioeconomic factors were then compounded to secure the composite par ticipation rates by activities. Because of intercorrela tions between the socioeconomic variables, such as income, education, occupation, etc., the net effect of all the variables in combination was found much more meaningful than the gross effect of each individual variable. The composite effect for future years was then multiplied by the 1960 observed rates to secure estimates per person for future years. These rates were then multiplied by the projected number of persons 12 years and over to ob tain the number or level of actual recreation activity. This was done for each of 17 outdoor recreation activi ties, and projections made to 1976 and to 2000, for the 126 United States and each of the four census regions sepa rately. However, due to the far greater degree of value judgments that were required to derive the opportunity indices, two sets of projections, both including and ex- j ! eluding the opportunity factor, were made.9 j It may be noted that a second nationwide survey j of outdoor recreation activities was undertaken in 1965.1® j The data for 1960 were gathered for the ORRRC while those j for 1965 were gathered for the Bureau of Outdoor Recrea- I i tion (BOR) for its use in developing a nationwide outdoor i i recreation plan. For the first time, a comparison of j national outdoor recreation participation at two points in time became possible. An analysis of the results of | these surveys will be made in the following chapter of | this report. | Unfortunately, however, the projections of future participation published by the BOR using both the 1960 and 1965 survey information are limited in scope and the I methodology used has not been presented. These projec- i j tions (available only for participation in 16 major sum- | mertime activities) are shown in Table IV-1. ®Ibid., pp. 14-22. 1®U.S. Department of the Interior, Bureau of Out door Recreation, The 1965 Survey of Outdoor Recreation Activities (Washington, D.C.: Government Printing Office, 1968) . TABLE IV-1 PARTICIPATION DAYS IN 16 MAJOR SUMMERTIME OUTDOOR RECREATION ACTIVITIES 1960-2000 1965 Bureau of Outdoor Recreation 1960 Outdoor Recreation Resources Review Commission Rank 1965 1980 2000 % Change 1965-2000 Rank 1960 1980 2000 % Change 1980-2000 Walking for Pleasure 1 1,030 1,539 2,581 151 3 566 922 1,569 177 Swimming 2 970 1,671 2,982 207 2 672 1,310 2,307 243 Driving for Pleasure Playing Outdoor 3 940 1,423 2,146 128 1 872 1,508 2,215 154 Games or Sports 4 929 1,594 2,940 216 4 474 915 1,666 252 Bicycling 5 467 617 860 84 8 228 314 452 98 Sightseeing 6 457 705 1,169 156 5 287 499 825 187 Picnicking 7 451 668 1,022 127 6 279 446 700 151 Fishing Attending Outdoor 8 322 422 574 78 7 260 364 521 100 Sports Events Boating (other than 9 246 352 535 117 9 172 272 416 142 canoeing or sailing)10 220 387 694 215 10 159 316 557 250 Nature Walks 11 117 173 274 134 11 98 167 263 168 Camping 12 97 173 328 238 12 60 127 235 292 Horseback Riding 13 77 111 179 132 13 55 88 143 160 Water Skiing 14 56 124 259 363 14 39 94 189 385 Hiking Attending Outdoor 15 50 89 159 218 15 34 72 125 268 Concerts, Plays Total 16 47 6,476 10 80 ,128 144 16,846 206 160 16 27 4,282 49 7,463 92 12,275 241 187 Source: U.S. Department of the Interior, Bureau of Outdoor Recreation, Outdoor Recreation Trends (Washington, D.C.: Government Printing Office, April, 1967), pp. 20-21. io - j 128 As may be noted in this table, the projections published by the ORRRC in 1962 for the years 1980 and 2000 are considerably lower than those published by the BOR in 1967. This is true for all categories of recrea tion activities except "Driving for Pleasure." In addi tion, participation in all 16 activities between 1960 ! \ and 1965 increased by 51 percent as compared to an in- j crease in population (12 years and older) of 8 percent. j I Critique.— One of the limitations of ORRRC and i BOR demand studies is that the data are aggregated for j S the nation as a whole or for broad census regions. More ! than that is needed at the state level in making deci sions regarding the provision of providing specific facilities in selected areas. In addition, the ORRRC and BOR limited their con sideration to the age group 12 years old and over. The rationale given is that this is the age group that exer cises discretion in deciding where it will go and in what recreation activities they will participate. However, with entire families engaging in outdoor recreation, serious errors in the magnitude of measured demand can result. This error could be on the order of underesti mating demand by 50 percent in the case of families with 129 two adults and two children under twelve.H Moreover, the ORRRC studies (as well as BOR studies) do not indicate the number and types and loca tions of facilities needed to satisfy these demands. Further, although the constraint of supply on demand was considered via the introduction of the "opportunity" fac tor, this consideration was weak at best. The opportunity factor did not include the cost of time, money or distance as affecting recreation demand. These items cannot be excluded until family incomes are unlimited or alternative choices for expenditures are drastically reduced. Even if one accepts the proposition that variables other than cost are relatively more important in determining recrea tion demand, it is necessary to know the degree of price elasticity. In addition, the participation rates themselves do not represent true demand quantities since they are generated from a hybrid of demand and supply influences. That is, participation rates are not strictly demand phe nomena but also represent, necessarily, available recrea- •^R. J. Daiute, "Methods for Determination of ; Demand for Outdoor Recreation," Land Economics, Vol. XLII, No. 3 (August 1966), p. 329. 130 tion supply opportunities.12 A final comment is that although the ORRRC studies have their limitations, it must be remembered that these works were pioneering efforts and were the first of their type undertaken on a large scale. In ad- | dition, virtually every statewide outdoor recreation plan I throughout the nation undertaken since the ORRRC projec- j tions and surveys were released have utilized the par- j ticipation rates derived by the ORRRC in developing recreation demand estimates. Reviews of two of the more prominent statewide ! outdoor recreation plans, i.e., for the states of Cali fornia and Arizona, are made below. California Recreation | and Parks Study | A valuable example of the application of the ! socioeconomic method of projecting recreation demand is the study undertaken in 1965 for the state of California, j A number of states preparing outdoor recreation plans i | | since that time have patterned their approaches to fore- 12 J. J. Seneca and C. J. Cicchetti, "User Re sponse in Outdoor Recreation: A Production Analysis," Journal of Leisure Research, Vol. I, No. 3 (Summer 1969), p. 238. 131 casting recreation demand and supply requirements after that developed in the California study. The California Department of Parks and Recreation, as part of its long range planning program to provide outdoor recreation facilities, commissioned the Stanford j t Research Institute to undertake a comprehensive statewide I i study of outdoor recreation.*3 The purpose of the study j was to estimate the potential demand for outdoor recrea- j i tion in California to the year 1980 and to "establish a j framework for estimating this potential d e m a n d . j The procedure used to estimate outdoor recreation i demand in this study was based on the data developed in the National Recreation Survey, which was conducted in 1960. Per capita participation rates for the western United States, by socioeconomic category, were derived from the national survey information by the Stanford Re search Institute. Coefficients of determination of each factor were computed (using cross products) and applied to the socioeconomic characteristics of the population Stanford Research Institute, California Recrea tion and Parks Study: An Element of the State Resources Development Program, prepared for the state of California Department of Parks and Recreation (South Pasadena: Stanford Research Institute, December 1965), Part I and Part II. 14ibid., Part I, p. 1. 132 (12 years of age and older) expected in 1970 and 1980. The resulting reweighted per capita participation figures for each of the 23 outdoor activities are then converted into total recreation demand (number of recreation occa sions) by multiplying them be the projected population. j This was done for the population in each of the eight | major metropolitan areas in California and for the state I as a whole.15 These calculations, however, gave only the i | annual demand for various recreation facilities. Through ; | a separate survey, the Stanford Research Institute deter- I ! mined the seasonal and daily peak demand patterns for the | 23 activities. By applying these relationships to the j projected 1970 and 1980 annual demand, estimates were made of peak demand for these years.16 Here, it should be noted that while peak demand figures are relevant to the planning of adequate recrea tion facilities, it is neither economical nor necessary to build to accommodate all the people who may want to j use the facilities on the few summer holidays or the ex- | treme peak weekends. This illustrates an important issue ! related to demand prediction — that of peak demand i ISrbid., Part I, pp. 31-41. 16Ibid., Part I, pp. 83-90. 133 versus year-round demand. Unless a recreational system is prepared to carry huge amounts of surplus capacity for much of the year, peak demand can never be accommodated. Consequently, once peak demand has been predicted, it becomes a policy decision as to what percentage of this j demand will be supplied. In the California study, it was j determined that the optimum capacity need equal only 1 j percent of the total summer demand for all activities j except camping, and 1-1/2 percent of the total summer de- j mand for camping.17 Although not indicated in the report, the relatively small portion of total summer demand to be accommodated may be due to the long season of good weather in the state that would tend to "spread out" or reduce the peak season demand in the summer relative to other seasons during the year. The California projections also attempted to incorporate the effects of travel time and distance con straints into an otherwise purely socioeconomic analysis. First, concentric travel-time zones were established from the center of population in each of eight metropol- i itan areas within the state. The travel-time zone bound aries were determined by the distance the average family Ibid. ? Part II, pp. 138-139. 134 was expected to travel for recreation to certain recrea tion areas (day use, weekend use and vacation use) within I specified time limits: i Zero-to-One Hour Travel Time Zone: For close in fractional day-use situations requiring a maxi mum of two hours round trip. ! One-to-Two Hour Travel Time Zone: Generally for full day-use situations requiring a maximum of four hours round trip. | Two-to-Four Hour Travel Time Zone: Generally for weekend or overnight situations requiring a max imum of eight hours round trip. Over Four Hours Travel Time Zone: Generally for vacation situations requiring over eight hours I round trip. This zone also includes out-of- state travel.18 Then, projected demand for each metropolitan area i j was apportioned over each time zone according to an esti mated fixed proportion of all activities expected to ! occur in each zone. By distributing demand over various | time zones, this approach has value for planning purposes. J It was found, for instance, that California residents ; engaged in about 41 percent of their total recreation ! activity within a one-hour travel time distance from | their homes. By including projected highway construction i plans into travel-time zones, the future geographic l®Ibid., Part II, pp. 112-113. 135 location of the needed recreation capacity could be identified. The facilities in each time zone for various rec- ! reational activities were then determined by dividing i ! j the total demand for that activity by the average or j j standard size of the group using the facility. By com paring the facilities requirements with the existing sup- | I i i - ply in each travel time zone, deficiencies were identi- ! ! fied.19 j i ! I While the approach utilized in this study is sig- j nificantly more valuable to recreation planning than any j j ; j of the others previously discussed, it still has the same ; limitations previously noted for the ORRRC studies dis cussed earlier. For example, demand is measured for the I population 12 years and over, not for the total popula tion, and is probably understated. Also, the demand | estimates, which were used to determine facility require ments for parks throughout California, were not adjusted to account for the proportion of outdoor recreational I activity that does not occur at a park. As found by the ; Survey Research Center studies, not all outdoor recrea tion occurs at a park and not all activities require ^ • 9Ibid. , Part II, pp. 139-140. 136 developed facilities for their enj oyment. 20 Consequently, the supply required to meet potential demand or, stated differently, potential demand for park facilities is 1 probably overstated. ! | Further, while distributing total demand over j i ! time zones helps identify geographic locations of demand, this was done by applying a fixed proportion for all | activities over each zone. This technique does not ac- | count for the effect of time and distance on the amount | | of participation in specific activities, which could vary j i j ! significantly and effect supply requirements. j I ; Moreover, the effect of existing supply on rec- j i reation behavior was ignored. Facility supply require ments to meet projected demand were determined using recreation standards to convert demand (in activity or | participation days) to a quantity of facilities (e.g., i number of picnic tables) and then comparing this amount | with the existing level of supply, the difference being i an undersupply or an oversupply. This technique does not | specify the optimum combinations or amounts of facilities, i or their proper location, to best satisfy demand, within ^Mueller and Gurin, op. cit. , p. 58. r 137 the framework of the attraction or influence of existing supply on recreation behavior. As such, it can lead to I j misallocation of facilities and is, therefore, inadequate for recreation planning. I i Arizona Outdoor ! i Recreation Plan i ! | Inasmuch as a case study will be made for Arizona j ! j i to provide data for formulating an approach to forecast- j i | ing outdoor recreation demand and supply requirements, | it is appropriate to review the recreation planning that ! i j | has been accomplished thus far for the state. There have j been two statewide outdoor recreation plans prepared for i Arizona, both of which have used the socioeconomic method for projecting recreation demand. | In 1965, a plan entitled, Initial Outdoor Recrea- | ; tion Plan — State of Arizona,21 Was prepared by the i i Arizona Outdoor Recreation Coordinating Commission j I (AORCC) . The AORCC was established under state legisla tion as the legally designated representative to prepare i j and keep up-to-date a comprehensive outdoor recreation 21 Arizona Outdoor Recreation Coordinating Com mission, Initial Outdoor Recreation Plan — State of Arizona (Phoenix, Arizona, September, 1965). 138 plan for the state.22 one of the primary objectives of j the plan was to determine and project (to the year 1980) ! the demand for outdoor recreation in the state and "cor- I I relate these needs and desires . . . with the present and | potential supply of outdoor recreation r e s o u r c e s ."23 j i j 1 I This initial plan was prepared with the technical assist- j ance of the BOR, and while adopted as the state's offi- j ! I cial outdoor recreation plan, was intended to serve as a ! preliminary document until a more comprehensive plan j i j i could be prepared. | | In 1967, a more detailed statewide outdoor recre- i | ation plan entitled, A Plan for Outdoor Recreation in I ; i I Arizona,24 was prepared by the AORCC with the assistance | of a private consulting firm. Among the objectives of j j the 1967 plan were the determination of: ] | . . . Present and projected demand for outdoor | recreation [to the year 1985]; present and po tential supply of outdoor recreation resources i . . .; and, present and projected needs for re sources, facilities and programs determined by relating demand to supply of outdoor recrea- j tion resources.25 22ibid. , pp. i-v. 23jbid. ? 4. 24Arizona Outdoor Recreation Coordinating Commis sion, A Plan for Outdoor Recreation in Arizona (Phoenix, Arizona, June, 1967). 25ibid., Section I, p. 1-1. 139 Methodology and Critique — The 1965 Plan In developing the methodology for estimating recreation de- | mand in the 1965 initial plan, it was recognized that j ! I changes in socioeconomic characteristics could produce j i changes in future demand and that the "resource-oriented" j • i \ \ i approach or use of visitation data was inappropriate. j | ! ! Thus: | : I I . . . A technique was sought which would consider j j all important factors in the computation of Ari- j zona's present and future demand. Since a pro jection of future demand should consider the ; I ability and desire of people to participate in | activities, the "public-oriented" approach was used in contrast to the "resource-oriented" | method. The "resource-oriented" approach is , merely a projection of existing use and does not consider unsatisfied demands or, in other words, | uses that would have existed or been more in tense if the proper facilities of the required I quality and quantity had been available. Since the supply of existing facilities was not in cluded in the evaluation of existing and future use, the demand estimates may or may not have ! any relationship to existing u s e . 26 The methodology to determine recreation demand i was to multiply the projected annual number of partici- i pation days per capita for 23 major outdoor recreation activities for the western census region as reported by i ! the ORRRC in 1962^7 by the total population estimated in ^ Initial Plan (1965) , p. 63. ^Outdoor Recreation for America. r 140 1965 and the year 1980. The result was the estimated recreation demand for each recreation activity expressed in terms of total number of participation days for these i ; time periods. Although total population was used in | these projections, rather than the number of persons 12 years of age and over as used by the ORRRC, no explana- | tion was given. In addition, while it was recognized that travel time from major centers of population effect i | participation in various types of recreation activities, : no account was taken of this factor due to "lack of in formation. ! Outdoor recreation "needs," defined as ", . . ; those lands, waters and facilities required to provide | adequate outdoor recreation opportunities to satisfy the | demands of the people,"29 were then determined. This was i done by first dividing the state into four recreation i planning regions that were delineated, based upon "... similarity of terrain, climate, recreational opportun- | ities, zones of influence and population."39 Population ! for each of these regions for the years 1965 and 1980 ^Initial Plan (1965) , p. 76. 29ibid. 30Ibid., p. 70. 141 were then estimated. However, "... due to unavail ability of information, resident needs could not be ! | estimated by type of recreation activity."31 Conse- | quently, certain recreation standards were used to con- ^ | vert population into the number of recreation acres | j required per 1,000 people. These "needs" were then sub tracted from the existing supply of recreation land to determine deficiencies or surpluses in recreation acreage i j | in each of the four planning regions in 1965 and by the j I j 1 year 1980. The various types of recreation facilities ! I required (e.g., picnic tables, etc.) were not determined, j ! Although the plan was prepared almost entirely I I by the BOR, the agency given the responsibility to pro- i | vide technical assistance to states in preparing compre- | | hensive outdoor recreation plans, it is clear that the j I methodology used for projecting outdoor demand and deter- i mining facility requirements was inadequate and weak at | | best. This is indicative of the "state of the art" that prevailed at that time, a situation which has since im proved, but only slightly. Future changes in the socio- j ! economic characteristics of Arizona residents, the effect of time and distance, the influence of existing supply 31ibid., p. 76. 142 and the proportion of participation expected to occur at parks were not adequately considered and the plan could I j hardly be considered a guide for comprehensive recreation i planning. | Methodology and Critique — The 1967 Plan.— The ! | i I methodology used to estimate and project recreation de- \ i | mand in the 1967 plan was considerably more comprehensive i than that used in the 1965 plan. Here, recreation demand was estimated for 25 outdoor recreation activities for ! ! i l the 1965-1985 period by 5-year increments. The point of j : I | departure for the demand estimation was the 1960 partic- ! i ipation rates derived by the ORRRC in the National Recrea tion Survey for the western census region. It was assumed that the annual per capita participation rates for Arizona would be the same as those for the western states. The logic for this assumption was that even though the per ! i capita income in Arizona was lower than the average for the western regions i | . . . Arizona's climate, and abundance and avail ability of natural recreation resources permitted a participation rate at least in line with that of the Western region. The income constraint, therefore, was considered to be offset and ob viated any reduction in participation r a t e s . 32 -^Plan for Outdoor Recreation (1967), Section 3, pp. 3-57. 143 The annual per capita participation rates for each recreation activity were then converted to total j j potential recreation demand in 1960 by multiplying these rates by total population. Here, it may be noted that j | total population was used (rather than only those persons j i ! | 12 years old and older) inasmuch as the ORRRC participa tion rates were considered too conservative. This point j i was also noted by the ORRRC in their determination of I participation rates, and would be particularly true in ' those geographical areas with mild winters that permitted j ! | year around recreation activities. Consequently, to ad- | | I | just for the understatement of annual participation, total potential demand for 1960 was computed using the entire Arizona population, regardless of a g e . 33 Projections of future participation rates were | then developed by applying weighted adjustments to the | 1960 rates based upon the expected changes in per capita I income for the Arizona population. These adjusted rates i | were then applied to the estimated population in 1965 and ! to the year 1985 to derive estimates of the total number of participation days for each recreation activity to the target year. These estimates were then reduced to 33ibid., Section 3, p. 3-59. account for the seasonality of demand, or the portion that is likely to occur during the peak season, and to | determine the portion of demand that should be accom modated. ! | | Here, it may be noted that only the demand for j i j Arizona residents was estimated, as a survey of out-of- ' state visitors passing through the state disclosed that most visitors stayed only a few days and visited certain j ! | scenic and other attractions, not usually frequented for outdoor recreation purposes by Arizona r e s i d e n t s . j The demand projections also attempted to incor- j i i | porate the effects of travel time and distance constraintsj into the demand analysis. Travel-time zones were estab lished from the centers of the state's two major cities, Phoenix and Tucson. Then, the projected peak season demand was apportioned over each travel-time zone based upon observed recreational travel patterns of Arizona recreationists. Recreation "needs" or the facilities required in each time zone were then determined by converting demand to acres of land and water and various types of facilities 3^Ibid., Section 3, pp. 3-71 to 3-94. 145 (e.g., picnic tables) using recreation standards. By comparing facilities requirements with the existing supply in each travel-time zone, deficiencies or surpluses were identified. I The 1967 plan attempted to use the concept of j potential demand and correct for the probable understate- j ment of demand that would have resulted if only population j 12 years of age and older were used. It also attempted ; i to consider the effects of future socioeconomic character-j isties on future participation. However, the use of per i capita income to adjust the participation rates is ques- i i tionable, given the substantial differences found by the I ORRRC in participation between various income brackets. Family income, by income bracket, would have been a more satisfactory measure to weight the participation rates. Also, there was no consideration given to the pro portion of participation that is park oriented even though the supply requirements were for park facilities through out the state. Hence, the demand for these facilities is | probably overstated. In addition, while the effect of travel time and distance on recreation demand was esti mated for specific activities and time zones, the effect of existing supply on recreation behavior was not deter mined and thus not used in estimating supply requirements 146 to meet this demand. Similar to the California Plan, facility requirements to meet potential demand were based | on recreation standards and compared with the existing ! level of supply to determine deficiencies. As a result, the optimum supply combinations and locations to meet i I projected demand were not determined. Therefore, this i I plan, while more comprehensive and complete as compared with the 1965 plan, is still deficient for recreation I planning. i j An Econometric Approach to Recreation Demand and Supply Analysis i The econometric analysis of the demand and supply j of outdoor recreation by Cicchetti, Seneca and D a v i d s o n 3 5 reports the results of one of the most important, major ! studies made in recent years in the field of outdoor rec- ! reation economics. The study was financed by the Bureau I I of Outdoor Recreation of the U.S. Department of the In- | terior, the Bureau of Economic Research at Rutgers Uni- i I versity, and the Center for Computer Information Services j at Rutgers. 35 C. J. Cicchetti, J. J. Seneca, and P. Davidson, The Demand and Supply of Outdoor Recreation; An Economet ric Analysis (New Brunswick, N.J.: Rutgers-The State University, Bureau of Economic Research, 1969). In developing their approach, the authors rec ognized that supply can constrain action of demand and supply variables. Using the results of the 1965 National Recreation Survey, they related participation in various i ! forms of outdoor recreation to a wide range of socio- I j economic variables as well as certain supply character istics. They then developed a model for forecasting rec- ; reation participation claimed to be useful for recreation planners at all levels of government. This was done using | theoretical economic constructs in a manner and on a j scale which is unique in the outdoor recreation litera ture. While much time and effort have been devoted to i i studies utilizing such data, each has fallen short of the analytical scope reported in the Cicchetti, et al., study. I | Due to the importance of this study, which is | amplified by financial assistance provided by the BOR, a ! comprehensive review of the methodology employed and a | critique of its validity are in order. | Methodology The underlying philosophy in the Cicchetti, et al., methodology is the authors1 that recreation planners 148 "... need policy variables which they can alter to change patterns of behavior if they deem this desir able."3^ While it would be difficult or impossible to alter participation by changing the socioeconomic char acteristics of people in an area, it would not be diffi cult to change the supply of facilities in an area. Thus, projections of future recreation activity based solely on socioeconomic (demand) factors, such as the ORRRC studies, permit only a limited range of possible policy alternatives. Projections based on both demand and supply factors would allow planners more freedom in developing future recreation projects; once supply fac tors are explicitly introduced, planners can then change the amount, type and mix of facilities to achieve either greater or lesser participation in all outdoor recreation or greater or lesser participation for specific activ ities. 37 Cicchetti, et al., begin their analysis with a theoretical demand and supply model of the outdoor recre ation market and within this framework, attempt to deal 3^Ibid., p. 43. 37Ibid., pp. 43-44. 149 with the zero price and identification problems previously noted in Chapter II of this report. They justify their use of a theoretical model in an empirical study on the grounds that: j . . . First, all a priori analysis and rationale i should come together and be expressed in a con- j crete form for the purpose of testing the logic of the many individual lines of reasoning; and second, the theoretical model should form the basis for empirical estimation and hypothesis j testing.38 i The model itself is based on typical demand and ■ supply equations as well as a market clearing equation. I The result is a theoretical equation to estimate recrea- j tion demand with supply implicit. A two-step empirical j demand model is then developed from the theoretical de mand equation. Cicchetti, et al., use a two-step ap proach because: Those variables or factors that determine whether a person participates in a partic ular activity or not are not the necessarily the same variables or factors which would determine the number of days that a par ticular activity might participate.39 The first step of the model yields the proba bility of a given individual participating in a given outdoor activity. The second step yields the average 38xbid., p. 59. 39Ibid.? p. 79. 150 annual days of participation in a given outdoor activity per participant in that activity. Cecchetti, et al., | define 78 variables which they attempt to "fit" for each I I outdoor recreation activity. These variables include i | measures of various socioeconomic characteristics, dis tance, supply and complementary goods. j Cecchetti, et al., complete their analysis with I an application of the empirical model. The outputs are i | total recreation days on a nationwide basis for each out- ! door recreation activity for the years 1980 and 2000. j i i j This model relies on population data for persons 12 years j ! t I j ! of age and older, by age and sex, organized into cells for analytical purposes, as well as projections of future I supplies of recreation land and water resources. The salient features of these models are de scribed below. j Theoretical Model j -... . . . . ■ The theoretical model of the outdoor recreation market developed by Cicchetti, et al., is a "basic dis aggregate demand and supply model for individual i in j time period t. "4° It is based on typical demand and 40ibid., p. 60. 151 I supply equations which are dealt with separately to help overcome the identification problem, j The hypothesized theoretical demand equation for recreation is stated as: ! | = ai + a2 (distance t_^) + a3 (quality and i ; quantity t-1) + a4 (socioeconomic t) + u* [1] i ! I where is the quantity demanded by the i*^1 indi- i ^ ; vidual in time period t; distance is a proxy for price lagged one period and is the physical distance be tween the it*1 individual' s point of origin and the | recreation supply; quality and quantity form a ! composite of the level of crowding and relative avail- i i i ability of recreational resources in the time period im- j I mediately preceding t; and socioeconomic ^ are characteristics of individual i (e.g., age, race, sex, income, etc.) in time period t. | In this equation, distance of the individual from I the recreation site, quality of the facilities and quan- ! tity or relative availability of the facilities are considered the costs to the recreationist and used as proxies for the zero price. They are lagged one period 152 to account for the "learning-by-doing" phenomenon, the development of habit patterns in participation, and the demand which might occur for a new site or a new activity. ! By lagging these factors, past knowledge of the value | (or price) of participation to the recreationist can be j | imputed. In addition, the implicit effect of supply in I a previous period on demand in the current period is con- i i | sidered. The socioeconomic factors are not lagged as j ; | | these variables are considered constraints on participa- j | ! I tion only in the current period.4^ | i i : I The relevant theoretical supply curve is given | I i j as: j I I j j = bjL + b2 distance t + b3 (quality t and | | quantity t) + 2u£ [2] i where is the quantity supplied to the i ^ indi- ; vidual in time period t; the independent variables are defined the same as in equation [1] except that they are | | the values found in time period t rather than t-1. I Again, by including distance as well as quality and quan tity of the facilities, Cicchetti, et al., introduce an 41Ibid., pp. 60-63. 153 implicit cost for supplying facilities which typically are provided free by public bodies. This is also con sistent with the Cicchetti, et al., argument that supply in a previous period affects demand in the current time period.^2 Price is, therefore, implicit in both equations and is defined as: P^,S = f (distance t, quality t and quantity t) [3] The market will clear if demand equals supply. The mar ket clearing equation is defined as: ^■pj? = iP? [4] t t To arrive at the reduced form of the model, Cicchetti, et al., first substitute the price relation ship (equation 3) into equations [1] and [2], the demand and supply equations, respectively. They then substitute the resulting two equations into the market clearing equation [4], By collecting terms and lagging the re sulting equation by one period, they then derive the reduced form of the demand equation for participation in the present period. This is defined as: 42Ibid.. pp. 63-65. 154 D*- = (Ci - Co + ^2^1) + 2A distance ■ . t 1 d2 d2 d2 b2 t-1 + £2 (quality ^_l ar*d quantity ^ d2 b3 + a^ (socioeconomic ■ ( - ) + [5] In equation [5], supply is implicit. This equation is the basis for the Cicchetti, et al., formulation of the empirical model to estimate future recreation participa tion.^ Empirical Model The Cicchetti, et al., empirical model is a two- step or twin linear probability model to estimate the total number of recreation days in a particular recrea tion activity in the nation as a whole as well as the four census regions of the country. In building the model, they start with the following identity: Total Recreation Days = Number of Recreationists x Activity Days Per Recreationist They use a two-step approach based on the ration ale that factors (e.g., socioeconomic characteristics) ^Ibid. , pp. 66-68. 155 influencing whether an individual participates (number of recreationists) in an outdoor recreation activity are j different from the factors (e.g., availability of facil- j | ities) that affect the length of time (number of days) he j recreates. Thus, the first step relates to determining | i ! 1 i | the causal factors explaining participation in each ac- j tivity? the second step relates to determining, for those | i who participate, the factors that effect the amount of . . . 44 I ! participation. I | In the twin linear probability approach, the j : I first function estimates the probability that individual j i ; "i" in time "t" will participate in an outdoor recreation ! activity? the second function estimates the expected I j value of the level of participation, given that the prob- j ability of participation by individual "i" in time period "t" is greater than 0. The technique used to fit j the model was classical least squares. The data used ! for demand factors were taken from the 1965 Survey of Outdoor Recreation? the supply data were derived from the | Bureau of Outdoor Recreation inventory of recreation i f resources. 44Ibid., p. 76. 156 To make the data more pliable, Cicchetti, et al., grouped the 24 recreation activities covered by the 1965 | Survey into seven groups. Classification of an activity i | into a group was based on: whether public policy could i be used to affect these activities; the type of policy I | j ; area (e.g., water, land, preservation of natural environ- j ments, etc.); and, whether the recreational activity is | passive or active. j Cicchetti, et al., then develop a step one and j ; two equation for all 24 outdoor activities. In fitting j ! i the 48 equations, they draw on 78 independent variables | i ; I significant in any given equation. All equations are | | listed in the text where, additionally, they interpret ! the results of the step one and step two swimming equa- | i tion. A synopsis of the swimming equation interpretation j j follows. i The Swimming Conditional Probability Equation — Step 1.— For all the activities, the first step equations have several characteristics in common: (a) they repre- ( i sent national behavior; and (b) the supply factors are deflated by population (acres per capita) to measure the relative availability of resources between different areas. 157 The step one conditional probability equation for swimming is: Probability [swimming (summer only)] = .5884 - .2157 age of sample person/10 + .0114 (age of sam- (discrete) pie p e r s o n / 1 0 ) ^ + . 1 9 7 4 race + . 0 3 4 9 sex - - . 3 2 3 size (discrete) (dummy) (dummy) of town of residence + .0545 education of sample (dummy) (dummy) person + .1684 income of family/10^- - .0329 (income (discrete) (dis- of f a m i l y / 1 0 4 ) 2 + . 0 2 6 7 home ownership - . 0 0 8 8 family crete) (dummy) (discrete) size + .0315 number of children in family aged 6-11 (discrete) - .0790 region of residence + .0329 occupation of (dummy) (dummy) person + .0730 sample person self-employed or not (dummy) + .0427 occupation of head of household + .0119 (dummy) population in the primary sampling unit/106 (discrete) - .0189 population in the home state/106 - .0000859 (discrete) distance to major body of water + .0962 acres of BOR (discrete) Classification I in primary sampling unit of residence (discrete) 158 per capita lagged one year + .1445 money spent at ancillary facilities in each state/10^ + .1673 number | (discrete) j | of recreation facilities in each state/10^ _ .0275 i (discrete) ! night visits in the most recent year per capita i (discrete) j + .0214 density of state of residence/10^ + .0274 median income of state of residence/10^. i | The predicted value of the probability of swimming | is determined by substituting values of the independent variables for particular individuals into this equation. The result is the probability that individual i will engage in swimming given the specific values for the in- i j A K dependent variables that characterize individual i. The coefficients of the independent variables, along with their respective signs, should be thought of as the percentage change (in deicimal form) associated I with a one unit change in the independent variable. For j instance, -.0088 family size is the eleventh term in the i equation. This means that for every additional family member, there is a .88 percent decline in the probability that a member of that family will go swimming. The next term in the equation, .0315 number of children age 6-11 45 Cicchetti, et al., op. cit., p. 101. 159 in the family, indicates that for each child between the ages of 6 and 11 in the family, the probability of swim- ! ming is 3.15 percent higher. Cicchetti, et al., note | that: Together these variables indicate that families with younger school children swim more than j either those families with infants and pre school children or those families with infants I and pre-school children or those families with teenagers and adults living at home. In addi- ; tion, ceteris paribus, larger families swim | less than smaller families.46 Beneath each term of the equation is written a word, either discrete or dummy. The former is quantita- | tive and refers to an actual number for or amount of the ; independent variable (e.g., the age of the sample person would refer to a numerical age). The latter refers to a qualitative code and indicates a characteristic of the i independent variable (e.g., the race variable would be coded 1 for white and 0 for nonwhite). The quantitative I | variables are all used with deflators (i.e., division by j a factor of 10, 10^, etc., or by population) so that the j discrete variables do not overpower the dummy variables i | by sheer order of magnitude. Several points in the Cicchetti, et al., step one swimming equation require special mention. One is the 46Ibid.. p. 108. 160 wide range of socioeconomic variables they found statis tically significant. Of the 24 independent variables used in this equation, 17 are socioeconomic variables. Another point is that within the socioeconomic variables, age, race and family income were found to have the greatest effect, ceteris paribus, on the proba bility of a particular person engaging in swimming. For instance, a one year increase in age will produce a 2.15 percent decline in the probability of a person swimming. Also, Caucasians having all other characteristics in common with a nonwhite were found to have a 19.7 percent greater probability of swimming. This was not the sole difference in demand between white and nonwhite, however, as there may be other possible reasons for lower partici pation (e.g., the distribution of the supply of re sources) . In addition, the effect of family income is opposite that for age. That is, as income rises, the probability of swimming increases so that swimming is a normal good in the economic sense. However, the proba bility of swimming increases with income at a decreasing rate. Cicchetti, et al., estimate the income level at which swimming would become an inferior good to be $51,000. They note, however, that this is a dangerous 161 extrapolation since there is no breakdown of empirical ] annual family income data beyond $25,000.47 Of further interest in the Cicchetti, et al., i model is the inclusion of factors for complementary | goods. One such variable is the acreage of BOR Class I j I ' land (deflated and lagged one time period) in the Primary i Sampling Unit. Cicchetti, et al., argue that the type of i ! land in this classification is characterized by intensive j ; | J development and is used as a measure of the availability j of recreation land which is complementary to swimming. This land classification includes developments "such as road network, parking areas, bathing beaches and marinas, j artificial lakes, playfields, and sanitary and eating i I facilities."48 Cicchetti, et al., argue that the avail- 1 ability of these complementary goods will explain the i | demand for swimming and hence they are included in the market clearing supply and demand equation. Another point is that total overnight visits at ' public recreation areas are negatively correlated with the probability of swimming. This could indicate swimming 47ibid., pp. 102-106. 4®u.S. Department of the Interior, Bureau of Out door Recreation, Bureau of Outdoor Recreation Manual, Grants-in-Aid Series, Nationwide Plan (Washington, D. C.: Government Printing Office, rev. September, 1966), p. 12. 162 is either a single occasion outing or that swimmers stay at private lodgings in resort areas rather than at public facilities at the site when they do stay overnight, or that camping in public recreation areas is a substitute i for swimming. Cicchetti, et al., are not clear as to j j which of the three explanations is the most valid. None- j theless, they include the variable in the model. j A final point in the step one equation is the in- ! | elusion of aggregate socioeconomic variables. Cicchetti, j et al., found two such variables statistically signifi- j I cant. The first variable was the population density in the state of residence of the respondent; residents of \ I more densely populated states were found to have a higher probability of swimming than residents of less densely populated states. The second variable was the median income in the respondent's state of residence; residents in states with higher median incomes had a higher prob ability of swimming. The Swimming Day Equation — Step 2.— In all step two equations, Cicchetti, et al., derive a separate day visit equation for each recreation activity for each of the four regions of the United States. They use these regional equations to determine and subsequently elimi- 163 nate the presence of interaction between independent variables, as such interaction would result in different functional relationships between the dependent and inde pendent variables and thus different slopes (but the same intercept) for the linear regression lines for each ac tivity in each region. The problem of regional multi- collinearity (different intercepts for the linear regres sion lines but the same slope) was eliminated in the step j one equation by using regional dummary variables to ac- ] count for the regional effect.49 j i The step two day visit equation for the western I region is given as: j Days of swimming per swimmer (western region) = 8.7102 - 6.2216 age of sample person/10 + .5105 (age of sample pe r s o n / 1 0 )^ + 6.6889 race + 4.4635 income of family + 3.4425 marital status - .0195 distance to major body of water + 59.20 percent of families with annual incomes less than $3,000. Particularly interesting in the swimming equation is the authors' use of a deflator (acres per recreation ist) 50 on the supply variables. They feel this ratio is Cicchetti, et al., op. cit., pp. 116-119. ^®An acre per capita deflator is used in the equation for all other activities. 164 a measure of the quality (e.g., crowding factor) of the supply whereas acres, acreas per capita, or acres per swimmer per capita, are more representative of the quan tity of supply. Cicchetti, et al., believe "it is qual- j ity and not availability of supply which determines the j number of days of swimming per swimmer."51 That is, j quantity is more important in the decision to swim or | not to swim, but quality (crowding) is more important to j I how often a person swims once he has decided to swim. ; Of additional interest is that supply variables are not ! i represented in the day swimming equations for the western i and southern census regions and appear only once in the north central census region equation. This omission was not discussed by the authors. Cicchetti, et al., provide additional detailed explanations regarding the regional step two swimming equation in very much the same manner as the step one equation. The salient features, however, have been noted above and need not be discussed further. The intent here was to describe the approach used in this important work, with sufficient detail to enable an understanding of the Cicchetti, et al., methodology and conclusions. ^^•Cicchetti, et al. , op. cit. , p. 121. 165 The Applied Model Cicchetti, et al., complete their study with an applied model based upon the empirical model previously developed. The underlying assumption in the applied model is that its structure is invariable in the long run. That is, it is assumed that an individual in the year 2000 will participate in outdoor recreation in a manner similar to his counterpart who possessed the same socio economic characteristics in 1965. For example, a Cauca sian between the ages of 12 and 18 years who lives in the year 2000 is assumed to behave in the same way with re spect to his recreational activities as a Caucasian in the same age bracket, ceteris paribus, in 1965. A similar assumption is made for the changing sup ply of recreation resources in the future. For instance, if additional recreation resources are made available in an area in the year 2000, the effect on the population living in the area at that time would be the same as if that population were living in the same area in 1965 and i could use these additional resources. In the authors' j ! words: i This cross-section assumption involves regarding groups of individuals [now and in the future] with specific socioeconomic characteristics as 166 homogeneous with respect to their use of avail able recreation facilities. It assumes that the experience of recreationists in one area of given recreation facilities simulates what similar people in any other area would "use" if the recreational resources in the latter region were the same. Under this assumption, we can transfer experience to predict behavior of populations in equivalent zones. This as sumption is common to all simulated demand curves that transfer experience from one group of people to another group, or from one time period to another time period.52 Inputs to the model are: population projections by race and age for persons 12 years old and older (as developed by the Bureau of the Census); and, projections (under two different hypotheses) of the future supply of recreation resources by the six BOR land classes. Each of the population subgroups are isolated in a race-age "cell" for forecasting and simulation purposes. Although Cicchetti, et al., indicate the size of the cell is "population," the exact dimension is not clear. The reader can only assume that the PSU (Primary Sampling Unit) which served as the main data unit in the empirical model is the cell used in the applied model. Cicchetti, et al., also indicate that race and age were the only socioeconomic variables used in the cell classification, because projections of other key variables, such as 52Ibid., pp. 209-210. 167 family income, are not reported by the Bureau of the Census with the same degree of expertise and statistical accuracy. Instead, "the projected average income was I applied to each of the fourteen race-age groups."53 j i However, the basis for "projecting" average income, the | analytical technique for "applying" income, and the defi- | i nition of "average" income are not given. The bases for projections of the future supply of j i land and water resources (in acres) in the model are less j c l e a r . C i c c h e t t i , e t a l . , m ake t h e s e p r o j e c t i o n s u s i n g J I two alternate hypotheses: | Hypothesis A assumes a continuation of past trends with a nationwide average percentage increase, varied by Bureau of Outdoor Recre ation land class. . . . Hypothesis B is based on adjusting the relation of partici pation to acreage by section of the country to a nationwide average and relating future acreage to an assumed amount of participa tion by section of the country.54 While their methodology is obscure, supply projections are included in the model. The outputs of the applied model are projections of the number of recreation days to the years 1980 and 2000. These forecasts are broken down for each recrea- 53Ibid., p. 212. 54Ibid., p. 229. 168 tion activity included in each of the seven activity groups for a specified year; they are made for the nation as a whole, except for swimming which are for the four [ census regions. In these forecasts, it is interesting I i to note that one activity, driving for pleasure, shows a J growth in participation to the year 1980 but a decline ! | from 1980 to 2000. Cicchetti, et_al., explain this by | j stating: "The reasons for this exceptional behavior are i I the negative effects of a growing real income plus the j ! projected increasing population and density to the year j | 2000. 1,55 | ! | As noted previously, however, income projections j were not included as one of the separately defined or | classified variables; thus, it would be difficult at best to determine the separate effect of this variable. I I | Critique I I | The Cicchetti, et al., study is one of the most i important works undertaken to date in recreation eco nomics. There are, of course, problems in any such undertaking. The major criticisms that may be noted are discussed below. ^ Ibid. , p. 214. 169 1. Cicchetti, et al. , list 78 predetermined var iables that they fit in the empirical demand equation for any given activity. They then examine and attempt to | 1 justify the significance (as well as the sign) of each t j | of these 78 variables. Typically, however, only about j | one-third of these variables were found statistically ! | j significant. In the step one swimming equation, which \ \ \ they use for expository purposes, there were only 24 sig- : | nificant variables. The technique of fitting all 78 j \ \ variables into the equation before determining their sta- i | I tistical significance, however, leads to a greater number | of errors in the resulting equation than if a smaller | number of variables, whose statistical significance were i predetermined, were utilized. j For example, based upon the "t" values shown in | the step one equation (which range as low as 2.3), it is assumed they use an "alpha" error (the error of accepting | a false hypothesis) of no greater than .05 to determine i the significance of each variable. This means that five i times in one hundred a nonsignificant variable will appear significant by chance. The authors' indiscriminant use of 78 variables, however, makes this probability of error much greater because a low "alpha" error (which no greater 170 than .05 is) implies a high "beta" error (the error asso ciated with rejecting a true hypothesis) which maximizes | the probability of rejecting a significant variable. i I More properly, Cicchetti, et al., should have justified ; the variables for each activity before fitting them into j I ' 1 i the equation to reduce the total number of incorrectly ! ! j : significant variables. While the levels of the alpha and J beta errors would remain the same, the absolute number of j ! I errors would decrease because the number of cases was re- ! | ! duced. Their failure to do this accounts for the some- j | what cavalier reasons posited for particular signs of j i i individual variables and their difficulty in justifying j | the significance of certain variables (e.g., the signif- ! icance of the variable for distance to a body of water I j with the insignificance of all the water variables). I 2. While developing their own relationships be tween socioeconomic characteristics of the population and i participation in outdoor activities, Cicchetti, et al., failed to reconcile their results with those of the ORRRC analysis of the 1960 national recreation survey.^6 For | example, in the Cicchetti, et al., step one swimming 56ORRRC Report 26. 171 equation, age and race were found to have the greatest effects on the probability of an individual engaging in swimming. This finding is quite opposed to the results i j of the ORRRC which found that family income had the i j | greatest net effect on participation in swimming, followed! ! by leisure time and education; the effects of age and sex were almost identical with those of occupation and place | of residence, all of which were found to have a very low ! i I net effect on p a r t i c i p a t i o n .57 The monumental importance j l , j of the 1960 recreation survey alone would dictate a thor- j : | ough discussion of any departures from the ORRRC results. i i | The reader of the Cicchetti, et al., study is left con- | fused as to why similar data results in dissimilar con clusions. In addition, Cicchetti, et al., use median family | income as a variable when both the 1960 and 1965 recrea- i tion surveys indicate substantial differences in partici pation between various income brackets. 3. Cicchetti, et al., use the number of acres of Bureau of Outdoor Recreation Class I land as one of the significant variables in the step one swimming equa Ibid., p. 28. 172 tion. This variable is one of the three variables used to constitute the complementary goods factor and is de fined as "a measure of the availability of high intensity I | recreation land — land which is characterized by water- | oriented activities in and around the land a r e a . j ! Cicchetti, et al., claim this variable is a measure of the j ! availability of recreation land which complements swimming i ] and encourages participation in swimming. | : ! : However, according to the BOR definition of Class I I land and physical criteria for classifying such land, i ! j swimming areas may be included; it is not mandatory that | ; i j they be included in this land classification. In fact, | | swimming areas may be in any one of the six BOR Classes of ! recreation land, each of which, according to their loca- | tional criteria, are increasingly more remote from popula- j | tion centers (i.e., the higher the class number the more i I I remote from population).59 The authors’ imprecise use of I this variable leaves considerable doubt regarding their ! assertion that land complementary to swimming areas actu- j ally encourages participation in swimming. Determination I i of this issue is paramount to recreation planners. ^®Cicchetti, et al., op. cit., p. 114. ^ Bureau of Outdoor Recreation Manual, Part 630. 4.3.J, pp. 12-20. 173 4. Another problem that arises from the authors1 j imprecise use of the BOR Class I variable is that if the i ! | acreage does include swimming areas, there is a contra- | diction with a conclusion they draw earlier, i.e., that I in the case of swimming, distance to water and not the ! amount of water is significant in determining swimming ' i | participation.An improved technique would have been to ’ i more precisely define the variables and use an initial ; i i screening procedure as noted above. j ! | 5. Cicchetti, et al., claim to have made a sig- ! i | : nificant departure from previous recreation demand studies ; ; by attempting to include supply in their model. As noted j i | in the previous discussion of the ORRRC studies, however, ! supply was included in one of the ORRRC's alternate series I I of demand projections. The ORRRC included supply as an | "opportunity" factor in a much more meaningful manner | inasmuch as this factor encompassed quantity and quality ; of resources, accessibility and development of resources, | and the relative degree of use made of them. The Cicchet- | ti, et al., claim is, therefore, not valid. i | In addition, the authors' indiscriminant use of a large number of variables has detracted from proper con- ^^Cicchetti, et al., op. cit., pp. 113-114. 174 sideration of supply factors. For example, even by their own interpretation, the step one swimming equation does j | not include a water supply variable because the supply of ; water was not found statistically significant. Although I j j they include BOR Class I land as a complementary goods j variable, use of this variable is, as noted previously, ; imprecise at best. j 6. Another criticism is that the authors' choice | of units of measurement for the supply factors has made i the real value of the supply analysis highly dubious. ; By using the BOR classification system to identify and ; measure supply, interpretation of the supply variables is | made extremely difficult. For instance, the primary cri- I | teria for the BOR recreation land classification system I | relates to location (i.e., proximity to population) and intensity of development (i.e., degree of facility devel- : opment) , not to size of area or recreation activities that i ! may be pursued. The nature of outdoor recreation is such | that most activities can be pursued at any type of area | irrespective of classification. An example of the fallacy j in using this classification system is the authors' find ing that BOR Class I land (a high intensity recreation area, highly developed and within or near major centers 175 of population) was not a significant variable for bicy- | cling whereas BOR Classes IV and V (natural areas and I j primitive areas, respectively, both of which have limited i i or no facility development and are located in remote | i J areas) were both found statistically significant (and with ! I ! | opposite signs). Gross measures of land are seldom good ! | measures of supply as a very small percentage of this j j j j acreage has an attraction, except for a few activities j I such as wilderness camping and wilderness hiking, which j | have extremely low rates of participation. j j Thus, a more narrowly defined concept of supply | i ! i should have been used. Measures such as the actual number j i of facilities or actual number of developed acres used would have been more meaningful if the Cicchetti, et al., model was to truly operate as a functional tool for rec reation planners. | i | 7. Cicchetti, et al., state they used distance | to a "major" body of water from a recreationist's home county as a proxy for price in the step one swimming equa- j t i o n . However, in their listing of variables, the dis tance variable (DIS) is indicated as "distance from ocean 61Ibid., p. 113. 176 or Great Lakes."^2 These are indeed "major" water bodies, and if correctly interpreted, distance to them would have little meaning as a proxy for price for recreationists not living in close proximity, i In addition, this same distance variable is used J ' I j for all recreation activities and in several instances un- i likely results were obtained. For example, distance was ! statistically significant in the step two equations for j such non-water oriented activities as Attending Outdoor j I Sports, Bicycling and Playing Outdoor Games and Sports. ! Further, even by attempting to use distance as a I proxy for price, Cicchetti, et al., failed to include ! other measures of "price" such as time, which they appar ently feel are irrelevant to user demand and should be ignored by public decision makers. This is not the case and will not be until time budgets are unlimited. I 8. In both the theoretical model and the empirical I | model, Cicchetti, et al., use a measure of "crowding" as | i | a surrogate for the quality of supply. They argue that i I | visitation and the amount of visitation to any given rec- I reation area is partially a function of one's own past experiences at that or similar recreation areas as well as 62Ibid., p. 172. 177 other experiences by friends or associates. Thus, next to distance, the degree of crowding experienced in use j I is the most important factor affecting utilization of that i I facility, and therefore represents the quality of supply. However, as Clawson and Knetsch note, quality of supply is related to a more subjective measure dealing j with the attributes, physical condition or design of the ' ; recreation area and its facilities.^3 The "crowding" j | factor more properly refers to quantity of supply rather | i ! than the quality, as overcrowding is directly related to | i I j the number of recreationists and the amount of recreation j i opportunities at the site. The measure Cicchetti, et al., j I use for quality of supply is therefore inadequate. ! 9. Cicchetti, et al., use population projections, by age and race, as inputs in their applied model to pro- | ject the number of recreationists. Family income, a key i | variable as found by the ORRRC, was not included as a sep- | arate variable as ". . .it is not a variable upon which | the Bureau of the Census bases population projections with j the same degree of expertise and statistical reliability such as age, race or sex. . . . 1,04 Granted, the Bureau 63ciawson and Knetsch, op. cit., pp. 164-168. 64 Cicchetti, et al., op. cit., p. 211. 178 of the Census does not make income projections? however, | j the claimed accuracy for the Bureau's population, age and I race projections is misleading. This is evidenced by the i : fact that the Bureau publishes four alternate series of such projections based upon varying assumptions and the ] I i choice of which series to use is at the discretion of the : I i demographer. I j In addition, while Cicchetti, et al., apparently j : 1 ! made some type of adjustment to race and age to account I | for the income effect on participation, the technique used | 1 i : is not specified. j ! i i 10. Another criticism is that the limitations of the analysis are not always spelled out. Cicchetti, et al., give lengthy rationales as to the effects of dif ferent variables on various recreation activities, but | have little to say regarding the low degree of explanation i associated with individual equations. For instance, the < coefficients of determination (r2) for the 24 step one I equations for various recreation activities are extremely i j low, ranging from .010 for Wildlife and Bird Photography I and Driving for Pleasure, to .33 for Swimming. Of these, 19 had r2 values of under .16? three were between .16 and .25? and two were over .25, one of which ranged as high 179 as .33. The values of the 23 step two equations ranged from .023 for Horseback Riding to a high of .25 for Bicy- | | cling. Of these equations, 17 had r2 values of under .16; | five were between .16 and .25; and, only one equation j j reached .25.^5 While Cicchetti, et al., admit to the sta- ; I ! ! tistical inefficiency of their classical least squares j j j estimator, and present a variety of reasons why they chose I ; i ! i | this statistical method over others in formulating his i j model, the reader is left to wonder why it was used at all I j | if it was known at the outset to be statistically inade- j quate. I ! : 11. Cicchetti, et al., claim their applied model, which parallels their empirical model, is a useful tool for recreation planning at various governmental levels. Further, that the model can be adjusted to bring the output from a national or regional scale to a state level. i i Insofar as being a useful tool for statewide rec- | reation planning, the model is weak and incomplete. The | output of the model, for example, is in terms of days of i i recreation activity for various recreational pursuits. i The number, types and locations of facilities needed to supply the recreation demands of the population are not ^ Ibid. . pp. 100-161. 180 specified. In addition, it is doubtful that even with considerable adjustments to bring the model to a state I scale that the output will be location specific within j ! a state. The model, therefore, not only fails to specify j I ; j facilities' requirements but also fails to specify proper : I ! j locations to best satisfy these demands. I i I 12. A final criticism is that the potential de- i ] ' mand of future recreationists has not been measured. The j I 1 ! authors use a reduced form market clearing equation which ! | specifically includes supply to project demand. Further, j ' all demand variables are held constant while supply is | manipulated to determine the effect of such supply changes on visitation or use. This methodology assumes supply creates its own demand and that the effect on visitation created by a given increase in supply is constant regard- j less of the magnitude or location of such an increase. I Consequently, the potential demand of recreationists based I 7 j | upon willingness to pay is constrained by the availability i | of supply and is not being measured. The result is yet ! ! another aspect of the identification problem. t I In addition, the Cicchetti, et al., model is geared toward controlling demand, for as the authors note, ". . . planners need policy variables which they can alter 181 to change patterns of behavior. . . ."66 pr0m the view point of public policy, this type of model permits the ! i recreation planner or policy maker to manipulate the ! "market" without allowing the "market" to express demand I in the true economic sense. Consequently, efficient re- | source allocation would have no firm basis since decision I making regarding the type, amount and mix of recreation facilities to be provided by a recreation investment pro- ! gram is left entirely to the value judgment of the deci- | sion maker. This could lead to inefficient allocation I of resources. ■ Moreover, this type of model is contrary to the spirit of the national legislation which created the BOR and more surprisingly, it is contrary to the requirements for such plans set forth by the BOR, the federal agency | which funded the Cicchetti, et al. , study. The BOR re quirements point toward first determining the future out- : door recreation demands of the public and then determining ■ the supply of facilities required to meet these demands.^7 i Thus, the Cicchetti, et al., model does not serve as an I | effective planning tool for recreation planning within the ^ Ibid. , p. 43. ^ Bureau of Outdoor Recreation Manual, Part 630, Chapter 4, p. 5. 182 context of stated national policy regarding the provision of outdoor recreation opportunities. 1 j j User Response Model I j A recent article by Seneca and Cicchetti presented I an interesting variation from previous recreation demand ! I ! t ! studies by suggesting a production function approach to j i | ; the analysis of user response for outdoor recreation. The j j purpose of their study was ". . .to explore the neglected j j ! j but important aspect of user response to recreation supply j I conditions."68 j The rationale for their approach is based on: the i identification problem; the overemphasis in previous rec reation research on estimating recreation demand; and, the relationship of visits to the supply of recreational I ! opportunities. The authors assert that an identification j | problem exists in the usual recreation demand analyses I ! because supply and demand are not considered separately, ! with the result that both factors are incorrectly included I in the category of demand. To overcome this problem, they introduce a user response function which they call a pro duction function for recreation sites. They claim a rec- 68seneca and Cicchetti, op. cit., p. 239. reation site can be viewed as a production process whereby recreation visits as the output of the site have an em pirical relation to the physical size and characteristics ! of the recreation area. Since recreation day visits are { not storable, the inputs of production (i.e., land, water, j I parking spaces, etc.) can be used to determine the tech- | ' nological function that explains the output of the facility! ’ i | (recreation days). The number of visitors at any given j ; park is the output of the park; any excess capacity is ' lost due to the non-storability of a recreation day. The j authors thus claim that the amount consumed is identical j ; to the total output of the p a r k . 69 They also assert that ! t j this approach will be useful to recreation planners as it will allow them to vary the mix of facilities to produce a given (or targeted) level of recreation user d a y s . 70 The user response function is defined as V = f (F,G) where; V is visitation or the amount of recreation I j visits; F is the extent of the physical facilities; and, i G is the physical area that exists at a recreation site. The user response function, then, is a quantitative rela- ! tionship between recreation use and the physical charac- 70lbid., p. 244. 184 teristies of the site. The data used to fit an equation to the user response function were based on the 1964 Bu- | reau of Outdoor Recreation (BOR) inventory survey of ! forest-oriented outdoor recreation areas in six eastern j ! i ! states. Included in the survey were data such as the num- i I ! i her of recreation visits at each site, the land area, j water area, number of parking places, the number of camp | sites, BOR water classifications, access mileage and | amount of entrance fee for each recreation site. j | ' The estimated equation that resulted was: log Vj = .1413 log Lj + .2681 log Wj | ' + 6562 log Pj + .7856 I SRj + 1.0764 Fj + 4.7910 C where Vj = total recreation visits at site j Lj = land acres j Wj = water acres j | Pj = parking places j i ! SRj = a binary variable coded: | | 0 - if swimming is not an activity at site j 1 - if swimming is an activity at site j Fj = a binary variable coded: 0 - if no fee is charged at site j 1 - if any fee is charged at site j C = a constant 185 The equation had a coefficient of determination (r2) of .738. The authors computed the order of homoge neity to be 1.035 Which indicated that there will be a constant return to scale; that is, at any given park, a 1 percent increase in the level of land, water and parking spaces will produce an approximately 1 percent increase in visitation at that park. Therefore, visits can be ex pected to increase proportionately with given percentage increments of land, water and parking spaces. Seneca and Cicchetti then use the relationships developed in the user response equation to develop a rec reation isoquant to determine optimum combinations of land and acres of water for the amount of visitation. That is, by holding visitation constant in the user response equa tion and solving for various levels of the land and water variables, a recreation isoquant of that fixed level of visitation is formed. They contend that this isoquant analysis will permit park planners to vary the mix of park inputs (as in the production theory of the firm) and thus become a useful policy tool. Seneca and Cicchetti have presented a creative extension of production economics into recreation eco nomics. However, their arguments fall short in several areas. For example, the "identification problem" which they attacked and attempted to overcome with their user i ] response function was not solved because the structural : I | demand equation was not identified. Market response to j | demand is not distinguished since the authors used visita- j ! I ; tion (consumption) and not demand in their model. Their ! i | i user response could just as easily be described as part i i I ; of a generalized demand function and is, therefore, just j | another aspect of the identification problem.71 | ! The return to scale observation made by the au- j j I i thors is also questionable. They claim that any increase ■ i ; | in the supply will necessarily result in a corresponding j [ ! | increase in the use of the facilities (in an approximately 1 to 1 ratio) regardless of where the new facilities are i located. However, visitation is also linked to demand i ■ ! | factors. Although the authors mention that some of the | costs associated with the participation in outdoor recrea- I tion, such as travel costs, do vary among individuals, i they do not have any factor in their equation to account ! for the fact that the location of new facilities will in- j j | fluence the amount of new visitation that will be gen- i erated. H-D. w. Fischer and J. M. Gates, "A Comment on •User Response in Outdoor Recreation: A Production Anal ysis,1" Journal of Leisure Research, Vol. II, No. 2 (Spring 1970), p. 135. 187 Moreover, output in the user response function is defined as visitation or use. This is equivalent to de fining output of the firm as sales and not production. i The production function does not forecast sales per unit ! i of time; it identifies an upper found on sales in the ab- j i sences of inventory accumulation. More appropriately, j i 1 the production function for a recreation site yields ca- j pacity, or the recreation opportunities possible, not the j i actual use. That is, the park or recreation site produces j opportunities for recreation each day, while use is some ! I I subset of these opportunities and is governed by factors I external to park production inputs. On any given day, for instance, the opportunity for recreation will exist at a given park in that the production inputs (land, water, campgrounds, etc.) have created opportunities at that park. However, visitation may be low or even zero if, for instance, there is inclement weather in the area on that day. Hence, it could not be said the park is not producing; yet that is the conclusion that must be drawn from the user response function. In other words, the au- thors1 assumption that the total amount consumed is iden tical to the total park output is incorrect.72 ^L. e. Gosse and R. J. Kalter, "User Response in Outdoor Recreation: A Comment," Journal of Leisure Re search, Vol. II, No. 2 (Apring 1970), p. 131. 188 Another issue is that Seneca and Cicchetti incor rectly calculated the order of homogeneity of their equa tion as 1.035; the actual figure, however, is 1.0656. This results in a slightly higher return to scale which further accentuates the above criticism. A final comment is that the user response model, which is similar to the previous model developed by Seneca and Cicchetti together with Davidson, 73 an<j discussed in the preceding section of this chapter, is geared toward controlling demand and not allowing the "market" to ex press its demand for outdoor recreation. This could, as indicated earlier, lead to inefficient allocation of resources. Other Methods of Demand Projection To complete the review of the various methods for projecting outdoor recreation demand, it is appropriate to note the methods that are typically used by recreation planners. These include the Visits Per Acre, Recreation Standards and Time-Income-Mobility approaches. 70 Cicchetti9 et al., op. cit. 189 Visits Per Acre Due to the complexities in measuring recreation demand, one of the most commonly used methods for esti mating future use of recreation areas is the projection of the number of visits to the areas in terms of visits i per acre. This technique attempts to use visits as a | measure of demand and translate this usage into acres j required. It has been used to project requirements for ! additional land acquisition and investments in facilities i to permit more intensive use particularly in the national j : parks, national forests and state parks. The procedure is to project historic data concern- | j ing per capita visits to existing recreation areas. The j projections are usually based upon past trends in visita- i | tion and then related to the number of acres in the par- | I ticular recreation area to determine the number of visits i | per acre. What might be termed the carrying or holding ! capacities of the recreation area is then estimated and j | compared with the projected visitation per acre to deter mine the amount of additional acreage needed. H. Landsberg, L. L. Fischman, and J. L. Fisher, Resources in America's Future: Patterns of Re quirements and Availabilities, 1960-2000, Resources for the Future, Inc. (Baltimore: The Johns Hopkins Press, 1963), pp. 223-230. 190 The use of "visits per acre" as a technique for projecting demand for outdoor recreation has the advantage i j of simplicity of application. However, the use of visita- : tion data, as indicated previously, has a number of seri- I i | ous limitations for outdoor recreation planning. One j i I ! problem, for example, is that it is limited to estimating i ' I future use at existing areas with no consideration given j to planned or newly developed facilities. Another is that ! ! ! it relates only to individual or specific areas, and i visits per acre vary greatly from park to park, especially j ; j those with wilderness areas.75 I j | ! Further, there is no consideration given to chang- ! ! i i ing population distributions, or changing consumer tastes i i j or other socioeconomic factors that are operative in af- | fecting recreation demand. The mere relation of visita- ; tion to the growth rate in population, for example, with no consideration given to the multiplicative or restraining effects of other factors, is far from satisfactory. In j 7 fact, the basic methodology of projecting past growth trends in visitation has been shown to yield very unlikely j | results.Moreover, the use of visitation data has, as i I i _______________ _ i ! ^Clawson and Knetsch, op. cit. , p. 120. ^Landsberg, et al. , op. cit. , pp. 224-225. 191 indicated previously, complicated the identification prob lem in recreation demand estimation. Thus, while this j method may be relatively simple to use, the disadvantages S ! outweight the advantages. | Recreation Standards j i i This approach, which is also widely used, attempts | to estimate demand for various recreational facilities i (parks, playgrounds, etc.) based upon assumed, standard I acreage requirements for each 1,000 or 100 population. | An example of such standards are those published by j National Recreation and Park A s s o c i a t i o n .77 These stand- j ards specify an estimated amount of space required for I i recreation areas and the facilities needed for particular | recreation activities based upon both the travel distance i | to a recreation site and an assumed proportion of popula- 1 tion that will engage in recreation. For example, the I j demand for regional parks is based upon the standard of i i 10 acres per 1,000 population. For each city containing 50,000 to 100,000 persons, therefore, there is a demand ^National Recreation and Park Association, Out door Recreation Space Standards (Washington, D.C.: National Recreation and Park Association, 1965). 192 for a regional park 500 to 1,000 acres in size within a 10 to 15 mile radius of the city.7® i i This technique, similar to the "visits per acre" method, does not consider the unsatisfied demands that would have existed or been more intense if the type and : ! quantity of facilities demanded by the public had been | | I made available. In addition, application of the typical j 1 standards to determine demand is unrealistic in that it ! i i | assumes a static situation by implying that no change is ! taking place in the variables such as population, the tastes and preferences of consumers, or other socioeconomic \ variables affecting recreation demand. According to | Daiute: It can be noted, parenthetically, that these standards have origins, years ago, in empir ical surveys. The standards have come to be used, however, as cheap and fast substitutes | for empirical studies. Standards have over simplified the solution to the problem of | first ascertaining demand; then relating de- j mand to supply; and finally, translating re- | lated demand and supply into acreage and design features.79 ! Thus, the use of recreation standards by themselves I | to determine the demand for outdoor recreation is unsatis- | factory. They are, however, useful in the sizing of facil ities, a matter which will be covered later in this study. 7®lbid., pp. 20-25. 79Daiute, op. cit., p. 335. 193 Time-Income-Mobi1i ty Researchers in the field of recreation economics | have made a good inductive case for the theory that lei- | sure time, income and mobility are causally related to participation in outdoor recreation.®® That is, income, j leisure time and mobility are associated factors to the extent that changes in rates of participation in outdoor i I recreation can be predicted from changes in these factors. The usual assumption is that each factor acts in dependently and fully, i.e., with an elasticity equal to ! one, and that the composite effect of the three acting ! together may be obtained by multiplication of the propor- ! tionate changes in the factors. For example, if the fac- i j tors are expected to rise 10, 25 and 50 percent, the j composite effect is equal to 1.10 x 1.25 x 1.50 = 2.06, ! and participation rates will increase 106 percent. This 106 percent increase is then applied to per capita activ- I ity measures of all kinds ranging from park visits to camping trips or swimming occasions. I | It is obvious that the time-income-mobility method | involves many arbitrary assumptions. In particular, the ! _______________________ ®®See, for example, Marion Clawson, "The Crisis in Outdoor Recreation," in American Forests (March-April, 1959). 194 assumption that each of the three socioeconomic factors will have exactly the same degree of influence on partic ipation is highly questionable. While the factors con sidered may be logically thought of as in some way j facilitating participation in outdoor recreation, it is j | also logical as well as desirable to estimate, rather than j I to assume, the degree of association between participation j and these factors as well as the degree of independence j and net influence of each. Such a procedure would enhance | the degree of rationality in the choices of recreation ! planners and decision makers. In addition, while these | three factors can reasonably be expected to exert an in- j fluence upon all measures of demand for outdoor recrea tion, there is no reason to expect that the degree of influence will be precisely the same for each measure. Conclusions With the completion of the two national recreation surveys, the socioeconomic methods of projecting recrea tion demand were developed and the direction of recreation research turned toward estimating demand and somewhat away from estimating benefits. The ORRRC studies which stem from these surveys established multi-variate relations 195 between recreation activity and various socioeconomic variables. The result was the ability to generate recre ation demand projections (in terms of participation days) which functionally related certain socioeconomic charac teristics of the population to a number of outdoor recreation activities. The socioeconomic methods of pro- j j jecting recreation all involve the assumption that current | relationships between recreational participation and so cioeconomic characteristics can be applied to the expected future socioeconomic structure of the population. | This method has been used in many statewide out- j door recreation plans prepared since the release of the j i ORRRC works. Demand projections using this approach typically apply the ORRRC participation rates (adjusted in some manner to account for variations assumed or observed in the characteristics of the population in the state under investigation) to the projected population to derive the expected number of total recreation participation days. These projections, however, have their limita tions. The ORRRC participation rates are not strictly demand phenomena as they also represent, necessarily, available supply opportunities. In addition, these rates 196 typically applied to determine demand only for a segment, and not all, of the population. Further, the projections do not account for the portion of various activities that do not occur at parks, which are the major supply compo- I ; nents for which the plan is concerned. Also, these ap- | j proaches do not take into account the effect of existing I i supply on recreation behavior or determine the optimum [ supply requirements, given this effect. Thus, these ap- ! proaches are deficient and can lead to gross overestimates | or underestimates of demand and supply requirements. In the past few years, econometric and probabal- ! istic models have been developed in attempts to overcome i the complex factors of the recreation market, and ap proaches geared toward statistical estimation of recrea tion demand and supply requirements have become the j ! primary area of emphasis. The attention in several promi- i nent studies has focused on methods to derive structural i | | demand functions and provide solutions to the "identifica- | tion problem" (non-separation of the effects of demand and | supply). However, these techniques depend on reduced form equations which specifically include supply variables in a demand model to project demand, or assume that supply al ways creates its own demand. For example, the reduced 197 form equations associated with a market clearing model and the user response models rest on the assumption that the effect on use of a given increase in facilities is con- ! j stant regardless of the magnitude or location of such an j increase. Because of the nature of such models, demand in t I I its true economic sense cannot be obtained and the end I result will be yet another aspect of the identification j ; problem. Construction of such models are in their infancy | and their application to statewide recreation planning is | extremely limited. i i Due to the complexities of measuring the projecting! j i I I j recreation demand, the "visits per acre," "recreation standards" and "time-income-mobility" approaches are com monly used. None of these approaches, however, consider potential demand and are essentially "consumption" oriented j j (i.e., based on previous visitation). I I i j Generally speaking, the socioeconomic methods are I the most appropriate, although modifications must be made i to overcome their deficiencies. Socioeconomic data change ! relatively slowly and are thus best suited for "target j | date" projections, ten, twenty or thirty years into the future. They are also well suited for large areas, such as the nation as a whole, large census regions and states 198 where adequate demographic data are usually available. It is also important that these data can themselves be predicted with some assurance. In addition, although I I these methods assume that current relationship between I I j recreational participation and current socioeconomic j characteristics can be applied to the expected future I j socioeconomic characteristics of the population, this ! assumption, although questionable, is probably necessary i 01 j if long-term projections of demand are to be made. 81 A. D. Little, Inc., Tourism and Recreation: A State of the Art Study, prepared for the Office of Re gional Development Planning, U.S. Department of Commerce (Washington, D.C.: Government Printing Office, 1967), p. 70. CHAPTER V | FACTORS AFFECTING THE DEMAND ! FOR OUTDOOR RECREATION i ! In developing an approach to forecasting future I ; i | recreation demand, it is necessary to analyze the factors | which affect the present demand for outdoor recreation among various subgroups of the population. From an under- j S standing of these determinants, projections can be made of ! the demand in future years. The purpose of this chapter is to identify which socioeconomic characteristics of the I ; | population, as well as other factors, are most relevant in j : l | projecting future demand and which may be disregarded. This information, together with that contained in the pre vious chapters of this report, form the background for the I approach for forecasting statewide demand for outdoor rec- i j reation that will be developed in the subsequent chapter. i Socioeconomic Factors ------------------------- i i | Economic analysis of the demand for most goods or i | services is concerned mainly with those variables which ! are reflected in the price of the commodity in question. Consumer preferences are usually assumed to be given or 199 200 held constant during the periof of analysis. However, with the absence of a market price for outdoor recreation, the assumption of constancy in preferences can be a severe | limitation if the demand model does not properly specify I those variables which underlie the consumer preference structure.1 ; There are several ways in which these preferences can be determined. One is that a behavioral science model ! might be sought that could be incorporated into the demand i model. A review of the literature in search of a behav ioral theory of outdoor recreationists^ suggests nothing sufficiently adequate for the needs of this study. This J is not to imply, however, that theorizing about leisure j time behavior has not been undertaken by behavioral scien- I I ! tists.3 a more rigorous framework would be necessary, | however, to serve the needs of this analysis. t 7 ^■Herbert H. Stoevener and L. J. Guedry, "Sociolog ical Characteristics of the Demand for Outdoor Recreation" in Cooperative Regional Research Technical Committee for ! Project No. WM-59, An Economic Study of the Demand for I Outdoor Recreation, Report No. 1 (San Francisco, 1968), ! p. 66. j | 2For a good bibliography, see Marion Clawson and | Jack L. Knetsch, Economics of Outdoor Recreation (Balti more: Johns Hopkins Press, 1966), Chapters II and III. 3M. H. Neumeyer and E. S. Neumeyer, Leisure and Recreation (3rd ed.7 New York: A. S. Barnes & Company, 1958) . 201 Another is that recent efforts in the area of mar keting suggest the analysis of the individual's entire decision process.^ Such an approach would allow the con sideration of product characteristics, the individual's environment and his perceptiveness, and the stimuli which he may receive, to be reflected in the analysis. Although promising, the usefulness of this approach has not yet been demonstrated in a conventional market situation and cannot therefore be adapted to outdoor recreation.^ A final approach would be along the lines sug gested by Katona.6 He indicates that for the study of consumer behavior, it is more appropriate to focus upon clearly defined variables and differentiate among groups of consumers on the basis of these variables, rather than ^F. M. Nicosia, Consumer Decision Processes: Mar keting and Advertising Implications (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1966). 6L. J. Guedry and Herbert H. Stoevener, "The Role of Selected Population and Site Characteristics in the De mand for Forest Recreation" in Cooperative Regional Re search Technical Committee for Project No. VJM-59, Economic Study of the Demand for Outdoor Recreation, Report No. 2 (Reno, Nevada, 1970), p. 88. ^George Katona, The Powerful Consumer: Psycholog ical Studies of the American Economy (New York: McGraw- Hill Book Company, 1960). 202 try to develop social-psychological models to reflect their behavior. Such a recommendation probably arises from the fact that: (a) psychological models are specific ^ to a given individual or group, and do not carry the gen- i | eral content exhibited by economic models; and (b) from j | a practical point of view, even if such a conceptual model ! could be developed, the measurement problems at present j | i may be prohibitive. j | Given the objectives of this study, the lack of j ! available behavioral theory in this area, information | ! i ! along the lines suggested by Katona was sought to deter- I I | mine which socioeconomic variables are important determi- ; nants underlying tastes and preferences of outdoor recrea tionists. The ORRRC and i BOR Studies j ! The significance of various socioeconomic charac teristics in explaining the demand for outdoor recreation I has been alluded to by numerous researchers. Clawson, | for example, explains that the estimates derived from his ; model are only an approximation of the demand curve for the total recreation experience because "... the popula tion in various distance zones may differ considerably in 203 terms of average income or of income distribution, as well as perhaps in other socioeconomic characteristics."7 i | In a more recent work, Clawson and Knetsch offer I I i a more elaborate discussion as to the role and signifi- ! cance of certain of these variables but do little more i | than describe past and future trends in population, lei- | ! sure time, transportation and income.® A part of their i I suggestions, however, arises out of the works sponsored j or conducted by the Outdoor Recreation Resource Review j 1 Commission (ORRRC) which contain the most extensive I I i efforts to collect information pertinent to the user pop- j ulation and its demand for given outdoor recreation j j facilities. The ORRRC studies have given impetus to the i | numerous suggestions that the inclusion of socioeconomic variables would increase the specification of recreation i demand models. They have also provided a data base for continuing periodic national recreation surveys, the i latest of which was that undertaken by the Bureau of Out- ! ! door Recreation (BOR). "^Marion Clawson, Methods of Measuring the Demand for and Value of Outdoor Recreation, Reprint No. 10, Resources for the Future, Inc. (Washington, D.C., 1959), p. 19. ®Clawson and Knetsch, op. cit., pp. 93-112. 204 Survey Research Center Survey.— The ORRRC spon sored a survey, undertaken by the Survey Research Center I (SRC) of the University of Michigan in November, 1959 and i May, 1960, of the leisure time activities of American adults.^ The sample used represented a cross section of I 2,750 household heads, 18 years of age and older, and ! ! their spouses throughout the nation. Questions were asked | : I to determine present patterns of participation, people's ! | wishes and aspirations to participate more frequently, i and barriers to increased participation in 11 different 10 recreation activities. ! I The socioeconomic characteristics of the respond- j i ents were noted and the respondents classified under 10 S socioeconomic groupings (family income, education, occu pation, age, etc.). The differences in participation be- i i tween the various subgroups of the population were then ; examined to identify the factors which made for high or low participation in outdoor recreation activities. The I concept was that where projections of future changes in I 9 Eva Mueller and Gerald Gurin, Participation in | Outdoor Recreation; Factors Affecting Demand Among Amer ican Adults, in ORRRC Study Report 20 (Washington, D.C.: Government Printing Office, 1962). l°Ibid., p. 3. 205 the relative size of these subgroups could be made, esti mates could then be derived as to how future participation would differ from present participation.1I An activity score was computed for each sample person by assigning values to whether participation in selected activities was mentioned spontaneously by re spondents interviewed during the survey or only after prompting, the number of activities engaged in during the j I previous year, and whether such participation occurred "a few times" or "often." Scores had a possible range from 0 through 26 and were a means for combining activ- j i ities into one measure of participation. This device was j not regarded as precise but was thought to be a satisfac tory method for ranking the population according to the degree of its participation in outdoor recreation. Differences in participation were examined in two ways: (a) by examining the activity related to each vari able without adjustment for the possible influences of the other socioeconomic variables; and (b) through multivari ate analysis to determine the separate effects of each independent variable on activity scores (the dependent variable) while holding all other factors constant. -t-^Ibid. , p. 10. | The SRC analysis of outdoor recreational activity ! by socioeconomic groups pointed to the geographic homoge neity of the American culture. Both urban-rural and re- ! gional comparisons revealed rather small differences in | participation in each outdoor recreational activity. ! Greater differences were found when men and women were j | compared and when age, income and other social status | i i i variables were used to classify people. ; ! I Participation in outdoor recreation was found to j ; I ' rise with income up to the $7,500-$10,000 income group and j : i j then declined slightly. Apparently in the lower income j | i I brackets, lack of money imposed some limitation on outdoor j I recreational activity. An increase in participation rates i over time could, therefore, be expected as more people move into the middle income brackets. j Breakdowns by population characteristics showed i | that women had distinctly lower participation rates than | men, and that participation declined sharply with advanc- ! ing age. Marital status and presence of children in the I I family did not seem to exert much independent influence | on outdoor recreational activity. ^ The second type of analysis was more important because its results could be interpreted causally. For i^ibid., pp. 27-28. 207 example, it could show the extent to which outdoor recre- ! ation activity levels are attributable to family income i | levels. Thus, by holding all other factors constant and ! | hypothesizing an increase in family income, the corre- ! i sponding increase in outdoor recreation activity could be i | | predicted. Proceeding from a survey of the factors af- I fecting outdoor recreation activity, therefore, a pre- I I dictive method could be derived. ! I I ; I The multivariate analyses on these scores yielded j ! relatively poor results. The coefficients of determina- j tion (r2) obtained were low — 0.30 for the overall re- j ! suits.in other words, only 30 percent of the variation j in recreational participation and vacation travel patterns could be explained by the variables used. This led the authors to conclude that: Factors other than socioeconomic characteristics j are major determinants of outdoor recreation ac tivity. Such things as time available, the goals ! and interests which the individual seeks . . . , the leisure time preferences of other members and i friends, physiological factors, recreational ex- j perience in childhood, interest in . . . compet ing activities . . . , or availability of facil- | ities come to mind readily.14 l^ibid., pp. 26-29. l^Ibid., p. 27. 208 In addition, the SRC report points out that per haps a somewhat greater proportion of the variance in outdoor recreation activity would have been explained if a more refined measure of participation could have been devised. Further, that a number of independent variables were intercorrelated, especially income, education, paid vacation, age and life cycle stage. Like all regression techniques, the multiple classification analysis cannot separate the individual explanatory value of several in tercorrelated variables with complete accuracy. However, the technique yields estimates of the net relation between the dependent variable and each of a set of intercorre lated independent variables which may be regarded as fairly good approximations. A detailed breakdown of the relationships between the socioeconomic variables and participation for outdoor recreation may be found in Table 22 of ORRRC Study Report 20.I6 These relationships are summarized in Table V-l. The National Recreation Survey.— The National Rec reation Survey conducted by the ORRRC staff in 1960 and 1961 obtained data regarding participation using nine ISxbid., p. 26. ^ Ibid. , p. 28. 209 TABLE V-l RELATIONSHIP BETWEEN SOCIOECONOMIC FACTORS AND PARTICIPATION IN OUTDOOR RECREATION Factor Influence on Participation Income Positively related Education of head Positively related Occupation of head Positively related to status Paid vacation Positively related Urbanization Negatively related Region West and North Central more active Age of head Negatively related Life cycle Negatively related to child im pedance and age Race Nonwhites less active Sex Males more active Sources Eva Mueller and Gerald Gurin, Participa tion in Outdoor Recreation: Factors Affecting Demand Among American Adults, ORRRC Study Report No. 20 (Wash ington, D.C.: Government Printing Office, 1 9 6 2 ) , Table 22, p. 28. 210 categories of socioeconomic factors, somewhat different in coverage and scope.^7 Recognizing that factors other | | than socioeconomic characteristics might affect recreation ! participation, the effects of expected changes in leisure | time and the opportunity to participate in recreational | activities (in terms of availability of facilities) were ' also included in the analyses.I8 The NRS obtained information by season of the | ' year, June, 1960 through May, 1961, from four separate I ) samples of population 12 years of age and over in four census regions in the continental United States. The regions included the Northeast, North Central, South and : West; the western region includes Arizona. Each sample | included approximately 4,000 persons and pertained to one i j season. The estimates used in the ORRRC projections dis- | cussed in Chapter IV of this report were primarily based on activities pursued during June-August, 1960, the sea son during which weather conditions are geographically | most uniform and participation in most activities is i i I --------------------- 170utdoor Recreation Resources Review Commission, National Recreation Survey. ORRRC Study Report 19 (Wash ington, D.C.; Government Printing Office, 1962). 18Ibid., p. 4. 211 greatest. The size of the NRS sample for June-August, 1960, was sufficiently large to show distributions of outdoor activity participation rates of the population by socioeconomic characteristics for 16 activities including: swimming, fishing, water skiing, hiking and others. In addition, estimates were prepared for hunting on the basis of September-November participation. The NRS survey showed variations in the relation ships between socioeconomic characteristics and recreation participation similar to those found by the SRC survey. The largest variations in participation were nearly always found within the income distribution, followed in order by the age-sex, place of residence and major geographical regional distributions. Income does not have the same effect on all recreation activities, however, as some activities are more responsive to changes in income than others. Per capita participation rates in playing outdoor games, swimming and sightseeing were found to increase with per capita income throughout the family income classes. In activities such as driving for pleasure, picnicking and camping, participation increases with in come up to the average income level and then levels off and even decreases for higher income levels. Only activ- 212 ities such as walking for pleasure, hunting and fishing showed no consistent pattern of response to change in per I Q capita incomes. i j The NRS study report also contained detailed anal yses which strongly suggested the existence of stable ) I associations between the socioeconomic characteristics j | i of the current population and the rates at which the pop- j ulation engaged in outdoor recreation. A factor analysis ! | | ; was made of intercorrelations between types of outdoor j | i recreation activities which enabled the 15 principal out- j ; | door activities to be aggregated into four homogeneous | ! i j t | groups. These were: (a) passive pursuits, (b) water- ! oriented activities, (c) physically demanding activities, and (d) backwoods activities. The purpose of this group ing was to identify a particular method for determining | similarities between individuals in their recreation pat- | terns. The object of the factor analysis was to test the I i validity of the four hypothesized groups.29 ! The analysis was carried out in great detail for I eight subgroups of the population and for 15 separate activities which could be classified, with varying degrees 19Ibid., pp. 62-68. 20Ibid., pp. 81-82. 213 of strength, into the four main groupings outlined above. | A "score" was calculated for each respondent reflecting j levels of participation in each activity. An attempt was I then made to "predict" this score. This was done by i | identification and analysis of 30 background variables i l I which statistically "explained" a significant part of the : variance in "activity scores." Thus, the data gave an index of the extent to which participation in any one activity was related to participation in the others and j to the possession, in varying degrees, of the 30 back- I ground variables. The 15 activities were grouped by means i j ! j of rates of participation and correlation. That is, the i | number of days (annually) of participation per person was j given, for each activity, as an index of relative popu- i j larity. Then, correlation coefficients were calculated i t j between each activity and the remaining 14 activities. j | This provided an index of the extent to which each activ- j ity was associated with the others and, hence, showed | which activities generally "go together." The analysis also indicated which of the 30 back ground variables were most significant for explaining levels of participation in each activity, and, hence, each activity grouping. This analysis would permit at least a 214 measure of prediction based upon the activity patterns that had been identified and their relationship to each of the 30 background variables. Again, however, the results were not very impres sive. Regression analysis performed on the socioeconomic factors for each sex and for each of the four regions yielded coefficients of determination ( r 2 ) of up to 0.44 in one case (western male participation in physically de manding activities), but in most cases the explanatory value of the variables fell between 10 and 20 percent (i.e., R2 of 0.10 to 0.20).21 From this analysis, however, it was possible to make some general statements regarding the pattern of dependence of the four activity groups on the socioeco nomic variables. These are listed in Table V-2. The 1965 Survey of Outdoor Recreation Activities. — Another survey, similar to the two previously discussed, was conducted for the Bureau of Outdoor Recreation by the Bureau of Census in September, 1965.22 This survey was patterned after the National Recreation Survey and con- 2-*-Ibid. , pp. 89-91. 2 2 U.S. Department of the Interior, Bureau of Out door Recreation, The 1965 Survey of Outdoor Recreation Activities (Washington, D.C.: Government Printing Office, 1968) . 215 TABLE V-2 RELATIONSHIP BETWEEN SOCIOECONOMIC FACTORS AND FOUR TYPES OF RECREATIONAL ACTIVITY Activity Group Passive pursuits Water-related activity Physically demanding activities Backwoods activities Pattern of Dependence The major variable affecting participation appears to be ed ucation; the participation rate increases with higher levels of education. Surprisingly, poorer health also goes with less pas sive pursuit activity. Nonwhites have lower scores, while those who live away from urban centers have higher scores. Occupational status is positively related to water-oriented activ ities. Participation in physically de manding activities is dependent almost entirely on age, with younger persons naturally having higher participation scores. Here, age and income are most strongly related, but the rela tionship is somewhat less clear than for the other types of activities. Source: Outdoor Recreation Resources Review Com mission, The National Recreation Survey, ORRRC Study Report No. 19 (Washington, D.C.: Government Printing Office, 1962), pp. 89-96. 216 sisted of almost 7,200 interviews with persons 12 years of age and over, with samples taken within each of the nine Census Divisions and the four Census Regions in the United States. Although somewhat different in technique i | and scope as compared with the NRS, the 1965 Survey ob- I | tained participation rates in individual outdoor recrea tion activities, by various socioeconomic subgroups of the ! population, as well as preferences for various outdoor ! activities. i i One of the major differences between the two sur veys was that the 1965 survey primarily covered the summer season (June, July, August and the first week of September i i through Labor Day). However, to obtain a measurement of non-summer participation, respondents were asked to esti- i | mate the number of days in which they participated in j various recreation activities from September, 1964, | through May, 1965. Another difference was that the 1965 i j survey covered a smaller sample of the population; how- | ever, the size of the sample for the summer season was larger as compared to that in the N R S .23 Moreover, sta tistical analyses to determine the association between the socioeconomic characteristics of the population and ^ ibid. , pp. 2-3. 217 the rates at which they participated in outdoor recreation activities were not made for the 1965 survey. However, as noted in the preceding chapter of this report, projections of future recreation demand were made by the BOR using the I data in both the NRS and the 1965 survey.^4 I Importance of Income | ! i I I In later analyses of both the SRC and NRS data, the ORRRC determined the net and composite effects of the six socioeconomic factors affecting future recreation de- j mand. It was found that income alone, as a single vari- ! able, could be used to estimate and project the composite j i effects of all six factors on future participation rates in outdoor recreation activities. This finding is not too surprising since there is a suspected high correlation between income and the other variables (education, occu- | pation, place of residence, age, sex, and to some extent, I ! leisure time). As stated by the ORRRC, "at least some intercorrelation with income can be deduced and in some 24 U.S. Department of the Interior, Bureau of Outdoor Recreation, Outdoor Recreation Trends (Washing ton, D.C.: Government Printing Office, April, 1967). 218 cases demonstrated for all other factors considered here."25 In their analyses, the ORRRC determined the degree ! of association between composite and individual factor ' effects. The coefficients of determination (R2) for the i ; 1960-1976 and 1960-2000 time periods were calculated ; across 17 activities for all two-at-a-time combinations | : of the composite and six factor effects. These data are j I indicated in Table V-3. Winter hunting was excluded from j I j i | the analysis to avoid double counting of this activity; ice skating and sledding or tobogganing were also excluded | I to avoid unusual geographical differences.2^ i A first order coefficient of determination, as j 7 | shown in Table V-3 is interpreted as the proportion of the variation in one variable which can be accounted for by the variation in another variable — in this case all variations are observed across the 17 activities. As was | noted earlier, intercorrelations among the six socioeco- | nomic factors considered here are often at least deduced. | | It should also be noted that these coefficients of deter- I | j - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - i ! 250utdoor Recreation Resource Review Commission, Prospective Demand for Outdoor Recreation, ORRRC Study Report 26 (Washington, D.C.: Government Printing Office, 1962), p. 12. 2^Ibid., p. 31. TABLE V-3 FIRST ORDER INTERRELATIONSHIPS — COEFFICIENTS OF DETERMINATION, R^ Factor i Factor j Income Education Occupation Residence Age-Sex Leisure 1960-1976 Period 6 Factor Composite 0.922 0.435 0.332 0.162 0.046 0.717 Income — .379 .222 .070 .065 .517 Education .379 _ _ .200 -.001 -.001 .190 Occupation .222 .200 — .310 -.173 . 315 Residence .070 -.001 .310 — .061 .176 Age-Sex .065 -.001 -.173 .061 — .001 Leisure .517 .190 .315 .176 .001 - 1960-2000 Period 6 Factor Composite .877 .398 .298 .181 .074 .665 Income — . 323 .157 .090 .090 .393 Education .323 — .160 -.001 -.211 .149 Occupation .157 .160 — .324 -.109 .241 Residence .090 -.001 .324 — .165 .122 Age-Sex .090 -.211 -.109 .165 — .001 Leisure .393 .149 .241 .122 .001 — Source: Outdoor Recreation Resources Review Commission, Prospective Demand for Outdoor Recreation, ORRRC Study Report 26 (Washington, D.C.: Government Printing Office, 1962), p. 31. to M v O 220 mination pertain only to the relationships among the net effects upon participation rates of changes in factors and not to the factors themselves. On this basis, little | intercorrelation exists and the net effects of the indi- i J vidual factors may be considered essentially independent. Comparison of the coefficients for the 1960-1976 I and 1960-2000 time periods indicates that linear relation-j j ships between composite effects and each of the factor I j effects are slightly closer for the shorter time period, ! but in both periods factors arrange in the same order of ! j i magnitude of this measure. The closest relationship | found for the composite effect is that with the income i : i j effect through which 92 percent of the 1960-1976 variation | and 88 percent of the 1960-2000 variation can be ex plained. As noted by the ORRRC: Estimates of total effects of six factors on activity participation rates based on the re lationship with net effects attributable to ! the income factor alone could, accordingly, | be offered as reasonably comparable to esti mates of composite e f f e c t s . 27 i ! Thus, by using the single variable income in projecting participation, the effect of the other five variables are also included. i 221 Sole use of the next most closely associated single factor — leisure — as the basis for estimating composite effects is clearly much less desirable than using income data for this purpose. Only two-thirds of j the variation in composite effects for the longer time period is explained through leisure alone. In addition, ! the changes expected over the next 40 years in each of | the remaining factors — education, occupation, place of | residence and age-sex — were estimated by the ORRRC to ! generally exert more modest net influences upon partici- j pation rates than the changes in either income or lei- | sure.28 i That income is the most important variable in recreation demand projections was also the conclusion i reached in the study made by Boyet and T o l l e y 2 8 and re viewed in Chapter III of this report. In their study, Boyet and Tolley attempted to project future recreation ! demand for the national park system based on a demand I ! analysis that included the effect of changing "demand j | shifters" on visitation. They started with the hypothesis i _____________________ 28ORRRC Report 26, pp. 31-32. 28Wayne E. Boyet and George S. Tolley, "Recrea tion Projection Based on Demand Analysis," Journal of Farm Economics. Vol. XLVIII, No. 4 (November 1966). 222 that income, age, education, residence and race were demand shifters. The regression equation they developed yielded visitation to a given park from a given state as j a function of distance, population, per capita income, j median age, median education, percent of state population I in urbanized areas and the percent of state population that was Caucasian. Because of a high cross correlation | with income, the variables age, education, percent urban ! and race were not found significant and were deleted from i the equation. The resulting regressions of park visits I on distance, population and income yielded high coeffi cients of determination ( r 2 ) with income having a greatly increased role as an explanatory variable. The function ! was further modified to visits per capita as a function of distance and income and this function was fit using least I | squares and used for the visitation projections. The regression coefficients were significant at a 1 percent | alpha error and most of the coefficients of determination | across parks were above .90, ranging as high as .96.30 j | Thus, income was again found to be a major ex planatory variable, along with distance and population, in projecting park visits. Further, by being able to 30Ibid., pp. 988-989. 223 omit other socioeconomic variables (such as age, educa tion, race, etc.) from demand equations, the estimating j ; procedure and data requirements could be simplified. The authors concluded that: | This simplification of the model has the addi- ! tional advantage of making it possible to ! estimate demand functions for visits in non- j census years since population and per capita i income estimates are available for noncensus I years.31 A similar finding was also made by the ORRRC in forecasting national park visits. In generating regres sion equations of national park visits per 100 persons j as a function of per capita income, per capita automobil travel, leisure time and time period, the ORRRC found a good relationship using all the variables together and using each variable separately, with the coefficients of determination ranging from 95.3 to 99.6. In its summary ( | of this analysis and the resulting equations, the ORRRC | ; noted that each of the income, leisure time and distance j traveled variable is "about equally correlated with vis its so that the addition of the other two to an analysis based on any one alone results in about the same improve ment in the amount of the variation explained."32 •^Ibid. , p. 987. •^ORRRC Report 26, p. 7. 224 Time and Distance Outdoor recreation, as found by the SRC survey, was the second most frequently engaged in leisure time pursuit (looking at television was first).33 However, as j I determined by both the Survey Research Center and the BOR i | survey, many people desire to participate more often in ! I outdoor recreation activities but are prevented from doing so primarily by lack of time (for the activity itself or for the trip necessary to reach suitable facilities). In j the SRC survey, for example, of the 60 percent of the j I people who indicated they would have liked to engage in I i more outdoor activity than they did, lack of time was the reason given by 52 percent of the respondents (see Table V-4) . In the BOR survey, of the 64 percent of those who indicated they did not participate in the activity as often as they would have liked, lack of time was the rea son cited by 58 percent of the respondents (see Table V-5). Although there are minor differences in reporting between the two surveys, the major importance of leisure time as a factor affecting outdoor recreation is evident. This importance was shown in the preceding analysis of ■^Mueller and Gurin, op. cit., pp. 31-32. 225 TABLE V-4 FACTORS PREVENTING DESIRED OUTDOOR ACTIVITY Reason Given Percent of Total Respondents Lack of time 52 Financial cost 17 111 health, old age 11 Family ties 11 Lack of available facilities 9 Lack of car 5 Lack of equipment 4 Miscellaneous 9 Don't know or not ascertained 4 Total 122a aTotal adds to more than 100 percent because respondents could mention more than one factor. Source: Eva Mueller and Gerald Gurin, Participa tion in Outdoor Recreation: Factors Affecting Demand Among American Adults, ORRRC Study Report No. 20 (Wash ington, D.C.: Government Printing Office, 1962), Table 7, p. 7. 226 TABLE V-5 REASONS FOR NOT PARTICIPATING AS MANY DAYS AS ONE WOULD LIKE — SUMMER SEASON Reason Given Percent of Total Respondents Lack of time 58 Financial cost 12 Facilities inadequate, too distant crowded or 18 Personal (age, health, fear) 5 Other 7 Total 100 Source: U.S. Department of the Interior, Bureau of Outdoor Recreation, The 1965 Survey of Outdoor Rec reation Activities (Washington, D.C.: Government Print ing Office, 1968), Table 5, p. 19. 227 the various socioeconomic factors affecting recreation demand. It may also be shown in terms of the increasing amounts of leisure time expected in the future. For ex ample, in 1930, the weekly hours of leisure per employed person was 15.0; by 1960, the number rose to 23.1. By 1976, the number of leisure hours per week per employed person is expected to reach 26.6, increasing to 30.6 by the year 2000.34 Effects of Time and Distance Closely interrelated with leisure time and par ticipation in outdoor recreation, however, is distance, i.e., where the facilities are located with respect to the recreationist. Distance effects demand because it influences the recreation decisions of the consumer, since "it is a characteristic of outdoor recreation that ordinarily the site is somewhat removed from the home."35 Normally, a consumer will use a given facility at a rate 34ORRRC Report 26, p. 6. -^Outdoor Recreation Resources Review Commission, "Outdoor Recreation and the Megalopolis," Vol. II, The Future of Outdoor Recreation in Metropolitan Regions of the United States, ORRRC Study Report 21 (Washington, D.C.: Government Printing Office, 1962), p. 15. 228 related to its distance from his home; the more remote it is, the less he will tend to use it. Or, given sev eral comparable facilities, all other things being equal, he will use the closer facilities more frequently than, or to the exclusion of, those more remote or distant.36 Although these relationships seem simple, they are actually quite complex. In any conventional analysis, outdoor recreation would be defined as a set of goods for which a consumer would give up part of his income, given the prices of all goods. Recreation, however, has an important pecularity in this respect. The consumption decision is not exclusively a matter of money and pro pensities to recreate; these factors can produce an effect on demand only if the consumer has the available time in which to engage in the activity. Distance becomes a prime barrier to consumption of the recreation "good" in that the consumer must transport himself to the facility (with about 90 percent using an automobile) ,37 an<^ this involves him in cost calculations. It not only costs 3^Harvey S. Perloff and Lowdon Wingo, "Urban Growth and the Planning of Outdoor Recreation," Trends in American Living and Outdoor Recreation, ORRRC Study Report 22 (Washington, D.C.: Government Printing Office, 1962), p. 91. ^ The 1965 Survey, pp. 29-33. 229 money to make the trip to and from the facility, but it also takes time. Since his supply of money and time is not unlimited, he must measure the satisfaction of the anticipated experience against that which alternative ! uses of his income and time would yield. In short, dis tance is a measure of the dominant costs confronting the recreationist, although travel costs are not necessarily the only costs. There may be costs for recreation equip- i ment, food, and so forth, but the travel costs, both in j time and money, cannot be avoided. Consequently, dis- j ! tance exerts an important influence on the consumer deci- | 1 sions through its implications for the relative location of facilities vis-a-vis the consumer and will thus effect the rate of participation in various activities. One of the main implications that may be noted relates to the location of future facilities. The great est amount of leisure time occurs in short time blocks, primarily at the end of the working day, on weekends and on holidays? while yearly vacations account for a large block of time, the total hours of leisure available dur ing this period are not as great.38 As a result, the greatest demand is for facilities located near the par ticipant's home for a day's or less than a day's outing 38ORRRC Report 21, p. 16. 230 (within a two-hour drive), followed by facilities that afford the opportunity for an overnight or weekend trip (within a four-hour drive).^9 Another major implication of the time-distance factor relates to the composition of the facilities them selves. As noted in Chapter II of this report, a charac teristic of an outdoor recreation outing is that it usually consists of participation in a collection of com plementary activities (hiking, sightseeing, etc.). However, considering two sites on which it is possible to j I accommodate the same collection of recreation activities, ! the level and type of demand confronting these sites will j depend on their location relative to population concen trations. 40 This implication has given rise to the func tional use classification of sites referred to in the recreation literature, i.e., day use, overnight or week end use and vacation use. The demand for day use recre ation facilities at a given location will be dominated by residents in the immediate area. Consequently, the 39The 1965 Survey, pp. 32-36. ^°Ivan M. Lee, "Economic Analysis Bearing Upon Outdoor Recreation," ORRRC Study Report 24, Economic Studies of Outdoor Recreatipn (Washington, D.C.: Govern ment Printing Office, 1962), p. 7. 231 composite of recreation activities provided for by day use facilities would normally be more heavily oriented toward day use activities such as picnicking, swimming, etc. On the other hand, camping is not ordinarily a day use activity and would take a much heavier weight in the overnight or weekend use composite. This is not to say that day use, overnight use, as well as vacation use activities will not take place at the same site. How ever, the demands for the various activities will be primarily related to the distance of the site from the residence of the recreationist. Thus, as seen from the above, time and distance have an important effect on the demand for outdoor rec reation in terms of the location and composition of facilities that must be provided. As noted by Lee, the inclusion of a distance variable in the demand relation would seem the appropriate way to recognize the effects implied (of both time and distance).41 Relationship of Dis tance and Cost Distance and the cost of getting to a recreation site have an important relationship to the location of ^ Ibid. , p. 29. 232 recreation demand. The importance of distance and cost to the location of economic activity has been made ex plicit in the works of location theorists. As noted by Isard, ". . . it is the distance factor that is the heart of locational analysis."42 Economic theories of location are, however, rela tively few when compared with the total body of economic j ! doctrine, especially in English-speaking literature. One j i of the chief reasons for this is that English classical economists typically assumed a static society, with time and distance given, in their analyses of economic activ- ! i ity.43 This tradition was continued to a considerable extent even in later years by renowned economists such as Marshall, who observed that the influence of time on mar kets was of more importance than the spatial aspects of economic activity.44 writers in the field of general equilibrium theory have also tended to ignore the aspect 42walter Isard, Location and Space Economy (Cam bridge, Mass.: The M.I.T. Press, 1956), p. 35. 43^dward H. Barker, "An Economic Analysis of the Changes in Retail Store Sales and Location in the Central Business Districts of One Hundred and Nine Central Cities in the United States, 1948-1958" (unpublished Ph.D. dis sertation, University of Southern California, 1963), p. 43. ^Alfred Marshall, Principles of Economics (8th ed.; New York: Mcmillan Co.^ 1920), p. 496. 233 of location on much the same grounds as the English clas sical and neoclassical writers. Economists such as Cassel, Pareto, Hicks, Lange and Samuelson assumed that the factors of production, the products and the consumers meet in one location, and as Newman observed, further assumed that "... there are no transport costs and no spatial resistance."45 j Development of a rational body of economic theory j i regarding location and a workable explanation of its im- j portance has been attributed to a group of German writers in what may be termed "The German School of Economic Lo cation. "46 The most prominent of these writers include: Johann von Thunen (the first writer of economic locational theory); Alfred Weber (the founder of this school of thought); Andreas Predohl, Oscar Englander, Werner Som- bart, Toward Palander, Hans Weigman and August L o s c h . 4 7 Von Thunen's theory regarding the effect of dis tance on location of economic activity was related to the location of agricultural production with respect to the 45phillip c. Newman, The Development of Economic Thought (New York: Prentice-Hall, Inc., 1 9 5 2 ) , p. 1 2 7 . 46;earker, op. cit. , p. 4 6. 47For an excellent review of the writings of the German and other location theorists, see Barker, op. cit., p p . 4 6 - 5 7 . 234 limits of cultivation for production of crops and trans portation costs. His basic hypothesis was that the type of crops raised will be in relation to their value, and these values will be effected by the costs of transporta tion.48 j Weber1s theory was based upon the location of man-j ufacturing rather than agricultural activities.48 In | establishing his theory, Weber claimed that the basic j s factors affecting the location of economic activity were: j transportation costs, labor costs, and, agglomeration forces (the advantage of producing \n one location as com- i pared to many locations). A firm will locate at a site I depending upon its evaluation of these factors.88 Modifications of Weber's thesis by other writers in this "school" were basically minor until the writings of Losch. Losch expanded Weber's thesis to include not only plant location but also urban concentration, markets, C 1 belts and regions. 48Isard, op. cit., pp. 15-19. 48Newman, op. cit., p. 128. 88Carl J. Friedrich, Alfred Weber's Theory of Lo cation (Chicago: University of Chicago Press, 1929), pp. 48-219. 51 August Losch, The Economics of Location, trans. by William H. Woglom (2d rev. ed.y New Haven: Yale Uni versity Press, 1954), p. 10. 235 C O Later writers, primarily American, have also made contributions to general location theory, two of the most prominent being Isard and Hoover. According to Isard, the major cause of the clustering of economic activity is the distance factor and the cost necessary to overcome it. In selecting a site location, a business firm employs the substitution principle in determining i the site of minimum cost in relation to transport costs ! I | (which include capital and labor outlays) and land j I costs.53 j Hoover asserted that the firm's location is deter- I I | mined by transportation and production costs. He claimed j that the type of industry that will locate on a given site is the one that can pay the highest rent over that of competitive users, given these costs.54 As seen from above, the primary construct in lo- ! cation theory regarding the location of economic activity j is transportation, and that transport costs to markets and natural resources are some function of distance. For a review of the more important of these writings, see Barker, op. cit., pp. 58-63. 53Isard, op. cit.. p. 33. S^Edgar M. Hoover, The Location of Economic Ac tivity (New York: McGraw-Hill Book Company, 1948), pp. 47-89. 236 Location theory, however, is not directly applicable to analyses of the location of outdoor recreation activity in the same manner that it would be for other economic ' activities. | Recreation is not marketed per se and is available | . at zero prices to the recreationists. There is, however, i a tendency to think of the locational potential for out- i door recreation areas as being heavily resource oriented. ! This is implied by the prominence given to the natural | endowment of a site or area for outdoor recreation uses. I Natural endowment may be important, but there is an impor- ! tant distinction between outdoor recreation produced at j a site and the more usual type of commodity produced in an economy. The outdoor recreation product is not trans portable. Final consumers of outdoor recreation must f I consume these services at the site of production. The i large bulk of ordinary goods produced has incorporated in I | its production a certain amount of transportation, and j the costs of transport become an essential component of ! | the costs of production. Ordinary goods are, on the whole, transported to consumers. In outdoor recreation, consumers transport themselves to the product.55 ^Lee, pp. cit., pp. 23-26. 237 While the relationship of distance and cost to | the location of demand for recreation activities is not j directly comparable with that for the location of most ! i | other economic activities, there are certain analogies I i with traditional location theory that may be drawn. This I | may be illustrated by the classical theoretical model of j j 1 agricultural location developed by von Thunen. In hypoth-j esizing the effect of distance and costs on the location i j > of certain crop production, von Thunen noted a pattern of I | | agricultural land uses in concentric rings around a mar- | ket center. He then developed a model which described I the limits of cultivation for the production of crops as i a function of distance from the market based on the market price per unit and the transportation cost per unit of each crop. The market price and transportation cost com- i bine to form what von Thunen called the "bid rent" func- ! tion or rent gradient. According to his theory, the crop | | with the highest bid rent function at a given distance from the market will take precedence over all other crops for cultivation at that location. The slopes of the bid rent curve for each of the economic activities thus deter mine the juxtaposition of their respective land u s e s .56 ^6Robert Sinclair, Von Thunen and Urban Sprawl," Cultural Geography; Selected Readings, ed. by Fred E. 238 A diagrammatic representation of von Thunen1s model, as devised from an analysis by Hoover,5? is shown in Figure V-l. ! The abstraction of the von Thunen agricultural i i model that is applicable to the outdoor recreation market { j is the importance of distance and travel costs to the lo- ! cation of demand for specific activities. If von Thunen's! 1 i | market place is thought of as a single population center, and the price per unit as the percent of population par- i I ! I ticipating in an outdoor recreational activity, a simple i outdoor recreation location model may be constructed as 1 j shown in Figure V-2. The rent gradient in this case would | relate to the percent of the population wishing to par- i ticipate in a given outdoor recreational activity and to the "friction" of distance per hour of travel.58 The | latter factor is a function of real travel costs and opportunity costs of the time spent in travel, as well as the population's collective perception of time spent traveling vis-a-vis enjoyment of participating in a given activity. Dohrs and Lawrence M. Sommers (New York: Thomas Y. Crowell Company, 1968), pp. 356-362. ^Hoover, pp. cit. , pp. 95-96. 58William Alonso, Location and Land Use (Cam bridge, Mass.: Harvard University Press, 1964). M A R K E T P R I C E P E R U N IT 239 R E N T G R A D I E N T F O R V E G E T A B L E S R E N T G R A D I E N T F O R W H E A T R E N T G R A D I E N T F O R L I V E S T O C K D I S T A N C E F R O M M A R K E T V E G E T A B L E S C E N T E R fc W H E A T L I V E S T O C K Figure V-l — Von Thunen Agricultural Location Model P E R C E N T O F P O P U L A T IO N P ART ICIP AT ING 240 G R A D I E N T F O R D A Y U S E F A C I L I T I E S G R A D I E N T F O R O V E R N I G H T F A C I L I T I E S G R A D I E N T FOR V A C A T I O N F A C I L IT IE S D I S T A N C E F R O M P O P U L A T I O N C E N T E R D A Y U S E F A C I L I T I E S N*. \ P O P U L A T I O N " C E N T E R ^ \ • / O V E R N I G H T F A C I L I T I E S V A C A T I O N F A C I L I T I E S Figure V-2 — Simple Outdoor Recreation Location Model 241 As the maximum distance that persons are willing to travel varies with the type of recreational activity that is being pursued, it can be inferred that various activities have different "bid rent gradients." While it i I is not possible to have all outdoor recreational sites available adjacent to place of residence, a consumer of outdoor recreation would still prefer to minimize move- j ment as there is still a matter of costs in terms of i | time, convenience and travel, complicated many times by j the congestion caused by modern transportation.59 Those I ' recreational activities for which there is the greatest j : frequency of demand may be expected to predominate in locations closest to the centers of population.^ I j i Valuation of Time | and Distance As previously noted in Chapter II of this report, one of the inherent problems in determining demand for outdoor recreation is that recreation is generally zero- priced and supplied by the public sector. In addition, 59Richard T. Ely and George S. Wehrwein, Land Economics (New York: The Macmillan Company, 1940), p. 71. 60Mueller and Gurin, op. cit., p. 39. 242 while contemporary consumer demand theory seems to apply to most commodities, it is not applicable to commodities such as recreation which require a high degree of consumer assembly (i.e., may not be purchased in a simple package) and entail the expenditure of blocks of leisure time.6* The travel-cost models developed by Hotelling62 and Clawson^ attempted to apply contemporary consumer demand theory to recreation, which has a high degree of consumer assembly and significant time requirements. Both authors, however, suggested the use of a monetary cost of the travel distance or distance of the facility from the residence of the recreationist as a surrogate for price. The cost of time was excluded from their anal yses. Time, however, is one of the costs that represents a restriction on the budget of an individual attempting f i 1 Robert R. Wilson, "Consumer Behavior Models: Time Allocation? Consumer Assembly and Outdoor Recrea tion," Cooperative Regional Research Technical Committee for Project No. WM-59, An Economic Study of the Demand for Outdoor Recreation, Report No. 2 (Reno, Nevada, 1970), p. 158. ^2H. Hotelling, Letter to Director of the U.S. National Parks Services (1947) as cited in Warren C. Robinson, "Economic Evaluation of Outdoor Recreation Benefits," Outdoor Recreation Resources Review Commission, Economic Studies of Outdoor Recreation. ORRRC Study Re port 24 (Washington, D.C.: Government Printing Office, 1962) . ^Clawson, op. cit. 243 to maximize his utility through the purchase of commod ities. The assignment of a cost to the amount of time used in recreation participation, while disregarded in the past, has been recognized by recent researchers as an important factor affecting recreation demand. Clawson, in a later writing with Knetsch, ranked { time along with money and distance as a cost in obtaining | the recreation experience.^ This conclusion is analo- I gous to the theoretical work in urban land economics by Isard, who, when describing variation in land sales prices with the distance from an urban core, measured distance in terms of "effective distance." As Isard noted, "ef fective distance is not synonymous with physical distance. Rather, it is physical distance adjusted in the time-cost dimensions."65 Whether a consumer can evaluate time and distance is a moot point. Thompson^ attempted to deter mine the extent to which individuals were able to evaluate driving distance and driving time, and whether or not ^Clawson and Knetsch, op. cit., p. 50. ^Isard, op. cit. , pp. 200-201. 66Donald L. Thompson, "New Concept: Subjective Distance on Store Impressions Affect Estimates of Travel Time," Journal of Retailing, Vol. XXXIX, No. 1 (February 1963), pp. 1-6. 244 their evaluation was affected by the character of the destination. He found that time and distance estimated made by respondents were grossly overstated or biased upward. The amount of bias depended on the individual's | | particular destination. | j Opportunity costs for time used in recreation have been observed to be a significant factor in demand | i j determination by Milan and P a s o u r . 6 7 Pearse6® attempted ! to place a value on time by multiplying the number of hours spent in travel to and from a recreation area by the individual's estimated wage rate. This is similar to the position taken by Henderson and Quandt who attempted | to show mathematically how the "... rate of substitution i j j of income for leisure equals the wage rate."69 67r . l . Milam and E. C. Pasour, "Estimating the Demand for an On-Farm Recreational Service," American Journal of Agricultural Economics, Vol. LII, No. 1 (February 1970). ®®Peter H. Pearse, "A New Approach to the Evalua- j tion of Non-Priced Recreational Resources," Land Economics, Vol. XLIV, No. 1 (February 1968). fiQ James M. Henderson and Richard E. Quandt, Microeconomic Theory (New York: McGraw-Hill Book Company, 1958), p. 23. 245 That travel time is considered a cost has also been borne out by studies made concerning the economics of highway congestion.7® Here, congestion is considered a factor since it causes an increase in the time it takes j for an individual to get from one location to another. | Congestion, in addition to causing increased driver irri- ; tation, tends to lower the utility which an individual might obtain from traveling, thus lowering the total util ity which he might gain from his total recreation experi- | ence. In addition, most motorists consider the possi- j bility of highway congestion when taking a trip or plan- I ' ning an outing and place a cost on time required totravel j to and from the area.^ ! A theoretical approach to the cost of the time j i different from the traditional theory, exemplified by | Henderson and Quant,73 and utilized by Pearse,73 was for- I j | - ■ - M . , , - - ■ ■ | | 7®M. Bruce Johnson, "On the Economics of Road ! Congestion," Econometrica, Vol. XXXII, No. 1 (February 1964)7 and Edward F. Renshaw, "The Economics of Highway | Congestion," Southern Economic Journal, Vol. XXVIII, j No. 4 (November 1962). i 7 - j I Renshaw, op. cit., p. 374. 7^Henderson and Quandt, op. cit. 7 * 3 /JPearse, op. cit. 246 mulated by B e c k e r .74 Becker's simplified model is based upon households being both producing units and consuming units which combine both time and market goods to produce i I commodities. Combinations of the commodities produced j are then chosen in such a way as to maximize a utility | function subject to a time and income restraint. Becker combines his time and income restraint into a "full income" restraint which represents the income that could ". . .be obtained by devoting all the time and other resources of a household to earning income, with no regard for consumption."75 He then defines a utility i function which is a relationship of direct time costs and j indirect time foregone in obtaining the combination of | commodities which would maximize utility subject to the i "full income" restriction.7^ According to Becker's theory, the importance of I the foregone wage rate is directly related to the foregone i ! earnings per hour of time and the number of hours used per ! unit of goods produced by households. He claims that the i --------- M ------------ Gary S. Becker, "A Theory of the Allocation of Time," The Economic Journal. Vol. LXXV, No. 3 (September 1965) . 75Ibid., p. 497. 76Ibid., pp. 497-500. 247 importance of foregone earnings is determined by the "... amount of time used per dollar of goods and the cost per unit of time."77 Therefore, the time intensive ness of commodities would determine the importance of foregone earnings if the cost of time was constant among commodities. According to Milam and P a s o u r ,78 however, the wage rate as a measure of time value may be too low when alter native uses for time yield more than the wage. For ex ample, if a consumer substitutes an afternoon of golf for an afternoon of work at a loss of that afternoon's pay, then the golf was more valuable than the work in terms of I money. However, if the consumer did lose the pay and en- ! | joyed the golf as well, the value of time is not revealed. If the consumer had the afternoon off and had no gainful opportunity for its use, the value of golf could not be | assessed, but would certainly be less than the wage. i i On the other hand, according to J o h n s o n , 79 the ! wage rate as a measure of leisure and travel time tend to i | ------------------------ 77Ibid., p. 503. 78Milam and Pasour, op. cit. 7%. Bruce Johnson, "Travel Time and the Price of Leisure," The Western Economic Journal, Vol. IV, No. 2 (May 1966). 248 overstate the time cost of these activities. Johnson, in his attempts to evaluate time for use in the study of travel time and commuting to and from work, developed a ! model by which the value of travel time could be esti- | mated. The analysis was based on the assumptions that: j | (1) . . . the individual's behavior is subject | to a time budget constraint as well as a money | budget constraint, and (2) that work and lei sure are distinct decision variables in the i utility function.80 j j Johnson also assumed that time is a scarce re- ! source to the individual, that the level of money income may be chosen, that goods and services have time as well j as money prices, and that trips cost both time and money.Based on his analyses, he found that "... the value of leisure and the value of travel time must be less than the money wage rate."82 At best, it can be said there is no rigorous method that has ready application to the evaluation of ! I time and distance in consumption. Further, as shown by I I the mathematical expositions and "variable proportions i time allocation model" developed by Wilson,83 distance, 80Ibid., p. 135. 81Ibid., p. 136. 8^Ibid., pp. 137-139. OO Wilson, op. cit. 249 recreational travel, travel costs and time costs are not prices. Rather, they are parameters in a demand function which do not necessarily have an assigned, explicit mone- ! tary value. As Wilson notes: There is no evidence that distance is properly a surrogate for price except that distance di minishes, one would expect both the amounts of activities and facility use to increase via | time substitution. The distance parameter may be viewed as a lower bound for recreational travel, an activity demanded jointly with ac tivities at each recreational site. Distance, recreational travel and travel cost are not I prices. The use of production parameters, ! activity quantities or total costs of activi ties as surrogates for prices would appear to lack economic justification. The distance I from the residence to the recreational site is a parameter in the production of activities from a facility. As such, it is a parameter in the consumer's demand function both for the facility and for activities associated with it.84 Conclusions | From the preceding, it may be concluded that in- ! come is the best single socioeconomic variable for pro- ; jecting future recreation demand. It is a good summary i measure of and highly correlated with a large number of socioeconomic factors influencing recreation demand. By using income, the effects of these other variables are ®^Wilson, op. cit., pp. 177-178. I 250 included. Moreover, by using income as the sole explan atory socioeconomic variable, intercorrelations with other j socioeconomic factors can be avoided. i ! Further, the use of only one socioeconomic vari- | able in the demand model (and the elimination of the | numerous, more complicated socioeconomic factors which | have little utility) makes the model easier to construct and use and hence a more practical tool for recreation I planning. In addition, the ORRRC and BOR participation j ! rates are given by various categories of income distribu- j tion, thereby easing the data collection and analytical | tasks of the recreation planner and allowing analyses of the income effect on demand between income levels, which j | varies significantly. Time and distance also have an important effect [ [ on recreation demand in terms of the location and compos- | ition of facilities that must be provided. The variables i | are not, however, surrogate prices for recreation. Rather, they are parameters in a demand function and use of the distance variable alone will account for the im plied effects of both time and distance. CHAPTER VI AN APPROACH TO FORECASTING STATEWIDE OUTDOOR RECREATION DEMAND AND | | SUPPLY REQUIREMENTS ! S I I A wide variety of approaches have been developed j to project outdoor recreation demand and supply require- ! ments. There are a number of deficiencies in these meth- I | ods and their applicability to statewide outdoor recrea tion planning is extremely limited. In this chapter, an i approach will be developed and an empirical model, using 1 ] theoretical microeconomic constructs as well as other | theories regarding recreation developed in recent years, j will be formulated to enable such forecasts. To provide data for development and application of the model, a case study will be made for one state — Arizona. | Scope of the Model j There are three critical questions that must be t j resolved by the recreation planner to enable preparation | of a comprehensive recreation plan: 1. How can a true measure of future recreation demand be determined, given current data limitations, the 251 252 complex nature of the recreation market, and the time and financial constraints typically faced by the planner?? i i i 2. What effect does the existing supply of fa- ; cilities have on recreation behavior?? and, | 3. What are the amounts, types and optimum com- ! binations of facilities that must be supplied, and where should they be located to best satisfy demand, given the ! effect of existing supply on recreation behavior? I Answers to the first question will depend on the i J type and extent of data available, the time and money available to collect new data, and the time and money ! , i I available to conduct the actual analysis.1 Unfortunately, i i there are still insufficient data regarding recreation ! i ! demand, particularly at the state level, to provide rea sonably accurate parameters to derive a true demand func- | tion for outdoor recreation. Moreover, the time and cost ! j to collect such data and conduct the analysis are still i i | prohibitive in almost all cases. However, at least some | j approximation of future demand must be made, given these constraints. On the other hand, the approximation must A. D. Little, Inc., Tourism and Recreation; A State of the Art Study, prepared for the Office of Re gional Development Planning, U.S. Department of Commerce (Washington, D.C.: Government Printing Office, 1967), p. 75. 253 be as accurate as possible to be useful for recreation planning, programming and budgeting. The demand estimating procedures commonly used have been deficient in a number of areas. For example, most studies usually consider only the demands of the population in a certain age group, rather than the de mands of the total population. Also, while consideration is sometimes given to the socioeconomic factors affecting demand, there is little or no proper account taken of the | effects of travel time and distance on demand. Although j I some approaches attempt to incorporate the effects of j I travel time and distance, they usually exclude the socio economic characteristics of the population and improperly use past visitation (or consumption) data to determine demand. On the other hand, when socioeconomic and dis tance factors have been included in some models to esti mate demand, reduced form market clearing equations, which explicitly include supply, are used to project de- ' mand. In any case, a true measure of demand, or for that matter a reasonable approximation of potential demand, is not obtained and the identification problem is not solved. Moreover, when estimates of potential demand are made (albeit inadequate), there is little or no account taken 254 of the proportion of participation that does not occur at public recreation areas, such as parks, although a j park and its related facilities are the primary type of ! supply component planned for and provided by public agen- i cies. Thus, current approaches to estimating future i 7 I j demand can either grossly overestimate or underestimate i potential demand and thus have limited value for recrea- ! tion planning. j Solutions to the second question regarding the | effect of existing supply on recreation behavior have been ignored in previous approaches. Consideration of this i i effect is crucial to recreation planning as the planner j i must know the influence or attraction that the amount, i 7 type and location of existing supply has on recreation behavior, as new facilities must be provided within this framework. Otherwise, optimum use of existing resources is not possible. i Answers to the third question concerning optimum supply requirements are closely related to the second « • > * issue noted above, but again, previous methods have been deficient in this aspect. Usually, determination of supply requirements is made by comparing the amount of facilities demanded (typically derived from a recreation 255 standard) with the existing level of supply, with the difference being a deficiency (under supply) or a surplus (over supply). This technique does not specify the opti- | mum combination of the amounts and types of facilities, j or their proper locations, to best satisfy demand given i f j the influence or attraction of existing facilities on the i ! behavior of recreationists. Consequently, this approach i is also inadequate for recreation planning. j I To help resolve these issues, the approach devel oped in this chapter will be geared toward enabling pro- i jections of potential outdoor recreation demand and the | optimum combinations of facilities that must be supplied I to meet this demand. The empirical model that will be formulated will rely on data that are currently available j or that can be obtained easily so as to minimize the j | costs of data collection and analysis. In addition, it will be designed to serve as a useful planning tool that will overcome the deficiencies noted in previous ap proaches and help provide answers to the critical ques tions currently faced by recreation planners. As such, it should be of assistance to other states in preparing comprehensive plans for outdoor recreation. 256 Underlying Theory of the Model The underlying theory of a planning model is the i set of functional relationships, stated or implied, that | is assumed to prevail between the "subject" of the model i j ; (or the activity it projects, allocates or manipulates) and the relevant variables. It is impossible to make a I "theory-less" model, as the model either derives directly ; from theory as a symbolic statement of it, or abstracts i ! certain phenomena to symbolic form and relates them i structurally, thus creating theory. Models are thus j either theory-based or theory-laden, or sometimes a com- I j bination of both.2 i j Where theory exists, the model builder usually relies heavily upon it for his organizing principle and deduces the model based on the theory. The result is a i "theory-based" or "deductive" model in which the relation ships between the activity it projects and the relevant j | variables are expressed by a series of precise, but often limiting, mathematical statements. Examples of theory- based models are those employing the principle of gravity, market theory or location theory. I ---------------------- ^Maurice D. Kilbridge, Robert P. O'Block, and Paul V. Teplitz, Urban Analysis (Boston: Graduate School of Business Administration, Harvard University, 1970), p. 13. 257 In the absence of theory, model builders strive ! to find regularities and derive relationships as best they can, frequently without fully understanding the na ture of the causality of derived statements. Examples of such models are those developed for biomedical programs. Relationships are put together pragmatically, frequently using multiple regression techniques or systems of simul- | taneous equations to create a planning model. Since the [ i mathematical statements comprising the model are theory- j i laden, the analyst has in effect created theory without j necessary reference to generality or logical coherence. | | Thus, such "inductive" or "theory-laden" models embody I theory implicitly.^ I | The model developed in this chapter is both theory-based and theory-laden. It relies heavily on ex isting theoretical microeconomic constructs as well as [ other theories regarding recreation that have been devel- | oped in recent years. However, due to certain data lim- i ! itations and the complex nature of the recreation market, other constructs must be developed which are theory-laden. The relevant constructs underlying the model are discussed in the following. 3Ibid., pp. 13-14. 258 Demand The model utilizes the theoretical concept of potential demand, i.e., the various amounts consumers | desire or are willing to purchase in a given market in i | a given period of time at all possible alternative prices | per unit, ceteris paribus. It is hypothesized that poten tial demand is the amount of a particular outdoor recrea tion activity (or various quantities) that a recreationist ] | (demander) desires or is willing to engage in (or pur- i chase) in a given market in a given period of time at all i possible prices per unit, ceteris paribus. Here, the quantity demanded will be expressed in | terms of visits to a recreation area inasmuch as visits | are the commonly accepted units of measurement in recrea- i i j tion economics. However, it should be noted that the | | total number of visits to a site is not the same as the number of days of participation in a recreation activity | as developed by the ORRRC to measure demand. The total number of visits refers to the total number of people that come to a park irrespective of their length of stay. For example, a person visiting a park for the purpose of camping may stay several days but his visit would only be counted once, when he entered the park. However, the 259 number of participation days is the number of days or parts of a day that an individual engages in a particular activity. ^ While visiting a park, a recreationist could stay for several days and each day would be counted as a | day of participation or a visitor day. j Inasmuch as the ORRRC data will be used in this ! study, visitor days (or participation days) will be used ! to determine potential demand and then converted to the ! j number of visits to provide comparability with certain j data that are available for Arizona. As to a "price" for recreation, it was noted ! earlier that there is no conventional market pricing for j outdoor recreation and that as a result, a wide variety j i of approaches have been developed in an attempt to over- | come this "zero price" problem. For example, the i "travel-cost" approach attempts to assign explicit dollar i | values to the cost of travel to a recreation site as a | proxy for price. This has been expanded by some econo- j | mists to also include the dollar cost of "time" either j ! for travel to the site or to participate in the activity, Outdoor Recreation Resources Review Commission, National Recreation Survey, ORRRC Study Report 19 (Wash ington, D.C.: Government Printing Office, 1962), p. 109. 260 as price surrogates. However, there are many unresolved problems in assigning such values and attempts to compile accurate cost data have been unsuccessful. Moreover, while recreation travel and time costs have an important effect on recreation demand, they are not prices per se as they cannot be assigned an explicit dollar value. i Rather, they can be parameters in a demand function which [ do not have an explicit monetary value, and the use of I | distance as a variable in a demand function would be an | appropriate way to recognize the implied effects of both i time and distance. Thus, distance of a recreationist from a recrea- i tion site may be thought of as an implicit price for ! j recreation in the demand function. Hence, potential de- t mand or the quantity demanded may be expressed as: | j VD = quantity of visitation demanded at a ! ^ 1 particular distance, d. In addition, it is hypothesized that potential demand for outdoor recreation will respond to most of the same factors that affect consumer demand for most other goods and services such as: tastes and preferences; socio economic characteristics, such as income; certain dis 261 utilities, such as the time spent to get to the recreation area; and, the availability of alternative uses of leisure time and money. Here, the disutility associated with the purchase of a good or service is assumed to be approxi- ! mated by the "implicit price," distance. Also, the avail- I ability of alternate uses of leisure time and money will J be assumed implicit in the measurement that will be used | to determine the tastes and preferences of recreationists | for various activities. ! Thus, an individual's demand for outdoor recrea- ! | tion, maY ke expressed as: i | V_ = f (tastes and preferences, socioeconomic I d S characteristics, distance) i i Although attempts have been made by some econo- I I mists to consider some measure of the quality and use of | supply as demand determinants,^ these efforts have been | unsuccessful. Measuring the quality or degree of excel- | lence of recreation areas and facilities is a difficult | matter as it is largely judgmental and highly variable in | that it involves individual tastes and preferences. It 5C. J. Cicchetti, J. J. Seneca, and P. Davidson, The Demand and Supply of Outdoor Recreationt An Econo metric Analysis (New Brunswick, N.J.: Rutgers-The State University, Bureau of Economic Research, 1969). 262 also involves individual ranking as to the attributes or character of an area cjnd how it can contribute to or effect the recreational experience. In addition, the use of an area is not a reliable indicator of quality as i people will use an area no matter how poor its quality | if there is no convenient alternative. i The difficulty in obtaining adequate measures of j quality was noted by the ORRRC. In attempting to develop ! "opportunity indexes" for recreation areas (which included ! a measure of quality and use), the ORRRC concluded that ; such measures were too judgmental and lacked the neces- ! sary precision for inclusion as independent variables in t its analyses.^ i ! Attempts have also been made by some economists I j to incorporate previous experience in various recreation j activities (the so-called "learning by doing" phenomenon) i • - ! as a demand determinant.' However, the ORRRC found this factor invalid as a demand determinant for all activities j except swimming.® I ^Outdoor Recreation Resources Review Commission, Prospective Demand for Outdoor Recreation, ORRRC Study Report 26 (Washington, D.C.: Government Printing Office, 1962), p. 38. ^Cicchetti, et_al., op. cit. 8ORRRC Report 26, p. 33. 263 In view of the above, potential demand in this study will be determined as a function of an individual's tastes and preferences, his socioeconomic characteristics, and the distance of a recreation site from his home. supply In microeconomic theory, supply is defined as the various quantities that the supplier would be willing to place on a given market in a given period of time at all | possible alternative prices, ceteris paribus. In this j | model, supply is defined in a somewhat different manner ! | and used in a special way although the basic concept is still within the framework of microeconomic theory. It is hypothesized that supply is the amount of facilities to accommodate recreation visitation (or vari- | ous quantities) that a public agency or recreation planner s | (the supplier) would be willing to place on a given market | | in a given period of time at various prices, ceteris pari- I bus. Here, visitation is considered a measure of quantity supplied inasmuch as a public agency provides facilities to accommodate visitation. This visitation, as will be shown later, will be converted into facilities' require ments. 264 In terms of price, it was noted in Chapter II that the unique characteristics of the outdoor recreation market also effect the supply function. Since outdoor recreation is normally zero priced and provided by the I public sector, the cost of providing recreation facilities does not ordinarily enter the supply function in the same | manner as it would for a private industry in microeconomic theory. Nevertheless, since the recreation supply is not I distributed uniformly, the cost of moving the recreation- j ist (the demander) to the source of supply must be added | [ to the supply function (marginal cost curve). In this I | sense, the recreationist produces some portion of the j recreation good. When the marginal cost at the site is | zero, then the costs of time and travel to the site be- i 7 I | come the marginal costs for the recreationist in producing j or providing himself with a unit of the recreation good. i j These costs, however, are traditionally ignored by economists in analyzing the demand for and supply of most | goods and services. If they are considered, the are usu ally netted out or subtracted from the demand price for a given quantity demanded, since they are viewed as costs (or reductions in demand) accruing to consumers. In other words, the traditional price-quantity demand curve 265 reflects the price net of delivery costs. However, the supply points for outdoor recreation (e.g., a park) are more sparsely and unevenly distributed than the supply points for other goods and services (e.g., a supermarket). Also, there is no usual market price facing the recrea tionist or a traditional price-quantity demand curve from which an explicit price for time and travel can be sub tracted. Since it is important to include the costs in volved when a recreationist "consumes" visits to insure that outdoor recreation is not incorrectly considered a "free" good, the delivery cost or cost of time and travel must be introduced into the supply function. Therefore, the costs of time and travel, represented by distance, will be considered an implicit price for supply. Thus, the quantity supplied may be expressed as: Vs^ = quantity of visitation supplied at a particular distance, d. It is further hypothesized that the quantities of facili ties a public agency would be willing to provide is a function of the facilities themselves. As stated earlier, a park is the typical supply component. Further, parks usually contain a composite of facilities to accommodate 266 a number of activities. In some instances, man-made lakes or "artificial" resources are provided, as well as the land area occupied by the park itself. These types of park characteristics accommodate and represent poten tial visitation; they can be supplied and quantified, while other characteristics, such as scenic beauty and quality of the park cannot. Also, while there may be budgetary or other limitations faced by the recreation planner that would effect the amount of facilities sup plied, the purpose of this approach is to determine the amount of facilities that should be supplied to meet potential demand. Consequently, this constraint will not be considered. For purposes of this analysis, then, the quantity supplied, Vg^, may be noted as: Vc = f (facilities, distance) bd ’ While the above formulation is within the basic framework of traditional economic theory, supply will be used later in a special way to determine optimum supply requirements to meet potential demand, with due account taken of the effect of existing supply on recreation behavior. This part of the model is theory-based as there is no existing theory for the formulation that will be derived. 267 The Market Clearing Equation and "Price" The primary theoretical construct underlying the model is the market clearing equation of microeconomics i | that equates demand and supply. That is, in a given mar- j ket, the usual market transaction involves the meeting of j ! demand and supply at a price where buyers are willing to I purchase a certain amount and suppliers are willing to S i j supply the same amount. At this market price or equi- ! ! librium point where the demand curve and supply curve i j intersect, the market is cleared. I Based upon the above formulations, demand and | supply are both expressed in terms of the number of vis- | its. Further, distance is the implicit price in both the demand and supply functions. Consequently, the market clearing equation is: vn = vq Dd d However, the manner in which the demand function can be estimated and equated with supply in the market clearing equation must recognize the identification problem. 268 The Identification Problem The identification problem in the determination ! of recreation demand has been linked with the use of time j I | series data regarding visitation to particular sites. | When using such data, it is difficult to tell whether the i supply or demand functions are being estimated, since j both functions can shift over time. As remarked in Chap- j I ter II, what is needed is a situation where the demand curve exhibits little variation while the supply curve | shifts over a wide range so that the points of intersec- ! tion or equilibrium more closely approximate the true | demand curve (see Figure II-5 [B]). | However, because it can rarely be assumed that a demand relationship will hold relatively constant over an adequate time span for collection of time series data (and this is especially true for outdoor recreation), and because the researcher cannot know a priori the degree of i shifting in such functions, attempts are often made to hold the demand function constant by the addition of ap- ! propriate "shift" variables to the model. Even if other variables are added to each of the functions, however, there still may be problems in identifying the relations. 269 Simply adding a variable to one of the equations may not help in identifying the system if the variable added does not materially influence the dependent variable in the equation to which it has been added, or if it does not ! exhibit much variation. I | Demand analysis using cross-sectional data, how ever, may have identification problems of a different I I sort. When using cross-sectional data, it is assumed | I that all the sample persons are homogeneous except for I random errors and the differences in the variables in- i eluded in the model. As noted by Klein: "A significant l fact about cross-section samples for a single period is that price variables, and indeed other market variables t such as interest rates and wage rates, are effectively held constant.1,9 However, in order to estimate the effect j of one variable on another, an independent variable must i assume at least two values. For most commodities, all j 7 i i consumers face approximately the same price during a given time period. For this reason, cross-sectional data I are often inadequate for estimating demand functions for g L. R. Klein, An Introduction to Econometrics (New York: Prentice-Hall, Inc., 1962), p. 55. 270 most goods and services since the price of the good is the key explanatory variable of the demand function. ! Because of the unusual nature of the costs asso- i ! ciated with the consumption of recreational "goods," j I however, derivation of demand curves for recreation using I j cross-sectional data is not only possible but appropriate. This has long been done through the "travel-cost" ap proach to recreation demand studies. Cross-sectional | ' statistical demand curves for recreation developed by the j j travel-cost approach are usually derived by assigning an explicit cost for the recreation experience, either esti- | mated or as some function of distance traveled, as a j proxy for price. Using this cost as the price variable ! I for a demand function thus insures an independent vari- I { able that will take on a wide range of values, since the { cost varies considerably. Moreover, because of the 1 cross-sectional nature of the data used, demand vari- i ' j ability is effectively removed and demand thus held more or less constant. j | Typically, however, such studies are site oriented, i.e., the demand curve is derived with reference to a specific site. Depending on the particular model and type of data used, such studies provide little or no vari- 271 ability in the associated supply function. The ability to trace the true demand relation is therefore reduced because the supply shifts are not observed and an equi librium point between the demand and supply curves is not determined. In addition, while variation in the price variable is assured, explicit account of the influence of the socioeconomic factors affecting demand is often omitted because of data limitations. Although a site-oriented demand curve may be use ful for developing operating policies for particular facilities, it is not sufficient for use in comprehensive statewide recreational planning. In such cases, the esti mation procedures used must be focused upon the demands emanating from a given population. Only in this way can information be obtained for making public decisions on the tradeoffs among the alternative policies necessary to implement a recreational program for a wider geographical area. Questions relating to the proper mix and spatial distribution of recreational facilities to best meet the preferences of people cannot be adequately answered with out the quantification of demand within this broader I orientation. i 1®Marion Clawson, Methods for Measuring the Demand for and Value of Outdoor Recreation, Reprint No. 10, Re sources for the Future, Inc. (Washington, D.C., 1959). 272 Thus, although certain aspects of the identifica- | tion problem are solved through the use of the travel-cost I approach, other aspects still remain. As noted by Kalter ■ and Gosse: i Excess dependence on site-oriented demand studies | as opposed to the more traditional market studies | undertaken for other commodities has been partially i responsible for our inability to solve identifica- i tion problems.H If, however, the travel-cost approach was modified to em phasize the traditional market type of demand study where ; potential demand is considered (i.e., based on the various socioeconomic factors of the population known to affect demand) rather than attempt to derive structural demand i equations (based on time series visitation or consumption i data), more appropriate informational outputs could be i obtained and the remaining aspects of the identification I i j problem could be solved. This could be done by formu lating a model which: 1. Is population rather than site specific to ! provide a measure of potential demand, and by empirically implementing the model by using distance as the implicit H r . J. Kalter and L. E. Gosse, "Recreation Demand Functions and the Identification Problem," Journal of Lei sure Research. Vol. II, No. 1 (Winter 1970), p. 50. 273 price variable. If this distance (implicit price) vari- ! able was derived based upon data regarding recreation | travel patterns obtained from a large cross-sectional sample of households residing in dispersed geographic areas, it would take on a wide range of values because of | | its dependence on the variation typically found in the location of supply. This variation stems from the fact that recreation facilities are not uniformly distributed and that the amount supplied depends not on market forces i but on the decisions of public bodies. The various dis tances (or implicit prices) that people would be willing to travel (or pay) for the recreation "good" would thus ■ help to identify the demand function. i 2. Relies on cross-section data regarding cer- ! tain socioeconomic characteristics of the population known to influence recreation participation, to determine potential demand. If cross-sectional data are used to estimate the demand curve, the function could be held more or less constant as supply variability with respect j to any given individual would essentially be removed. ! These data, combined with those regarding the various recreation travel distances (or prices) as determined above would thus provide a measure of potential demand at u . 274 a "price," or location by a travel time zone, that people would be willing to pay (or travel) for recreation, and thus help identify the demand curve. 3. Derives a supply function which is effectively "shifted" by the recreation planner to exactly meet the amount of potential recreation demand at the "price" (or distance) people would be willing to pay (or travel) for recreation. Provision of the proper amount of supply to exactly meet potential demand at this "price" thus estab lishes an equilibrium point between demand and supply or a "market price" for recreation. It also allows the "market" to express itself and not be manipulated or con trolled by the planner. This "shifting" of the supply curve by the public recreation planner can be accomplished independent of any shifts in the demand curve since the provision of recreation facilities is subject to the de cisions of public bodies and not to the usual market forces. Thus, construction of a model using individual cross-section data to estimate a traditional type of mar ket demand function, rather than one that is site spe cific, would provide both a measure of potential demand and a possible solution to the identification problem. 275 The special nature of the costs associated with partici pation in outdoor recreation activities and the spatial distribution of recreation facilities makes such a model feasible. This type of model will be developed in the I following sections. | | i Model Development and Data Sources j i The model will be formulated in two steps. First, | j a methodology to estimate potential demand will be devel- j oped. Then, based upon certain supply analyses, a pre dicting equation will be formulated that will be used to project the optimum combination and location of future supply requirements to meet demand. The important point to note here is that demand and supply are treated sepa rately. That is, a demand function is estimated sepa rately and includes only those variables affecting rec reation demand. Demand and supply are related only after potential demand has been quantified. With this tech nique, potential demand is not constrained or controlled by supply and is not combined with supply variables in a reduced form market clearing equation to project demand. This feature of the model helps eliminate the identifica tion problem and avoids the shortcomings of previously i 276 discussed models. It also allows the "market" to express its demand in the true economic sense. Demand i i , | In accordance with the underlying theory of the j i I model, cross-section data regarding certain socioeconomic j characteristics of the population known to influence par- j j ticipation in recreation will be used to help derive a j : i traditional market demand function. First, however, the j proportion of participation expected to occur at a park j i must be determined. ! Park-Oriented Activities.— One of the major defi ciencies of previous demand studies undertaken by or for | public agencies is that no account is taken of the propor- I i tion of participation that occurs (or conversely, that does not occur) in parks, the major type of supply compo nent provided by these public agencies. While park visi- i tation has increased dramatically in recent years, not all I outdoor recreation occurs in a park. For example, there are many occasions where activities such as driving for j pleasure, sightseeing, swimming, picnicking, etc., do not | take place at a park, and would not take place at a park even if there was a shortage of facilities for these 277 activities. This is not to deny that parks do provide facilities for these and other activities and that parks are widely used. The point here is that facilities do not have to be provided to meet the total potential demand j for all activities. To provide facilities for the total potential demand in all activities expected in the future < could lead to serious overestimates of demand and misal- ! location of natural and financial resources. This is especially true in Arizona which has a relatively low de gree of urbanization and many undeveloped areas where a variety of outdoor recreation activities may be pursued. | Based upon the SRC survey data, 11 primary park- j oriented recreational activities were selected for analysis: boating; camping; picnicking; hiking; horseback riding; nature walks; pleasure driving; sightseeing; swimming; pleasure walking; and, water skiing. These activities are considered the most representative of j those that occur in Arizona parks and for which facilities would normally be provided. For other states, additional activities could be added or some deleted. The method ology used to determine potential demand would, however, be the same. The SRC survey data were also used to estimate 278 | the proportion of participation in these 11 activities j I * 1 n expected to occur at a park.-1 -^ These proportions, which I range from 33 percent for walking for pleasure, to 90 percent for camping, are listed in Table A-l in the Appendix to this report. | i | Recreation Tastes and Preferences.— There are no I i data concerning recreation demand or the tastes and pref erences of Arizona residents regarding participation in various recreation activities. Determination of these i i i tastes and preferences, or consumer desires, are difficult j | to measure or quantify and appropriate data are costly to obtain. However, some of the socioeconomic character- : i ! j istics of a person have been found influential in deter- j ! ! mining the types and amounts of recreation desired. For ; ; ! ! instance, as found by the ORRRC studies, certain charac- j I , I teristics such as income, age, sex, education and others j I were found to be important influences on the type and j quantity of recreational services demanded by an indi- j vidual.^3 Sociological and psychological studies have i . . . . . . . . . . . ! i 9 Eva Mueller and Gerald Gurin, Participation in i Outdoor Recreation; Factors Affecting Demand Among Amer ican Adults, ORRRC Study Report 20 (Washington, D.C.: Government Printing Office, 1962), p. 58. 130RRRC Report 26. 279 have also indicated that various components of a person's ! socioeconomic status are positively correlated with his ! tastes and actions.^ Socioeconomic status as a single characteristic i associated with each person is, however, a complex vari able. Since groups of persons who have similar socio economic characteristics tend to have a common outlook on life, socioeconomic variables have some predictive value with respect to the actions of the consumer. Since de mand is in part a function of taste and preferences, and they in turn relate to socioeconomic variables, demand may be considered a partial function of these variables, i The most complete and extensive cross-section ! data on socioeconomic characteristics of recreationists i were gathered during the Survey Research Center Survey,*5 i 1 £ % the National Recreation Survey, ° and the Bureau of Out- j --------------------- --- a . Mead, "The Patterns of Leisure in Contem porary American Culture," Annals of American Academy of Political and Social Sciences, Vol. 313, No. 14 (September 1957), pp. 11-15; R. Havinghurst and K. Feigenbaum, "Lei- | sure and Life Styles,1 1 American Journal of Sociology. ; Vol. LXIV, No. 4 (January 1959), pp. 397-411; and R. M. Williams, "Individual and Group Values," Annals of Amer- | ican Academy of Political and Social Sciences, Vol. 371 | (May 1967), pp. 20-37. I l^Mueller and Gurin, op. cit. 16ORRRC Report 19. 280 door Recreation Survey of Outdoor Recreation Activities.-*-7 These surveys were based on large samples of the popula- | tion and provide considerable information on the prefer- i I ences and behavior of people with respect to outdoor I recreation. In addition, extensive cross-sectional multi variate analyses were made by the ORRRC^® of the SRC and NRS survey data which revealed a high degree of correla tion between certain socioeconomic characteristics of the population and participation in various outdoor recreation activities. Further, through multivariate analysis, the separate and composite effects of changes in socioeconomic i status on participation in each of these activities were 1 estimated and projections of annual and seasonal partic- | ipation rates in selected activities, for each of the | four census regions in the United States, were made to the year 2000. These data are particularly useful for purposes of determining and evaluating the underlying factors which shape and structure demand for outdoor rec- | reation. Consequently, the participation rates as devel- i I oped by the ORRRC will serve as the point of departure to j measure potential demand for recreation in Arizona. ■ * - 7U.S. Department of the Interior, Bureau of Out door Recreation, The 1965 Survey of Outdoor Recreation Activities (Washington, D.C.: Government Printing Office, 1967). ORRRC Report 26. I 281 Selection of this data source is consistent with the intent of this model since: i j 1. These data are readily available and inexpen- j ! sive to obtain? | j j j 2. The participation rates were derived from I i i cross-section data obtained from large samples of the | ! population in wide geographic regions? and, 3. These are the only meaningful data available regarding the various propensities of people to partici- I pate in outdoor recreation. j The manner in which these participation rates will| be used, as well as their limitations, will be noted j below. j | ! i Participation Rates.— A participation rate, as j ! defined by the ORRRC, is the number of separate days on I which a person 12 years of age and over participates in | a given recreation activity in a given year.19 The rate is in per capita form and was classified by the ORRRC according to various socioeconomic and demographic char acteristics of the population such as family income, age, ! sex, occupation, place of residence, census region, etc. | l^ORRRC Report 19, p. 4. 282 In this analysis, the 1960 NRS annual participa tion rates for the Western Census Region (which includes Arizona), by income category, for each of the 11 recrea- i ! tion activities considered in this study, were utilized as the data baseline. Annual participation rates (for the entire year as opposed to rates for each of the four seasons of the year) were used since the mild climate and abundance of natural recreation resources in Arizona per- i mit year-around participation in the activities considered ! in this study. The participation rates for the Western Census Region were used since Arizona is located within i this region. The basis for this assumption is that in 1960, per capita income in Arizona ($2,037) was comparable ! to that for the Western Region ($2,191). Further, with | Arizona's favorable climate and abundance of natural rec- i reation resources, it is logical to assume that the par ticipation rates for its residents would be in line with ! those for the Western Region and any income constraint I would be offset. Consequently, no adjustment to the par- | ticipation rates was considered necessary for the purposes | | of this analysis. Annual participation rates by income category were used since it was found by the SRC, ORRRC and BOR 283 i surveys that the largest variations in participation were ! nearly always found within the income distribution. Fur ther, income was found to be a major explanatory variable j | affecting park visits. In addition, income was found to be the best single socioeconomic variable for projecting recreation demand and would take into the composite effect of the other five most important socioeconomic variables affecting demand — education, occupation, residence, age-sex and leisure time. Consequently, arrangements of the participation rates by income category enables the projection of participation rates by these categories. The 1960 participation rates by income category for the 11 recreation activities were then projected to i the years 1970, 1980 and 1990. The starting point for j these projections was the ORRRC forecasts of future par- j ticipation rates, by Census Region, to the years 1976 and j 2000.2® The ORRRC's projections were based upon the asso- ! j ciations observed in 1960 between socioeconomic character- j istics of the population and their rates of participation j | in outdoor recreation. These observed associations, as j i well as the availability of alternate uses of a recrea- | ^Outdoor Recreation Resources Review Commission, Outdoor Recreation for America (Washington, D.C.: Govern- ment Printing Office, 1962), p. 221. r 284 tionist's leisure time and money, were assumed to con tinue in the future.21 Forecasts of participation rates i were made to the target years based upon expected changes ! ; in the factors shown importantly associated with varia tions in rates of participation, i.e., family income, ! place of residence, age, sex and leisure t i m e .22 However, as shown by the BOR 1965 Survey of Out door Recreation Activities, participation in outdoor I recreation was increasing at a faster rate than antici- | pated. New projections to the years 1980 and 2000 were made by the BOR in 1967 based upon the same variables used by the ORRRC in 1962.^ The projections used in this model are based upon the 1960 ORRRC participation I rates and the trends of the latest BOR projections. ! I These projections, by income category, are shown in i Tables A-2 through A-12 in the Appendix. ! Participation Rates and Potential Demand.— It is I ' ' important to note that the 1960 and 1965 participation 1 _____— | I 21qrrrc Report 26, pp. 2-3. ^ Ibid. , p. 27. 23U. S. Department of the Interior, Bureau of Out door Recreation, Outdoor Recreation Trends (Washington, D.C.: Government Printing Office, 1967), pp. 20-21. I 285 rates developed by the ORRRC and BOR were based on per sonal interview surveys in which the respondent was asked what outdoor recreation activities he had participated in during the three months preceding the interview. The re- | suits of this approach are clearly of an ex post nature because they measure consumption and not potential demand. The ORRRC and BOR projections, however, did consider the correlations between past participation and key socio- j | economic variables. Further, the projections were based | upon expected change in these variables as well as future levels of leisure time. By considering changes in these I i factors over a long period of time, the projected par ticipation rates are reflective of the propensity to ! participate in outdoor recreation, thus minimizing the j supply constraints implied by past consumption. This is i ! especially true since the projected participation rates j included the effect of increasing amounts of leisure time | which help free the future rate of participation from past consumption patterns. The relationship implied by the future levels of participation in outdoor recreation I | can be thought of, in the abstract, as having an inter cept in the ex post sense (for 1960) and a slope in an ex ante sense (for future years). Therefore, the use of 286 projected participation rates may be considered a reason able approximation to a measure of potential demand for outdoor recreation planning. Aside from the theoretical considerations men tioned above, another way in which the participation rates imply potential demand is that they are computed for each activity in the Western Census Region by income category. This approach allows the consumption or par ticipation patterns of the typical recreationist in the west, in a particular income category, to be used as a norm against which to measure the participation patterns of a typical Arizona recreationist, in the same category. Thus, if there was a deficiency in the supply of recrea tion opportunities in Arizona relative to the west that would have constrained participation or resulted in a lower participation rate, use of the western region par ticipation would at least insure that potential demand for the Arizona recreationist was not understated as com pared with that of the typical recreationist in the west in the same income category. Moreover, even if the western region participation rates were lower than those in other regions, they could be normalized against those of the region having the 287 highest participation rates to arrive at a better repre sentation of potential demand. This procedure, as will j be recalled from Chapter IV of this report, was under- ! taken by the ORRRC in one of their alternative series of projections. The ORRRC calculated an "opportunity" fac tor for each region across the nation and normalized it against the "best" region. The purpose here was to help correct for the possible under statement of participation ^ in some regions where there were insufficient facilities available for participation and thus lower participation rates. This procedure was not required for this study, as the average opportunity indexes (which considers quan- : i tity and quality of physical resources, the accessibility j and development of these resources, and the relative de- j gree of use made of them) were higher in the west than | j in any other region for all of the activities considered j i in this analysis, except swimming and boating, which were j 1 highest in the Northeast Census Region.24 &n adjustment J ! in participation rates for these two activities was not | i considered appropriate for this study since both boating and swimming were found to be highly correlated with 24ORRRC Report 26, p. 46. 288 i income, and incomes in the Western Region (including j Arizona) were lower than those in the Northeast Region. j I Thus, the participation rates used in this study j I are considered as reasonable an approximation to the rec- | I reation desires of Arizona residents as can be obtained, given the limited availability of data. These rates will ' therefore be used to determine potential demand. Population.— In the model, only the demand from ! the two major population centers in Arizona — the Phoenix and Tucson Standard Metropolitan Statistical Areas (SMSAs), Maricopa and Pima Counties, respectively — will be considered. One reason for using these two areas rather than the state as a whole is that the ORRRC and i BOR national recreation surveys showed that the greatest j demand for outdoor recreation is generated within and j immediately adjacent to urban areas where the population is geographically concentrated. A lack of facilities near home leads to travel to find opportunities to sat isfy demand elsewhere. However, the number of persons who will travel to participate in any recreation activity j declines rapidly as the travel distance increases. Locating recreation facilities near the home of the poten tial participants will increase the use of such facilities 289 as compared with the use of identical recreation facili ties located at a distance from the place of residence. Therefore, in order to allocate facilities in relation to the potential demand, emphasis must be placed on the development of recreation opportunities in or near densely populated areas. In Arizona, the greatest bulk of potential demand is currently located, and in the future will be located, in the Phoenix and Tucson SMSAs. These areas are the only two SMSAs in the state and their residents accounted for approximately 74.4 percent of the total Arizona population in 1970 and will account for an even greater proportion, 84.2 percent, by the year 1990 (see Appendix Table A-13) . Thus, to establish the outdoor recreation demand for the state is, in effect, to establish the demand for these two SMSAs. Consequently, only the demand from the popula tion in these two SMSAs will be considered for exposition purposes in the model. However, the model could easily be expanded to include other planning areas if necessary. It should be also noted that demand will be esti mated for the entire population in both the Phoenix and Tucson SMSAs. Total population (rather than only those persons 12 years of age and over) will be used to elimi- 290 nate the possible under statement of demand. The ration ale here was that entire families, as noted in Chapter IV of this report, typically engage in outdoor recreation activities and serious errors could result in estimating the magnitude of potential demand if the portion of pop ulation under 12 years old was not considered. In addition, only the recreation demand generated by Arizona residents is considered. Nonresidents or out- of-state visitors are excluded as the demand for outdoor recreation facilities generated by these visitors was con sidered negligible. For example, as determined in the surveys made in 1967 for the Arizona Outdoor Recreation Plan, most visitors to the state traveled by automobile and stayed only a few days or were just passing through. This finding, although contrary to popular opinion, is certainly logical when it is considered that Arizona is in excess of 600 miles in width. Thus, depending upon where the motorist started his journey, it would take at least one to two days driving time just to cross the state. Further, the above noted studies determined that the primary purposes of the visitor's trip were not oriented toward outdoor recreation but were more directed toward visiting scenic attractions (e.g., the Grand 291 Canyon, Petrified Forest and various Indian reservations, etc.).25 It may be noted that a similar finding was made in the studies undertaken for the California Outdoor Rec reation Plan. Here, it was estimated that only 5 percent of the demand for all outdoor recreation activities orig inated from out of state.25 Consequently, the portion of recreation demand emanating from nonresidents is assumed negligible and hence omitted from future consideration. This omission is not a serious limitation of the model in that it could easily be expanded to include demand from nonresidents in those states where the impact of these visitors on outdoor recreation facilities is of greater consequence. Income.— It was found by the ORRRC that family income alone, as a single variable, could be used to determine and project the composite effect of the five OC ^Arizona Outdoor Recreation Coordinating Commis sion, A Plan for Outdoor Recreation in Arizona (Phoenix, Arizona, June 1967), Section 3, pp. 71-94. Stanford Research Institute, California Recrea tion and Parks Study; An Element of the State Resources Development Program, prepared for the State of California Department of Parks and Recreation (South Pasadena: Stan ford Research Institute, December 1965), Part i, pp. 113- r i 292 i other major socioeconomic factors affecting recreation j j demand — education, occupation, residence, age-sex and j leisure time. Therefore, the income variable may be ap- j i plied to the participation rates, by income category, to j I | help measure potential demand. Income data for the | Phoenix and Tucson metropolitan area, projected to the i I year 1990, are shown in Appendix Table A-14. ; Recreation Travel Patterns and Time Zones.— As indicated earlier, information concerning the various dis tances that people would be willing to travel for recrea- , tion would establish an implicit "price" for recreation and help identify the demand function. Fortunately, such j I cross-section data are available for Arizona. A study j | I j conducted in 1965 by the BOR in conjunction with Maricopa ! and Pima Counties and the cities of Phoenix and Tucson I determined the recreation travel patterns for 23 recrea tional activities engaged in by the residents of the Phoenix and Tucson metropolitan areas. This study indi- : i I cated that 75 percent of all outdoor recreation partici- j i pation by the residents of these two major population centers occurred within 100 miles from these centers, | 17 percent within 200 miles and 8 percent beyond 200 miles.^7 27Arizona Outdoor Recreation Coordinating Com- 293 Based on this information, isochronic lines were developed for two-hour travel time distance intervals or travel-time zones from the population center of gravity ! of the Phoenix and Tucson SMSAs. These travel-time zones | are shown in Figure VI-1. It should be noted that these zones are not cumulative? each one originates from the center of gravity of the two large population concentra tions combined. Furthermore, these zones represent the actual driving times out of Phoenix or Tucson based upon the state's existing road systems as of 1965. This tech- j nique takes into account such factors as allowable driving! speed, road capacity and traffic congestion which will affect travel time and influence demand in a particular i I i I zone. i | In developing these time zones, it was estimated ! that the zero to two-hour travel zone (a maximum round- trip travel time of four hours) was equivalent to a 100 j mile driving distance (one way) zone wherein about 75 j i percent of the total participation in the various outdoor j i recreation activities engaged in by the residents of the ! Phoenix and Tucson SMSAs occurs. The two to four hours travel-time zone (four to eight hours round trip) was mission, Initial Outdoor Recreation Plan — State of Ari zona (Phoenix, September 1965), pp. 66-68. I I UTAH STATE OF ARIZONA ^ % ^JPHOENIXV • c e n t e r 7 O F ( ^ gravity T U C S O N 0 - 2 HR T R A V E L TIME Z O N E 2 - 4 HR T R A V E L T I M E Z O N E Figure VI-1 — Travel-time zones from the population center of gravity of the Phoenix- Tucson metropolitan areas 295 considered equivalent to a 200 mile driving distance (one way) zone that accommodated about 17 percent of the total outdoor recreation activity. It is within the first four hours from home that 92 percent of all the recreation participation takes place. Any recreation trips over four hours (eight hours minimum round trip) were considered synonomous with the over 200 mile driving distance (one way) zone. Based upon these data, the percent of participa tion in the 11 park-oriented activities, by travel-time zone from the population center of gravity of the Phoenix and Tucson SMSAs, was computed as shown in Table VI-1. It should be noted that Arizona presently has a relatively lower degree of urbanization and traffic con gestion than that found in other states. As the highway networks are improved, these zones will expand. In the long run, however, as the Phoenix and Tucson areas and the remainder of the state become more highly urbanized and the frustration of penetrating traffic barriers and congestion inhibits recreation travel, these travel-time zones may be reduced and the demand for outdoor recrea tion facilities will be manifest even closer to the core of population concentrations. 296 TABLE VI-1 RECREATION ACTIVITY BY TRAVEL TIME ZONES Percent of Participation ________ by Zone^________ Activity 0-2 hrs. 2-4 hrs. Over 4 Boating 60 30 10 Camping 60 30 10 Driving for Pleasure 75 20 5 Hiking 60 25 15 Horseback Riding 85 10 5 Nature Walks 75 20 5 Picnicking 85 10 5 Sightseeing 80 15 5 Swimming 80 15 5 Walking for Pleasure 85 10 5 Water Skiing 70 20 10 aTravel time zone from the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. Source: Developed from data contained in: Arizona Outdoor Recreation Coordinating Commission, Initial Out door Recreation Plan, State of Arizona (Phoenix, Arizona, September, 1965), pp. 67-68. 297 Projections of Potential Demand.— The above noted | I variables represent the complete array of recreation de- j ! 1 mand determinants hypothesized in this study for an j I 1 Arizona resident, i.e., ■ I V D ^ = f (tastes and preferences, socioeconomic j characteristics, distance) j The participation rates, by income category, rep- ! | resent the differences in tastes and preferences among ' recreationists according to differences in their incomes as well as the amounts or quantities of a recreation ac tivity they desire. Family income represents a composite ; of the most important socioeconomic characteristics in- j i i fluencing recreation demand. The various distances rec- j I reationists are willing to travel for various activities | and the proportion of participation in each distance or j ; i travel zone help establish the amounts of activities de- ! ! sired at various implicit prices for recreation. These j : i factors thus represent a recreationist's potential demand ! i or the quantities demanded at various implicit prices, i.e., the potential demand of an individual in a given income class in a given period of time for a particular recreational activity in a particular travel or distance zone. i 298 Market demand of a given population in a given time period may thus be determined by summing each of the individual's demands. This may be done through the following formulation: 8 11 3 Vn =VA.. (vi)= 2 2 2 [(Pm)(Rjm)(°j)] T±j D °mji ra = 1 j = 1 i = 1 J J J where: VD = potential market demand in a given time period for a given population expressed in number of visits V 1 = potential demand in a given time period •^ji for a given population in income level m for recreation activity j in travel zone i, expressed in visitor days Vj = constant to convert visitor days in activity j to the number of visits Rjm = participation rate for activity j by income level m Pm = population in income level m Oj = constant representing the proportion of activity j that is park oriented Tj ^ = constant representing the proportion of activity j occurring in travel zone i 299 m = income level which varies from 1 to 8 where: 1 = Under $1,500 2 = $ 1,500-$ 2,999 3 = $ 3,000-$ 4,499 4 = $ 4,500-$ 5,999 5 = $ 6,000-$ 7,999 6 = $ 8,000-$ 9,999 7 = $10,000-$14,999 8 = $15,000 and over j = a particular recreation activity which varies from 1 to 11 and includes: 1 = Boating 2 = Camping 3 = Picnicking 4 = Hiking 5 = Horseback Riding 6 = Nature Walks 7 = Pleasure Driving 8 = Sightseeing 9 = Swimming 10 = Pleasure Walking 11 = Water Skiing 300 i = a travel-time zone from the center of gravity of the Phoenix-Tucson metropolitan areas which varies from 1 to 3 where: 1 = the 0-2 hours travel-time zone 2 = the 2-4 hours travel-time zone 3 = the 4 hours and over travel-time zone Prom the above, it may be noted that a constant term, Oj, was used to account for the proportion of each activity expected to occur at a park. The individual values for these proportions were indicated previously. Also, that a constant term, vj, was used to convert vis itor days (which result from using participation rates) to the number of visits. Since there are no data con cerning the length of stay in a park for each of the ac tivities, the value of vj was assumed to be unity. When such data become available, this constant may be deter mined by the recreation planner. Additional constants may be incorporated to account for other effects, such as the proportion of annual demand expected to occur during the peak recreation season and others of importance to the recreation planner. The basic methodology would still, however, be the same. I — ------------------------------------------------------------------- 1 j ! 301 | The market demand "schedules" for the 11 park- | oriented recreational activities, by travel zone, for i | I the years 1960-1990 (by 10 year increments) are shown in j ; j Tables VI-2 through VI-5. Aggregate potential demand j ! over all time zones for the 1960-1990 time period is i 1 depicted in Figure VI-2. Figure VI-3 illustrates the market demand curve for 1970. Here, it may be noted that for purposes of plotting the curve, the midpoint values for the 0-2 hour | zone and the 2-4 hour zone, unity and three, respectively, were used; for the over 4 hour zone, a distance value of I 5.25 hours was specified. It may also be observed that the negative slope j i i | ! of the market demand curve conforms to that expected for j ! a "normal" good. For the usual commodity, the slope of j the demand curve reflects the degree of price elasticity I ; ] of demand or the responsiveness of the quantity taken of | a commodity to changes in its price. The degree of elas- i | ticity of a curve is usually measured between two adjacent i | points along it in the range that is of interest to the economist or in the "relevant range."^8 When a unit j 28 E. J. McCarthy, Basic Marketing: A Managerial Approach (3rd ed.y Homewood, 111., Richard D. Irwin, Inc., 1968), p. 209. 302 TABLE VI-2 TOTAL POTENTIAL DEMAND FOR ELEVEN OUTDOOR RECREATION ACTIVITIES IN VISITOR DAYS 1960 Travel Time Zonea Activity 0-2 hrs. 2-4 hrs. Over 4 hrs. Boating 466,230 233,100 77,770 Camping 1,017,600 508,690 169,630 Picnicking 2,166,430 254,990 127,360 Hiking 269,120 112,240 67,190 Horseback Riding 490,170 157,540 28,820 Nature Walks 1,353,000 360,890 89,940 Pleasure Driving 7,381,570 1,968,470 492,060 Sightseeing 2,835,180 531,710 177,080 Swimming 2,578,790 483,450 161,350 Pleasure Walking 4,526,190 532,440 266,210 Water Skiing 321,380 91,690 46,000 Total 23,405,660 5,135,210 1,703,410 aFrom the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. 303 TABLE VI-3 TOTAL POTENTIAL DEMAND FOR ELEVEN OUTDOOR RECREATION ACTIVITIES IN VISITOR DAYS 1970 Travel Time Zonea Activity 0-2 hrs. 2-4 hrs. Over 4 hrs. Boating 1,146,160 573,020 190,750 Camping 2,930,020 1,465,070 488,460 Picnicking 4,601,380 541,360 270,500 Hiking 670,560 279,450 167,520 Horseback Riding 951,000 111,930 55,790 Nature Walks 2,264,810 603,870 150,920 Pleasure Driving 11,257,320 3,001,890 750,550 Sightseeing 6,762,590 1,268,080 422,560 Swimming 6,351,690 1,190,770 396,720 Pleasure Walking 11,319,940 1,332,060 665,880 Water Skiing 980.310 280.070 139,890 Total 49,235,780 10,647,570 3,699,540 aFrom the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. 304 TABLE VI-4 TOTAL POTENTIAL DEMAND FOR ELEVEN OUTDOOR RECREATION ACTIVITIES IN VISITOR DAYS 1980 Travel Time Zone3 Over Activity 0-2 hrs. 2-4 hrs. 4 hrs. Boating 2,239,860 1,119,910 373,650 Camping 5,413,700 2,706,640 902,190 Picnicking 6,839,820 804,770 402,540 Hiking 1,160,900 483,750 290,230 Horseback Riding 1,488,610 174,980 87,760 Nature Walks 3,499,290 932,920 233,910 Pleasure Driving 17,481,290 4,661,880 1,165,140 Sightseeing 11,037,860 2,069,060 689,650 Swimming 11,500,780 2,156,320 718,550 Pleasure Walking 17,268,860 2,031,420 1,015,810 Water Skiing 2,236.790 639,190 319,120 Total 80,167,760 17,780,840 6,198,550 aFrom the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. 305 TABLE VI-5 TOTAL POTENTIAL DEMAND FOR ELEVEN OUTDOOR RECREATION ACTIVITIES IN VISITOR DAYS 1990 Travel Time Zonea Over Activity 0-2 hrs. 2-4 hrs. 4 hrs. Boating 4,025,940 2,013,000 670,670 Camping 8,995,980 4,497,960 1,499,330 Picnicking 9,884,140 1,162,160 581,520 Hiking 1,901,050 792,050 474,720 Horseback Riding 2,434,790 286,520 142,790 Nature Walks 5,401,800 1,440,370 360,620 Pleasure Driving 25,761,660 6,869,510 1,717,550 Sightseeing 17,635,270 3,306,240 1,102,740 Swimming 19,586,120 3,672,550 1,224,630 Pleasure Walking 27,278,850 3,209,830 1,603,860 Water Skiing 4.296.120 1,227,280 614,080 Total 127,201,720 28,477,470 9,992,510 aFrom the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. 306 V I S I T S ( IN M I L L I O N S ) 200 1 5 0 100 5 0 I 9 6 0 1 9 7 0 1 9 8 0 1 9 9 0 Figure VI-2 — Aggregate potential demand: 1960-1990 307 T R A V E L Z O N E ( H O U R S ) 6 5 4 3 2 0 20 3 0 4 0 V I S I T S ( M I L L I O N S ) 5 0 Figure VI-3 — Market demand curve for eleven park- oriented recreation activities: 1970 308 I | i change in price produces a proportionately smaller change in quantity taken in the price range relevant for the j i product, demand is relatively inelastic or unresponsive j to price changes. When a unit change in price in a par- ! ticular price range produces a proportionately larger response in the quantity taken, demand is relatively elastic. Unitary relative elasticity of demand is found where a given change produces an equal, proportionate change in quantity taken in the price range under consid eration.^ With respect to the market demand curve for out door recreation shown in Figure VI-3, the relatively elastic portion generally lies within the travel time or j "price" range from 0-3 hours. Within this range, quantity| demanded changes more than proportionately to changes in | | travel time. Also, as "price" or travel time decreases, | the responsiveness to quantity taken becomes even greater.i I Beyond three hours travel time, demand is relatively in- ! elastic and increases in price have less than proportion ate changes in quantity demanded. This is as expected since recreation demand is inversely related to price. 29Kenneth E. Boulding, Economic Analysis. Vol. I; Microeconomics (4th ed. y New York: Harper & Row, Pub lishers, 1966), pp. 181-183. 309 For the recreation planner, however, the "relevant range" is probably somewhat narrower, say within the 0-2 hour | time zone. It is within this zone that unit changes in : travel time, say from two hours to one hour, produce pro- J I portionately greater changes in quantity demanded as com- { I pared with distance changes, say from three hours to two hours. Further, it is within the 0-2 hour range that potential demand is greatest and where the bulk of facil ities should be supplied. The above traditional type of market demand curve is considered a reasonable approximation to a true demand i ! curve, given the limitations of the existing data. While i ! a structural demand curve cannot be formulated, the above demand function is useful for planning purposes. As | noted by Kalter and Gosse: If projections of use are the purpose of a study and if the researcher is willing to assume that I use is highly correlated with some observable ! variables, then the projections may not be sig- | nificantly distorted even though the structural equations are not identifiable.30 | In addition, the relative values computed for j | potential demand agree with the consensus as to the direc- | tion and location of future demands for outdoor recreation I j in Arizona. It must be emphasized again, however, that it 30 Kalter and Gosse, op, cit., p. 44. is not the magnitude of the values per se that is impor tant. It is the methodology used in determining these values and the variables considered that are of paramount importance and the primary focus of this study. Moreover, while there are limitations to the fore-j going methodology, it is felt that it will likely lead to more accurate results than any alternative method and will be useful for recreation planning. Commenting on this point, after having suggested their methodology for estimating recreation demand, the noted resource econo- | | mists, Clawson and Knetsch, observed: At this point, a non-economist might well exclaim: i "If you must so carefully further qualify your al- j ready somewhat novel approach to the estimation of j demand for outdoor recreation, is not the whole ! procedure too complicated and too arbitrary for j use in practice?" While this reaction might be j natural enough, it misses the main point. It is j not the procedure as such that needs qualification | and caution, but rather the basic recreation expe- j rience and the available data about it which are i complicated and not easily interpreted. Anyone attempting to analyze the same basic experience and same data by any less rigorous or superfici ally simple method would encounter the same prob lems of interpretation and analysis; moreover, he might easily be led astray by a less adequate method and thus reach less dependable conclu sions. 31 31 Marion Clawson and Jack L. Knetsch, Economics of Outdoor Recreation (Baltimore: Johns Hopkins Press, 1966), p. 89. 311 supply In accordance with the underlying theory of the model, it was hypothesized that supply, or the amount of facilities (or various quantities) that a public agency or recreation planner (the supplier) would be willing to place on a given market in a given period of time at all possible distances (implicit prices), ceteris paribus, is a function of distance (implicit price) and certain facility (supply) variables. In addition, that visita tion may be considered as a measure of the quantity sup plied. Further, that the amount to be supplied at that distance (implicit price) will be that amount which ex actly meets potential demand at that distance (implicit price). In other words, the public agency or recreation planner will "shift" the supply curve to meet the various distance (implicit price) — quantity demanded points along the demand curve. Hence, the market will clear where: Inasmuch as one of the main purposes of the model is to forecast the optimum supply requirements to meet potential demand, the supply function will be developed 312 as a predicting equation, with due account taken of the effect of existing supply on recreation behavior. That | is, it will be developed to forecast the optimum combina- ' I 1 ' I tion of facilities in a set of hypothetical parks to meet j potential demand, by travel-time zone from the population j | center of gravity of the Phoenix-Tucson metropolitan ! areas, to the year 1990. This combination of facilities will thus be at the equilibrium point of potential demand i i and supply where the market will clear. i Before proceeding with development of the equa- ! i tion, it is important to note the sources of the supply ! variables that are available and that will be used in ; the analysis. j | I ; j l Data Sources.— Supply data were collected for all ! I | ! major public parks (a total of 45) in Arizona. Included j ! i I ; | was information concerning the characteristics of each j ! park in terms of: total land and water acreage; recrea- I | tion facilities (e.g., number and size of picnic areas); i recreation units (e.g., number of picnic tables); and, | visitation to the park. These data were grouped by the i | various park administering agencies in Arizona which in- i ! elude: the U.S. National Park Service, U.S. Forest Serv- I ice, Arizona State Parks Board, the Maricopa County Parks 313 and Recreation Department, and the Pima County Parks and Recreation Department. The location of each park was also determined in relation to the travel-time zones from the population center of gravity of the Phoenix- Tucson metropolitan areas. This supply inventory is shown in Tables A-15 through A-30 in the Appendix to this report. It may be noted that the data are available only for the 1960-1965 time period. Further, there were no changes in the supply inventory during that period; only the visitation changed. Moreover, there were no major changes in the road system or access to the various parks. Thus, the supply of recreation facilities and the travel time required to reach the parks are constant for the 1960-1965 period. In addition, the visitation data are for visits to sites which offer a collection of facilities for a wide variety of recreation activities; breakdowns of visitation to specific types of facilities or for par ticular types of activities are not available. It may also be noted that there are other public recreation suppliers in the state, such as the Bureau of Land Management, U.S. Fish and Wildlife Service, Arizona Game and Fish Department, the Mohave County Parks Depart- 314 ment, various Indian reservation, and various Arizona cities and towns. Supply information from these sources was so fragmentary and incomplete as to render it useless. Exclusion of these sources from the inventory is not a limitation to the analysis since: these agencies (e.g., the Bureau of Land Management, various Indian reserva tions, etc.) have extremely limited facilities available or developed for recreational use by the public; and, are far removed from the Phoenix-Tucson metropolitan area. Upon review of the supply inventory, certain other exclusions were made. For example, five of the twenty U.S. National Park Service parks were eliminated — the Grand Canyon National Monument, Grand Canyon Na tional Park, Lake Mead National Recreation Area, Navajo National Monument, and Tuzigoot National Monument. These parks were omitted from the analysis because of: lack of facilities; orientation primarily to tourists with negli gible use by Arizona residents; location beyond 6 hours driving time from the Phoenix-Tucson population center of gravity; or, some combination of the above. In addition, five of the nine Arizona State Parks were eliminated: Havasu Lake, Buckskin Mountain, Picacho Peak, Painted Rocks, and, Jerome. While these facilities are used by 315 Arizona residents, they were not activated until the latter part of 1965 with the result that minimal visita tion occurred. A summary of the park characteristics used in ; this analysis, located by one-hour travel time or distance zones from the population center of gravity of the Phoenix-Tucson metropolitan areas, is contained in Tables VI-6 and VI-7. j While there are certain limitations to the data, j they are the best that are currently available. While a i detailed, complete supply inventory, with visits broken down to show visitation to various types of facilities or i j participation in various kinds of activities, would be j desirable, the time and costs involved to collect such j information for use in this study would have been pro hibitive. This is evident by the fact that the various supply agencies themselves do not maintain such informa tion, a situation which prevails in almost every state. However, it must be emphasized that these data limitations do not affect the validity of the technique that will be used to derive the predicting equation. Every possible type of recreation area or facility typ ically provided in a park is represented in the data; TABLE VI-6 SUMMARY OF SUPPLY VARIABLES Supply Variable Travel Time Zonea 1 Hour 2 Hours 3 Hours 4 Hours 5 Hours 6 Hours Total Total Recreation Land Acreage 93,258 2,544,363 2,307,162 1,711,378 3,567,581 3,942,670 14,166,412 Total Recreation Water Acreage 1,145 21,215 3,000 8,680 1,695 88,640 124,375 Acreage by BOR Class BOR I 404 14,461 1,351 4,524 9,393 2,803 32,936 BOR II 666 75,773 12,071 35,313 23,297 1,109,063 1,256,183 BOR III 30,585 1,961,375 2,245,917 1,421,966 3,266,846 2,096,892 11,023,581 BOR IV 10,228 12,420 14,228 52,702 90,818 516,871 697,267 BOR V 51,093 280,853 36,170 122,350 174,866 261,534 926,866 BOR VI 283 2,596 423 876 4,056 44,148 52,382 Swimming Beaches Number 1 1 2 1 1 6 Acres 4 3 3 2 1 3 16 Picnic Areas Number 10 32 21 37 51 80 231 Acres 177 144 53 242 305 81 1,002 Tables 229 713 606 669 1,250 930 4,397 aFrom the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. u> M <n TABLE VI-6 — Continued Supply Travel Time Zonea Variable 1 Hour 2 Hours 3 Hours 4 Hours 5 Hours 6 Hours Total Boat Access Number 1 8 4 5 7 11 36 Acres 3 17 4 9 9 38 80 Parking Spaces 15 794 189 302 103 2,346 3,749 Tent Camps Number 5 24 17 14 31 37 128 Acres 15 208 113 96 332 715 1,479 Tent Spaces 47 349 367 339 525 1,442 3,069 Trailer Camps Number 4 3 12 3 8 22 52 Acres 15 8 77 20 52 253 425 Trailer Spaces 54 48 216 82 94 952 1,446 Marinas Number 3 1 1 5 10 Acres 3 1 1 14 19 Slips and Moorings 105 36 29 898 1,068 Horse Trails (miles) 30 728 601 340 234 277 2,210 Foot Trails (miles) 65 80 12 79 213 385 834 Vista Points Number 4 3 12 8 11 38 76 Acres 9 8 11 11 14 33 86 Sources: Appendix Tables A-15, A-16, A-17, A-19, A-20, A-21, A-23, Ar-24, A-25, A-27, iU28, and A-29. U) -4 TABLE VI-7 SUMMARY OP VISITATION TO ARIZONA RECREATION AREAS Number of Visits Travel Time ; Zonea 1 Hour 2 Hours 3 Hours 4 Hours 5 Hours 6 Hours Total 1960 285,400 2,002,800 818,200 1,229,100 1,946,500 4,184,300 10,466,300 1961 318,100 2,063,800 933,500 1,382,600 1,921,700 4,230,400 10,850,100 1962 330,000 2,098,800 980,000 1,479,200 2,059,800 4,944,000 11,891,800 1963 373,400 2,247,400 1,505,100 1,658,300 2,358,900 5,995,700 14,138,800 1964 459,900 2,517,500 1,821,300 1,847,800 2,687,000 6,741,800 16,075,300 1965 508,500 2,401,100 1,959,600 1,990,700 2,741,600 7,242,400 16,843,900 aFrom the population center of gravity of the Phoenix and Tucson Standard Metropolitan Statistical Areas. Sources; Appendix Tables A-18, A-22, A-26, and A-30. 00 i - * oo 319 additional data would only increase the magnitude of each i | variable and not the number of variables. As remarked J earlier, it is the technique or methodology used and the ! type of variable considered that are of primary impor tance, not the magnitude of the variables themselves. i j A brief description of the supply variables that were measured follows. Description of Variables Measured.— Total recrea tion land acreage and total water surface acreage were measured for each park administered by the selected pub lic agencies. The land area variable was included to show whether the size of a park is an important considera tion in visitation to that park and, if so, what relative 1 ! | size park attracts the most visitation. Similarly, the ! | I ! amount of water acreage presents a certain attraction, j ! i i if available, and depending upon the amount, can con- | i | strain the water-oriented activities that can occur at the park. | The land and water acreages were further cate gorized for each park according to the six Bureau of Out- ! I O O j door Recreation classifications. ^ These classes are 19 U.S. Department of the Interior, Bureau of Out door Recreation, Outdoor Recreation Grants-in-Aid Manual (Washington, D.C.: Government Printing Office, 1965), Part 630, pp. 12-20. 320 briefly described below: Class I — High Density Recreation Areas Areas intensively developed and managed for mass use. Physical requirements are topography and soil type that are adaptable to intensive recreation use. There are no specific size criteria. Their location is usually within or near urbanized areas accessible by automobile. Activities are intensive day or week end type, such as picnicking and water sports. De velopment is intensive, and includes roads, parking areas, picnic areas, restaurants, bathing beaches, bathhouses, marinas, artificial lakes, playing fields, golf courses and sanitary facilities. Class II — General Outdoor Recreation Areas Areas subject to substantial development for a wide variety of specific recreation uses. Physical requirements are varied topography, interesting flora and fauna in a natural or man-made setting, adaptable to wider range of recreation opportunities than Class I. There are no size criteria and the areas vary widely in size. They are usually lo cated 20-40 miles from urban population centers 321 and easily accessible by highway. Activities are extensive day, weekend, and vacation use types, such as camping, picnicking, fishing, water sports, and nature walks. Development may have nodes of Class I types of activity and there may also be small portions of Class III areas. Class II also includes parking areas, picnic areas, camp grounds, bathing beaches, marinas, streams, and artificial lakes, cabins, stores, museums, playfields, ski tows, marinas, lodges, riding stables and hotels. Class III — Natural Environmental Areas Encompasses various types of areas that are suitable for recreation in a natural environment, usually in combination with other uses. Physical requirements are varied and include interesting land forms, lakes, streams, flora and fauna in natural settings. There are no size criteria and location depends on existing natural features. Activities are the extensive day, weekend and vacation type used which depend on the quality of a natural environment for sightseeing, hiking, cycling, nature study, picnicking, camping, boating, fishing and hunting. Concentrated use 322 is discouraged and the landscape retained in as natural a condition as possible. Development may include trails, simple picnic and camping opportunities. The focus is on the environment and the resource instead of the user. Class IV — Outstanding Natural Areas These are areas of outstanding scenic splendor, natural wonder, or scientific importance. Physical requirements are outstanding natural features that merit special protection and management to insure their preservation. There are no size criteria and location is dictated by where such features are found. Activities are the enjoyment and study of the natural attractions to preserve the quality of the natural features. Food and lodging facil ities which could impair scenic or scientific values are excluded. Development is limited to the minimum required for public enjoyment, health and safety. Access roads are located on the edge of the area and the visitor encouraged to travel on foot. Facilities other than trails and sanitary facilities are outside the areas. Improvements are held to the minimum required for public safety and 323 protection of the resource and are subordinate to the natural setting. Class V — Primitive Areas This category includes undisturbed roadless areas characterized by natural, wild conditions including "wilderness areas." Physical require ments are extensive natural, wild and undeveloped areas and settings removed from civilization or inhabitied areas. The natural environment of these areas must be undisturbed by commercial utilization and the areas accessible only by foot. There are no specific size criteria. Location is usually remote from population centers. Activ ities include camping and "roughing it" without mechanized transportation and permanent shelter or other conveniences. Development is limited to trails; no public roads, permanent recreation facilities or mechanized equipment are allowed except to control fire, insects and disease. Class VI — Historic and Cultural Sites These sites are of major historic or cultural significance, either local, regional or national. 324 Physical requirements are based on the signifi cant history, tradition or cultural heritage of | the Nation, State or local area which merits ! I preservation or restoration. There are no size j 1 criteria and location of the features establishes I ! : i location of the site. Activities are sightseeing, j j appreciation, enjoyment and study of the historic or cultural features. Kinds and intensity of use j i I are limited to activities appropriate to the fea- i ture. Development is limited to necessary pres- j ervation, restoration and interpretation of the I site. Access to the area should be convenient, especially for tourists. | j As may be noted from the above, the BOR classes ! | do not relate to the size of the area or to specific | types of activities. Rather, they relate more to the in- ! tensity of development (particularly in BOR Classes I and I | II) as well as the physical character or environmental j I i aspects of the area (particularly Classes III, IV and V) . | ! i j Consequently, this classification system cannot be used j i to measure the "complementary goods" aspects of supply as attempted in the Cicchetti, Seneca and D a v i d s o n 3 3 econo- "^Cicchetti, et al. , op. cit. 325 metric model of the outdoor recreation market. The pur pose of using the BOR classes in this study is to deter mine whether the intensity of development or physical characteristics of the parks themselves are determinants of visitation. The facility variables for water-oriented activ ities measured at each park include: the number and acreage of swimming beaches; the number of boat access points, their acreage, and number of associated parking spaces; the number and acreage of marina facilities; and, the number of developed slips at marinas. The facility variables for land-based activities at each park include: the number and size (in acres) of tent camps and group camp grounds as well as the number of developed tent spaces; the number and size of, as well as developed spaces in, trailer camps; the number and acreage of pic nic areas including the number of picnic tables. Other facility variables measured for land-oriented activities were: miles of horse trails and foot trails; and the num ber and acreage of vista points. The final variables are visitation to each park for the years 1960 through 1965. Here, it must be noted that these data refer to the total number of visits and 326 not visitor days. Moreover, they do not refer to visita tion to any particular facility? rather they refer to visitation to the park itself. Predicting Equation.— Using the above data, a relation was sought that would give visitation as a func tion of distance and facilities. In developing this re lation, a series of correlation studies were first made to determine which of the 58 supply variables were most highly correlated with visitation and with each other during the 1960-1965 period. The purpose here was to "screen" or predetermine which of the wide variety of sup ply variables were the best predictors of visitation. The value of this "pre-screening" process is that it allows a predetermination of the statistical significance of the individual variables and a determination of the collinearity of certain variables before fitting them into the equation. As such, it reduces the number of errors in the resulting equation and avoids the pitfalls noted in other models that attempt to indiscriminately fit a large number of variables into the equation before determining their statistical significance. Based upon these correlation studies and a succes sive screening of variables, only 17 of the 58 supply 327 variables used in the analysis showed sufficiently high correlation to warrant further investigation. The land and water acreage by BOR class was not, contrary to the findings of the Cicchetti, et al.,34 study, found to be j statistically significant. A correlation matrix for • ! these 17 variables is shown in Table VI-8. I As may be noted, the number of picnic areas, i acres of picnic areas and number of picnic tables all | showed high correlations with visitation. However, all of these variables are essentially part of the same supply; component and highly interrelated. To eliminate collin- | earity among these picnic-related variables, only one, the number of picnic tables was selected as the most var iable for recreation planning purposes. The same situation was observed with the number of tent camp grounds, acres of tent camps and number of I tent spaces. Due to high interrelation among all of these variables, the number of tent spaces was selected as the critical camping variable. A similar situation was observed for the number of boat access points, acres of boat access, number of automobile parking spaces at boat access areas and acres •^^Cicchetti, et al. , op. cit. TABLE VI-8 CORRELATION MATRIX FOR 17 SUPPLY VARIABLES to ( 0 <D CM M <D 0 < U C M O n ) (D -H ■P A a C O 6 o •H 3 -H Q 2 & Distance 1.000 0.920 Number of Picnic Areas 1.000 Acres of Picnic Areas Number of Picnic Tables Number of Boat Access Points Acres of Boat Access Number of Boat Parking Spaces Number of Tent Camps Acres of Picnic Areas Number of Picnic Tables Number of Boat Access Points Acres of Boat Access Number of Boat Parking Spaces I Number of Tent Camps ■4.065 0.891 0.782 0.648 0.657 0.837 3.192 0.908 0.915 0.879 0.804 0.915 1.000 8.197 -0.149 -0.104 -0.371 -0.155 1.000 0.889 0.683 0.660 0.951 1.000 0.897 0.918 0.962 1.000 0.911 0.809 1.000 0.807 1.000 oj to o o 6Z£ 2! > 21 2J > 0 0 0 O • H • H H C D 0“ C D 0 C O 0 C D 0 1 l - h Hi H 0 0 f f l Hi Cd 0 Hi 0 0 Hi 0 1 W 0 1 ►d rt 0 rt »d F- 0 1 F- o •d r t > o 3 0i o 3 F- a > o H- 0 O C D O H- O 0 1 t? 3 C D C O f - 3 n iQ C O C D C O •d O' 0 1 0 H- H C D 0 1 0 1 3 0 1 o rt C D C O c n S 5 if C D ■ 1 o Hi ►d H- o 3 H- 0 & C D ( 1 1 0 1 0 H- 0 1 rt 0 1 3 O C D o o o o O o o o • • • • • • • • vO 00 00 VO vO F* vO 00 00 CO cn cn to cn cn cn 00 <1 vo F* 4* 4^ O -0 o o o o O 1 O O o • • • • • • • • vO 00 00 VO 00 to VO vO U1 -J CO CO vO cn 4 * - F* to F* to VO 4 * . cn cn 00 o o o o O i o o o • • • • • • • • U1 4* to 4» 4* cn cn cn C T i F* vO to cn cn 4* cn to F* F* F» 4* o cn 00 o o o o O l o o o • • • • • ■ • • U1 4 * > 4» 4 * . 4* cn cn <! OJ t —1 O O cn o O CO to F» cn 4- cn F» 4* o o O o o l O O O • • • • • • • • Cn cn cn 4* CO cn cn cn 4 = > vO cn 00 VO to vo cn o cn M VO cn o 00 o o o O o o i o o o • • • • • • • • vO -j -o VO vO F* 00 00 cn VO o CO 4» cn cn 00 vO CO 00 -0 00 to cn o O o o O [ o o o • • • • • • • • cn VO VO -J 4 * . to cn 4» t - * F* -J to cn vo 00 cn F» CO o 00 F* o O o o o o o l -o o O • • • • • • • • VO 00 vO VO 00 VO VO 00 cn VO 4- 00 to cn F» o CO cn 00 to 00 cn ~0 o O o o o 1 o O o • • • • • • • • VO vO 00 vO 00 to vO 00 4^ O VO 4 » - cn cn cn <1 cn CO CO vO o -o to 00 Acres of Tent Camps Number of Tent Spaces Number of Trailer Space Acres of Trailer Camps Number of Trailer Spaces Acres of Land Acres of Water 1960 Visits 1965 Visits TABLE VI-8 — Continued I - * I — * > > * a > a a > c O c D O o ri o c c S O 0 1 0 1 * 1 M 5 M § § M cn O ( D C D o ' C D o ' o C D 0 1 0 1 C D 0 1 C D C D 0 1 < < M M M H - H - 0 0 0 0 0 1 0 1 M i M i 0 M l 0 0 M i H - H - M l M i M i r t f t S f ►3 h3 0 1 0 1 0 1 0 • - 3 M • - 3 • - 3 C D r t 3 M 0 1 M C D 3 ( D Q j 0 1 H - 0 1 3 r t M H - M H - r t t - > C D M O C D M M C D M & 1 O 0 1 * d 0 1 QJ Q o 0 1 * d § C D 0 ) t i § 0 1 o 0 1 t ) C D 0 1 0 1 ( - > o o o M o o CD o - J o o M o o • • • o 0 1 O l o - o I — 1 o O l 4 * i — * o o o • • • • o CD 0 1 cn o CD - J CD o O - J 0 1 o O o O • • • • • o CD CD O l o cn C O M C O o I - * 0 1 0 0 C O ( - » o O O o o • • ■ • • • O 4 * . 4 * CO CD o - J 0 1 0 0 CO C O o C O * » • O l CO ( - • o o O O o o • • • • • • • O cn cn CO 0 0 - o O l o cn 0 0 cn CO h - * 0 0 o CD 0 0 O ( — * cn cn I - * o o o O o o O m • • • • • • * o < 1 0 0 cn 4 = - 4 * . CD CO o 0 0 0 0 cn 0 0 0 0 4 ^ cn o 4 * o 0 1 O cn Oi o o o o o O o O • • • • • • • • • o CO ^1 CO O! 0 1 CO CD o 0 1 0 0 o CO O! 0 1 CD <1 O 0 0 CO o 00 I — 1 o l-» O l Acres of Tent Camps Number of Tent Spaces Number of Trailer Space Acres of Trailer Camps Number of Trailer Spaces Acres of Land Acres of Water 1960 Visits 1965 Visits TABLE VI-8 — Continued 331 of water. For the reasons given above, the number of water acres was selected from among these boating-oriented variables. Distance was also observed to be highly cor related with visitation and although highly correlated with the number of picnic tables and tent spaces, was included as a critical variable since it was not a part of these supply measures. It should be observed that, although the total number of acres of land had a high correlation with vis itation, land was excluded from consideration. The logic for this exclusion is that the sheer magnitude of the land variable would overpower the other variables in attempting to fit the supply function. Further, the total number of acres of land is not considered an appropriate indicator of supply since the majority of this land is in large federal parks, only a small portion of which are developed for outdoor recreation. Moreover, the appropriate number of acres of land for outdoor recreation can more properly be determined by first knowing the types of facilities that must be constructed. Thus, for purposes of this analysis, the supply variables — distance, d, number of picnic tables, P, number of tent spaces, T, and number of water acres, W — 332 will be considered as the key or critical variables which best predict visitation. The screening procedure used to select these variables helps eliminate or at least reduce the problems of collinearity among related independent variables. The variables selected are also appropriate in terms of the type of supply components that must be provided by a recreation planner. For example, a picnic table is a physical unit that must be placed on a site; a tent camp space is a physical unit that must be con structed (e.g., a parking pad); and, a lake, particularly in Arizona, is typically man-made due to the relative scarcity of water in this arid region. By knowing the number of picnic tables, tent spaces, and water acres necessary, the total land area or size of park to accom modate these facilities can be determined by the planner by applying pertinent recreation standards. This will be discussed later in this chapter. Thus, visitation may be expressed as a function of these key variables or as: = f (d, P, T, W) [1] It must be noted that the supply data for the 1960-1965 period indicated no change in facilities; thus, time 333 could not be included as a variable. Exclusion of the time variable does not, however, limit the basic con structs of the model. When sufficient data concerning changes in facilities over time become available, the function could be developed to include the time variable. Visitation could then be expressed as: where: y = a particular year or time period over which changes in facilities would be observed. In fitting the equation, an exponential form for the unknown independent variables was chosen where vis itation (the dependent variable) may be expressed as: where: i = index denoting driving or travel time from the Phoenix-Tucson population center of gravity, with values shown below: vSdy “ f (d’ PdY’ TdY’ Wdy) [ 2] [3] i di (hrs.) Average d^ (hrs.) 1 2 3 4 5 6 0-1 2-3 3-4 4-5 5+ 1-2 0-5 1.5 2.5 3.5 4.5 6.0 334 i e = base of the natural log system, a constant, chosen for convenience in later calculations."^ x-^ = exponents, to be determined to explain the contribu- i tion of each term to visitation, where k varies from j i 1 to 5. | Strictly speaking, the independent variables I should be time oriented. For example: Where y denotes the year, i.e., y = 1960, 1961 . . ., 1965. The same is true for T and W, and even d, if travel-time zones were observed to change due to an im provement in the road system. However, since facilities did not change over the 1960-1965 time period, nor were there any major improvements to the road system that would affect travel time to the various facilities, it is necessary to use: piy = pi> Tiy = Ti> wiy = wi> and diy = di- -^The purpose of the constant term is to account for the contributions of all other variables not included in the equation, e.g., the supply variables that were ex cluded due to collinearity or statistical insignificance. 335 However, population as well as total visitation did change between 1960 and 1965. Since supply did not change during this time period, a numerical change in j supply cannot be directly equated with a numerical change in visitation. However, the proportion of supply varies i | over each distance zone, as does the proportion of vis itation. Thus, by using proportions or percentages, rather than numerical terms, visitation (the dependent i variable) can be equated with the supply factors (the in- I | dependent variables). As a checlc, the percent of visitation to each ! zone was calculated for each year and found to be highly consistent from year to year. This consistency was fur ther checked by performing a correlation analysis of I ! visitation for each year with that of every other year. i The results of this analysis further verified this con sistency as all coefficients of correlation (r's) were j above .96. The correlation matrix is shown below: Year 1960 1961 1962 1963 1964 1965 1960 1.0000 .9988 .9978 .9826 .9777 .9679 1961 1.0000 .9991 .9875 .9826 .9739 1962 1.0000 .9897 .9854 .9791 1963 1.0000 .9995 .9969 1964 1.0000 .9983 1965 1.0000 336 On the basis of this correlation study, Vg, in equation [3] for each i, where i = 1, 6, may be determined as a percentage of total visitation over all years and all zones. This has the effect of averaging the visitation over all of the years, giving each year equal weighting. It also has the effect of increasing the number of variables that may be used to fit the equa tions. In addition, the equation will be representative of any year during the 1960-1965 time period. This pro cedure is reasonable in view of the extremely high corre lations given above. The exponents x^, ..., xs in equation [3] may be determined using a least squares fit of a given data: percent Vs^ (as specified above); average driving time d^; number of picnic tables P^; number of camping tent spaces T^; and, number acres of water Wj., for each of the six zones i = 1, ..., 6. Since there are six zones, there are six non-linear algebraic equations for the five un knowns x j _ , . . . , X 5 . Although the problem is non-linear, it can be handled exactly (with no approximations) via linear theory using logarithms. By taking the natural logarithm in equation [3], the result is 337 In Vgi = xi + (In d^) x2 + (In P^) x3 + (In Ti) x4 + (In W^) x5 [4] for each i = 1, 6. i | The above system of six linear algebraic equa- ! tions may be written in vector-matrix form as Ax = y where: A is the 6 by 5 matrix A = 1 In dl In Pi In Ti In Wi 1 In d2 In P2 In T2 In w2 1 In d3 In P3 In T3 In W3 1 In d4 In P4 In T4 In w4 1 In d5 in P5 In T5 In W5 1 In d6 in P6 In T6 In W6 x is the five-dimensional vector of unknowns x, x x = 2 x3 x4 x5 [5] 338 and, y is the six-dimensional vector of the known percent visitation, for each zone y = The least squares method consists of determining | that solution of x in equation [5] which best fits the ' equation in the quadratic sense. Since the system of j i I equations is overdetermined (i.e., six equations with j five unknowns), there is no exact solution, Thus, the solution that must be determined is that which minimizes I the sum of the squares of the "residuals" (or that amount of variation in Vs_ ^ not explained by the independent variables). Vl V- vc 339 6 J = 2 Ej [6] i = 1 1 J.V where: = itn residual = Y± - (Ax)i = In Vs> - - (In d^) x2 - (In Pi) x3 - (In Ti) x4 - (In Wi) x5 for each distince zone i. Determination of that value of Xi, .. . , X5 which minimizes J in equation [6] yields the least squares fit. It is known-^ that the minimizing value of the vector x is x = (A1 A)"1 Ay [7] where "prime" denotes the matrix transpose and ( )”* the matrix inverse. Using the data for d, P, T and W, the solution of x was found to be: Xi = -9.562 x2 = -0.133 x3 = 0.572 x4 = 0.417 x5 = 0.167 36r . Deutsch, Estimation Theory (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1965). 340 This gives the empirical solution to equation [3] as: v _ e-9. 562 d-0.133 p0.572 T0.417 w 0.167 f L J To test the validity of the fit of the equation, the estimated visitation for each year during the 1960- 1965 time period was calculated and compared with the actual visits. As may be seen in Table VI-9, the fit is good as there are only relatively minor variations be tween actual and estimated visitation using the equation. This predicting equation, it may be noted, shows the trend of visitation as a function of the existing supply. The various exponents for each supply variable approximate the proportionate influence of each supply variable, i.e., the effect of travel time and the critical supply variables on the spatial distribution of visita tion, and thus serve to indicate the effect of existing supply on recreation behavior. The effect cannot be exactly specified due to the existence of the error term, e. However, when the equation is used with the actual supply data, which have large numerical values, the size of the error term is not out of proportion. Thus, the relative effects of these variables, as represented by their exponents, is considered appropriate. TABLE VI-9 COMPARISON BETWEEN ACTUAL AND ESTIMATED VISITATION Travel Time 1960 Visitation 1961 Visitation 1962 Visitation Zone Actual Estimated Actual Estimated Actual Estimated 1 2 3 4 5 6 Hour Hours Hours Hours Hours Hours 285,400 2,002,800 818,200 1,229,100 1,946,500 4.184,300 302,489 1,747,700 1,050,270 1,261,190 1,796,430 4.373.520 318,100 2,063,800 933,500 1,382,600 1,921,700 4.230,400 313,582 1.811.790 1.088.790 1,307,440 1,862,310 4,533,900 330.000 2.098.800 980.000 1.479.000 2.059.800 4.944.000 343,688 1,985,730 1,193,320 1,432,970 2,041,100 4.969,190 Total 10,466,300 10,531,599 10,850,100 10,917,812 11,891,000 11,965,998 Travel Time 1963 Visitation 1964 Visitation 1965 Visitation Zone Actual Estimated Actual Estimated Actual Estimated 1 2 3 4 5 6 Hour Hours Hours Hours Hours Hours 373,400 2,247,400 1,505,100 1,658,300 2,358,900 5.995,700 408,629 2,360,950 1,418,800 1,703,730 2,426,780 5.908,140 459,900 2,517,500 1,812,300 1.847.800 2,687,000 6.741.800 464,596 2,684,310 1,613,120 1,937,080 2,759,160 6,717,340 508,500 2,401,100 1.959.600 1,900,700 2.741.600 7.242.400 486,810 2,812,650 1,690,250 2,029,700 2,891,080 7,038,510 Total 14,138,000 14,227,029 16,075,300 16,175,606 16,843,900 16,949,000 GO 342 In addition, it may be noted that the exponents also set the parameters for the equation. That is, at a given level of visitation, a change in the magnitude of one variable, say P, requires a change in the magnitude of other variables, say T, in order to meet this level of visitation. However, the changes in magnitude of these variables must be in relation to the proportionate influence of the other variables (as represented by the exponents) in order to equal the given magnitude or level of visitation. Thus, given a new level of visitation and solving for the level of facilities required, assures the optimum combination of facilities to meet this level of visits within the parameters or relative influence of the other variables. Of further interest in the equation is the nega tive exponent for the distance variable, d, which supports the hypothesis that distance is inversely related to vis itation. This exponent may be thought of as an implicit "cost" for recreation in the supply function. Also, the exponents for picnic tables, P, tent spaces, T, and water acres, W, are all positive indicating these variables are also directly related to visitation, a result that would be expected. In addition, the exponent for P is greater than those for T and W. One reason for this might be that picnicking, as found by the National Recreation Sur- vey, is often associated with other park-oriented outdoor J recreation activities such as walking for pleasure and j driving for pleasure.37 On the other hand, there is a greater potential demand for swimming than for picnicking , and the water variable might have been expected to have a ; greater influence on visitation than the number of picnic ; j tables. i The greater influence of picnic tables as compared I to water acres, however, appears logical since the most of the water bodies in Arizona (over 80 percent of the , total water acres) are located in excess of two hours i driving time from the Phoenix-Tucson population center of j gravity, which tends to discourage participation in swim ming which is primarily a day use activity. This points to a possible shortage of recreation water acres within the 0-2 hour travel time (day use) zone. It may be noted that as the existing supply of facilities changes, the predicting equation will change. That is, depending upon the amounts, types and locations of facilities added to the supply inventory, the values 37ORRRC Report 19, p. 108. 344 of the exponents will change and perhaps additional sup ply variables may indicate sufficiently high correlations with visitation to warrant their inclusion in the equa tion. This can easily be accomplished by the recreation planner as the methodology used to derive the equation will still be valid. Forecasting Optimum Combinations of Facility Requirements Before applying the predicting equation to deter mine the optimum combinations of facilities necessary to meet future levels of potential demand, it is important to recall several key, underlying assumptions. Assumptions 1. In accordance with the underlying theory of the model, it is assumed that the quantities of facilities a public agency or recreation planner will provide at a particular distance (implicit price) will be exactly equal to the quantities demanded at that same distance (implicit price). In effect, the planner will "shift" the supply curve to meet potential demand so that VD^ = Vg^, or, using the distance zone index notation, to where = Vg^. 345 I j 2. As shown earlier, the correlations of visita tions from year-to-year during the 1960-1965 time period were extremely high, indicating that the equation would hold for any year during the period. For purposes of this; analysis, 1960 will be used as the base year. 3. The predicting equation implicitly includes | the assumption that all facilities were utilized to their | full capacity. Review of the actual visitation data for ! j the 1960-1965 period indicates this was not the case. ; Actual visits did increase over the period while the num- ! ber of facilities remained more or less constant. ! Strictly speaking, an adjustment should be made to account; for the under utilization of facilities, or, even "excess capacity" at the various levels of visitation. However, ; this adjustment was not possible due to lack of data and ! I the addition of another constant term to the equation or I j a "utilization factor" was not made. The manner in which j such a factor could be applied will be described later in this section. Thus, for purposes of this analysis, it is assumed that existing facilities were being utilized I to full capacity during the base year. Consequently, for any increase in demand (in visits) relative to actual visits in the base year, an increase in the existing level of supply is necessary. 346 4. The predicting equation represents the rela tive influence or "weights" associated with each supply variable in the 1960-1965 time period. As observed ear lier, if the existing supply of facilities changes, or if travel-time zones are expanded, the influence of such variables (the values of the exponents) could be changed to reflect the influence of the new data. For purposes of this analysis, it is assumed the relationship deter mined for the 1960-1965 time period will continue in the future until modified as just indicated. Application of the Predicting Equation Since Vd^ = Vg^, the optimum level of supply or combinations of facilities necessary to meet potential demand in some future year in a particular time zone are determined by substituting these levels of demand (ex pressed in number of visits) for Vg^ in the predicting equation and solving for P, T and W. However, as stated earlier, Vg^ was, of necessity, determined in percentage rather than numerical terms. Further, the predicting equation was determined based upon the actual visits and existing level of supply in the 1960-1965 period, or for the purposes of this analysis in the base year 1960. 347 It will be recalled that: Xl X2 X3 X4 Xc vSi = e ai pi Ti wi [!] Since Vj)^ = Vg^, and if Vs^ is converted to numer ical terms, the resulting equation can then be solved to determine the new level of supply or facility combinations necessary to meet potential demand in future years, rela tive to 1960 as the base year. Simply stated, Vg^ in percentage terms is: 0 / o _ vs (number of visits in a particular zone) Vs (total visits over all zones) Therefore, to convert Vg^ to numerical terms and solve for facility requirements needed to meet potential demand in 1960, Vs (visits in a particular zone in 1960) = xi x2 X3 X4 Xc | Vg (total visits in 1960) e di Pi Ti Wi [2] i ! 1 where: Vg in visits equals potential demand, in visits, j in 1960. Since Vg (total visits in 1960) has an absolute value, X1 as does the constant e , let X1 Vs (total visits in 1960) e = K, a constant. 348 By substituting K in equation [2], the result is: Vs. (visits in a particular zone in 1960) = x2 x3 x4 x5 K d±Z T± W± [3] This gives the combination of facility require ments to meet potential demand in the base year and through use of the constant factor K, Vg^ is converted to numerical terms. It may be noted that the same value of K, which is derived on the basis of total potential demand in 1960 (30,244,280 visits, see Table VI-2) is used in equation [3] to determine supply requirements or facility combina tions to meet demand in various time zones in all years beyond 1960, since it was assumed that an increase in the level of existing supply is necessary to meet an in crease in the future level of demand. It may also be noted that the equation could be adjusted to account for the under utilization of facilities, or for that matter, excess demand at facilities. Such data could be obtained from observation of park usage, i.e., whether the facili ties were only partially used or whether recreationists were turned away because of insufficient facilities. If this information was obtained, a "utilization" coeffi 349 cient could be applied to both sides of equation [3]. In determining optimum facility requirements, all levels of supply variables will be specified, except one, i and the quantities of this variable will be allowed to ! vary based upon specified changes in the others. The dis-; i tance variable will be specified since it is desired to determine facility combinations within each distance or travel-time zone. For ease of computation and conform ance to the two-hour travel-time zones previously delim ited from the Phoenix-Tucson population center of gravity,, average travel time values will be specified such that in the 0-2 and 2-4 hour zones, i = 1 and i = 3, respectively?! for the 4 hours and over zone, i = 5.25. ! 7 J In addition, a certain amount of water acres j j will also be specified. Given this amount, various levels| of tent spaces will then be specified and the number of | I i picnic tables allowed to vary to achieve the optimum j combination of all facilities to meet potential demand ! I in a particular time zone. Any of these variables could have been selected for specification. However, it was reasoned that the amount of water acreage that could be provided would be severely constrained by physical lim itations (e.g., where a reservoir could be constructed) 350 as well as financial limitations (e.g., the high costs typically involved). Logically then, this factor should be specified by the recreation planner. A similar rea- ! soning was made regarding specification of tent spaces j since these facilities are more expensive than picnic I I tables and there is less demand for camping than for pic- ^ nicking. Again, however, any of these variables could be ! selected for specification at the discretion of the I planner. Efttpirical Results Various combinations of facilities that must be supplied to meet projected levels of demand from 1960 to I 1990, by travel-time zone, are shown in Tables VI-10 I through VI-13. These facility combinations were obtained j by substituting the quantities of potential demand, Vd^ \ i (the number of visits expected in the target years, by j I travel zone), for Vgj in the predicting equation and j I solving for the number of picnic tables, given alternate levels of tent spaces and a fixed level of water acreage. Since these combinations were determined so that Vg^ = Vjj, they are, therefore, optimum facility combina tions of supply-demand equilibrium. These analytical results were derived in large part with the aid of a computer. 351 TABLE VI-10 FACILITY COMBINATIONS OF SUPPLY-DEMAND EQUILIBRIUM 1960 Travel Time Zone: 0-2 Hours 22,360 Acres of Water Picnic Tables Tent Spaces 500 17,683 820 8,971 1,040 2,140 6,475 2,407 3,020 1,500 4,120 980 5,000 751 25,000 Acres of Water Picnic Tables Tent Spaces 500 16,910 8,579 6,192 820 1,040 2,140 2,301 3,020 1,435 4,120 937 5,000 719 28,000 Acres of Water 30,000 Acres of Water Picnic Tables Tent Spaces Picnic Tables Tent Spaces 500 16,160 500 15,719 820 8,198 820 7,975 1,040 5,960 1,040 2,140 5,756 2,140 2,199 2,139 3,020 1,371 3,020 1,334 4,120 895 4,120 871 5,000 687 5,000 668 352 TABLE VI-10 — Continued Travel Time Zone: 2-4 Hours 11,680 Acres of Water Picnic Tables Tent Spaces 100 7,793 265 2,047 395 1,184 590 683 850 414 1,045 312 1,500 190 14,400 Acres of Water Picnic Tables Tent Spaces 100 7,166 265 1,882 395 1,089 590 628 850 381 1,045 1,500 287 175 16,000 Acres of Water Picnic Tables Tent Spaces 100 7,166 265 1,882 395 1,089 590 628 850 381 1,045 287 1,500 175 17,600 Acres of Water Picnic Tables Tent Spaces 100 6,612 1,737 265 395 1,005 590 579 850 351 1,045 1,500 265 161 20,000 Acres of Water Picnic Tables Tent Spaces 100 6,282 265 1,650 395 954 590 550 850 334 1,045 251 1,500 153 353 TABLE VI-10 — Continued Travel Time Zone: Over 4 Hours | j 90,335 Acres of Water 92,600 Acres of Water ] Picnic Tables Tent Spaces Picnic Tables Tent Spaces ! 10 6,852 10 6,785 20 2,648 20 2,622 40 1,023 40 1,013 60 587 60 581 80 395 80 392 100 291 100 288 115 240 115 238 93,800 Acres of Water Picnic Tables Tent Spaces 10 6,750 20 2,608 40 1,008 60 578 80 390 100 287 115 237 95,000 Acres of Water Picnic Tables Tent Spaces 10 6,715 20 2,595 40 1,003 60 575 80 388 100 285 115 236 354 TABLE VI-11 FACILITY COMBINATIONS OF SUPPLY-DEMAND EQUILIBRIUM 1970 Travel Time Zonet 0-2 Hours 22,360 Acres of Water 60,000 Acres of Water Picnic Tables Tent Spaces Picnic Tables Tent Spaces 2,000 17,709 2,017 11,926 4,280 6,237 4,318 4,200 8,043 2,626 11,905 1,045 11,801 1,552 17,585 612 15,561 1,062 25,181 374 24,975 555 30,856 283 39,988 291 40,332 196 Travel Time Zone: 2-4 Hours 11,680 Acres of Water 35,000 Acres of Water Picnic Tables Tent Spaces Picnic Tables Tent Spaces 398 7,539 402 4,857 796 2,913 1,666 690 1,651 1,070 4,250 191 3,359 404 7,735 84 6,785 154 12,822 42 9,364 99 15,634 32 17,806 41 18,189 26 Travel Time Zone; Over 4 Hours 90,335 Acres of Water Picnic Tables Tent Spaces 16 26,033 221 524 1,009 90 1,484 53 2,495 26 3,400 17 5,007 10 100,000 Acres of Water Picnic Tables Tent Spaces 16 25,000 41 7,000 548 200 4,865 10 5,724 8 8,064 5 9,488 4 355 TABLE VI-12 FACILITY COMBINATIONS OF SUPPLY-DEMAND EQUILIBRIUM 1980 Travel Time Zone: 0-2 Hours 22,360 Acres of Water Picnic Tables Tent Spaces 17,547 4,723 7,116 10,326 23,003 38,127 63,196 142,439 10,000 6,000 2,000 1,000 500 164 105,000 Acres of Water Picnic Tables Tent Spaces 4,798 6,220 7,950 9,355 15,505 25,701 144,067 9,995 7.000 5.000 4.000 2.000 1,000 94 Travel Time Zone: 2-4 Hours 11,680 Acres of Water 55,000 Acres of Water Picnic Tables Tent Spaces Picnic Tables Tent Spaces 943 7,896 955 4,246 1,152 6,000 1,100 3,500 1,420 4,500 1,406 1,654 2,500 1,909 3,000 2,000 4,252 1,000 2,742 1,000 7,048 500 4,544 500 71,078 21 Travel Time Zone: Over 4 Hours 73,433 11 90,335 Acres of Water 105,000 Acres of Water Picnic Tables Tent Spaces Picnic Tables Tent Spaces 38 27,494 38 25,886 57 16,000 77 10,000 66 13,000 8,000 127 5,000 94 360 1,200 259 2,000 683 500 712 500 2,207 23,070 100 24,074 4 4 356 TABLE VI-13 FACILITY COMBINATIONS OF SUPPLY-DEMAND EQUILIBRIUM 1990 Travel Time Zone: 0-2 Hours 22,360 Acres of Water Picnic Tables Tent Spaces 10,673 13,964 17,222 26,435 85,457 141,646 427,161 17,350 12,000 9.000 5.000 1.000 500 110 Travel Time Zone: 2-4 Hours 11,680 Acres of Water Picnic Tables Tent Spaces 2,122 8,025 2,624 6,000 4,349 3,000 9,688 1,000 16,057 500 51,908 100 206,957 15 Travel Time Zone: Over 4 Hours 90,335 Acres of Water Picnic Tables Tent Spaces — - — ■, Picnic Tables Tent Spaces 10,842 8,700 12,705 7,000 16,237 5,000 23,563 3,000 52,488 1,000 87,000 500 434,872 55 70,000 Acres of Water Picnic Tables Tent Spaces 2,163 3,900 2,619 3,000 3,520 2,000 5,834 1,000 9,670 500 31,259 100 167,498 10 110,000 Acres of Water Picnic Tables Tent Spaces 87 28,000 87 112 20,000 105 185 10,000 5,000 175 306 290 991 1,000 937 5,309 100 5,020 47,147 5 44,589 26,000 20,000 10,000 5.000 1.000 100 5 357 Usefulness for Recreation Planning The usefulness of the model to a recreation plan ner may be shown by the following example. Suppose, for instance, that for purposes of developing a long-range comprehensive statewide outdoor recreation plan, a plan ner wished to determine the number and type of facilities that must be supplied to meet potential demand in the year 1980 in the 0-2 hour travel-time zone (80,167,760 visits, see Table VI-4). Suppose further that due to physical, budgetary or other limitations that no new water areas, in addition to the 22,360 acres already ex isting in this zone, could be provided. The planner could then specify some new level of tent spaces or pic nic tables he felt could be provided, given these con straints, and, through use of the predicting equation, solve for the various combinations of these facilities that must be supplied. For instance, if 6,000 tent spaces were specified, then 10,326 picnic tables would be required (see Table VI-12). These supply requirements could then be compared with the existing number of tent spaces and picnic tables in this 0-2 hour zone to deter mine the deficiencies. In this instance, there are only 396 tent spaces and 942 picnic tables in this zone (see 358 Table VI-6), indicating a shortage in both types of facilities. j If, however, the planner was able to provide ad- 1 ditional water areas in this zone, say a total of 105,000 I ! acres, and specified another number of tent spaces, say ! | 5,000 spaces, then 7,950 picnic tables would be required (see Table VI-12). These figures also indicate a short- j age of both tent spaces and picnic tables over the exist- | ing levels. ] Graphic illustrations of the facility combina tions of supply-demand equilibrium for the 1960-1990 tar- I get years, by distance zone, are depicted in Figures • j VI-4 to VI-7. The negatively sloping lines shown were | derived by plotting the various combinations of picnic j tables and tent spaces (shown in Tables VI-10 through VI-13), given certain levels of water. These graphs are j particularly useful for recreation planning. By selecting a particular level of water acres (or interpolating be tween those levels shown) and choosing a particular num- i I ber of tent spaces, the number of picnic tables required i 7 ! for all facilities to be in supply-demand equilibrium can j i be read directly from the graph. Or, by selecting a cer tain level of picnic tables, the required number of tent spaces can be determined. TENT S P A C E S 359 d = 0 - 2 H O U R S W = 2 2 , 3 6 0 A C R E S 2 - 4 H O U R S W = 1 1 , 6 8 0 A C R E S 4 H O U R S +• W = 9 0 , 3 3 5 A C R E S 100 1,000 10,000 100,000 P I C N I C T A B L E S Figure VI-4 — Facility combinations of supply- demand equilibrium: 1960 TENT S P A C E S 1,000 10,000 100,000 360 d = 0 - 2 H O U R S W = 2 2 , 3 6 0 A C R E S W = 6 0 , 0 0 0 A C R E S d = 2 - 4 H O U R S W = 1 1 , 6 8 0 A C R E S X C W = 3 5 , 0 0 0 A C R E S d = 4 H O U R S -t~ 100,000 \ A C R E S 100,000 10,000 100 1,000 P I C N I C T A B L E S Figure VI-5 — Facility combinations of supply- demand equilibrium: 1970 TENT S P A C E S 1,000 10,000 100,000 361 d = 0 - 2 H O U R S -W = 2 2 , 3 6 0 A C R E S ' W = 1 0 5 , 0 0 0 A C R E S d = 2 - 4 H O U R S W = 1 1 , 6 8 0 A C R E S \ 5 5 , 0 0 0 A C R E S d = 4 H O U R S + V ^ W = 9 0 , 3 3 5 V v A C R E S d = 4 H O U R S + — W = I 0 5 , 0 0 0 A C R E S o o 100 1,000 10,000 100,000 P I C N I C T A B L E S Figure VI-6 — Facility combinations of supply- demand equilibrium: 1980 362 d = 0 - 2 H O U R S W = 2 2 , 3 6 0 A C R E S -W= 1 2 5 , 0 0 0 A C R E S d = 2 - 4 H O U R S W - l l , 6 8 0 ACRES S r W = 7 0 A C R E S d = 4 H O U R S + W = 9 0 , 3 3 5 L A C R E S h _r d = 4 H O U R S + W = 1 1 0 , 0 0 0 100 1,000 100,000 10,000 P I C N I C T A B L E S Figure VI-7 — Facility combinations of supply- demand equilibrium; 1990 363 Hence, the model output provides the recreation planner with the flexibility of determining the number of facilities, in a variety of combinations, that must be supplied to meet potential demand in any travel-time zone. These combinations can be altered depending upon the constraints facing the planner, such as budgetary limitations, physical limitations (e.g., size of an area that can be developed), availability of natural resources (e.g., land and water bodies for recreation), and others. Further, facilities and areas complementary to those specified can also be determined. For example, the number of boat access ramps, acres of swimming beaches, i j etc., can be determined from the number of water acres j specified and the amount of visitation in water-oriented ! activities; the number of tent space parking pads and acres needed for tent camps can be determined from the number of tent spaces. This can be done by applying various recreation planning standards used by planners. In addition, once the number of man-made facilities and land areas are determined, the planner can then ascertain the total land area required to support or accommodate all of these facilities as well as other recreational activities that normally occur in a park but do not need 364 or require only minimal facilities for their enjoyment (e.g., nature walks, walking for pleasure, etc.). These requirements can also be obtained by applying pertinent i recreation standards. I i Once all these future facility requirements have been determined, including the number of man-made recre- i ation units and acres of land and water, the information ! can be used to help develop a long-range statewide rec- j reation plan. This information would be useful to the decision maker in determining guidelines for the quantity j of resources and facilities that must be supplied in the j future. It would assist in decisions regarding expansion i of or additional development within existing recreation j areas. If the forecasts indicate the need for additional recreation areas, the land and water resources can be reserved in advance while they are still available; otherwise, they may be preempted for other uses. This i information would also help determine the type, amounts, and locations of facilities that must be supplied to best satisfy future demand, and assist in establishing a pro gram of priorities for and financing of facilities de velopment. Also, once the model is programmed, it would be available for "off-the-shelf" use by other agencies 365 for a wide variety of purposes (e.g., benefit-cost anal yses, etc.). It must be remembered, however, that the model calculates the results to be expected from projections of recreation behavior. Subjective criteria (e.g., types of facilities, certain constraints, use of various rec reation planning standards, etc.) must still be used to develop overall facilities and park requirements? public policy is not, therefore, made obsolete by the computer. Conclusions The foregoing approach to forecasting statewide outdoor recreation demand and supply requirements helps overcome the deficiencies of other studies and provides answers to the critical questions that must be resolved by the recreation planner. Demand is treated separately from supply and includes only those variables affecting demand. It is not constrained by supply or combined with supply variables in a reduced form market clearing equa tion to project demand. This feature, plus use of cross- section data to determine demand, helps avoid the identi fication problem and the shortcomings of previous ap proaches. 366 The methodology to measure future potential out door recreation demand also considers the usual data limitations faced by recreation planners. Moreover, the data utilized are those that are currently available ! and/or easy to obtain at minimum cost. The only addi tional unpublished information peculiar to a particular state that may have to be obtained relate to the recrea tion participation and travel patterns of state residents. In addition, the approach enables determination of the effect of existing supply on recreation behavior. j i It allows the recreation planner to estimate the relative ; influence of travel time and certain critical supply var iables on the spatial distribution of visitation, and I i determine supply requirements for new facilities within j i this framework. As such, it can be of assistance in optimizing use of existing resources. Further, the approach allows determination of the amounts, types and locations of selected facilities that must be supplied to meet potential demand, with consid eration given to the effect of existing supply on recrea tion behavior. It also permits the optimum supply re quirements to be determined in a variety of combinations to enable the planner to provide needed facilities within 367 the limits of physical and financial constraints. It may also be noted that the foregoing approach is part static and part dynamic. The static parts are the assumptions that: the relationships observed in 1960 between certain socioeconomic characteristics of the pop ulation and the rates at which the population participated in outdoor recreation will continue in the future; and, the effect of the existing supply on recreation behavior in terms of the influence of travel time and the critical supply variables observed in 1960-1965, will continue in the future. The dynamic or forecasting parts are: the account taken of the probable changes in the socioeconomic factors in the future and the effects on participation; and, the predictions of how future levels of supply must be changed to meet future levels of demand. The static assumptions regarding demand may not be serious limitations to the model, as socioeconomic factors change relatively slowly and can themselves be predicted with some assurance. On the other hand, it must be realized that when the most important and least predictable factor of all is considered — the recreation tastes and preferences of the public — the relationships between recreation participation and certain socioeconomic 368 factors could change drastically. What the public likes and dislikes now may hold true only in the most general terms in the future, and then again, their tastes and I preferences may not change at all. Thus, synthesizing projected factors affecting demand for outdoor recreation I into recreation demand figures cannot be precise. How ever, such projections can serve as a valuable guide for j planning the types, amounts and locations of recreation facilities until changes in tastes and preferences are i j observed or more current data become available. Further, j the static assumption concerning the effect of existing supply on recreation behavior is not a serious limitation | since the effect can be reanalyzed whenever there is a I change in the supply inventory, distance zones or visita- j I tion patterns. i i Thus, while the foregoing approach developed to j | forecasting statewide outdoor recreation demand and supply | : requirements has its limitations, it can be a valuable | tool for recreation planning, given the usual data limita- | tions concerning outdoor recreation. Although developed for one state — Arizona — it can easily be modified for I | use in preparing comprehensive outdoor recreation plans for other states as the basic methodology would be the same. CHAPTER VII SUMMARY AND CONCLUSIONS The major objective of this study was to develop an approach to forecasting statewide outdoor recreation demand and supply requirements. The approach formulated was designed to serve as a tool that would help provide solutions to the critical issues currently faced by state recreation planners regarding: determination of potential demand; the effect of existing supply on recreation be havior; and, the optimum combinations and locations of facilities that must be supplied to best satisfy demand. A summary of the major findings that provided the back ground for developing this approach, as well as the con clusions that may be derived, are presented below. Summary of Major Findings There are a number of problems that make it dif ficult to analyze the demand for outdoor recreation using traditional economic constructs. The basic problems are the lack of quantitative data regarding recreation demand and the complex nature of the recreation market. Adequate 369 370 j data regarding potential demand at the state level are ! | usually nonexistent or costly to obtain. Further, there i is no conventional market pricing for outdoor recreation. ! In microeconomic demand-supply analysis, price is common ( : to both the demand and supply sides of the market and the ! I s market or equilibrium price determines the quantities j I bought and sold. However, recreation is not marketed; i it is "zero priced" as a matter of public policy. There | are no entrance fees or only minimal fees to use public j recreation facilities and the costs of supplying these j facilities are borne by public agencies. However, recre- i ation is not a "free" good and does not fall neatly withinj the classification of a public good. I Other complications related to the nature of the | recreation product in that it is "jointly produced" (i.e., contains facilities for a collection of activities) and "jointly demanded" (i.e., a recreation outing to a | site consists of participation in a collection of com- I plementary activities). These demand-supply conditions make it difficult to separate and analyze the effects of | each component of demand and supply. Also, due to the | lack of adequate data regarding potential demand, time series data of past visitation to recreation sites have 371 incorrectly been used as measures of demand. Past visi tation is consumption, not potential demand. Use of such time series data to estimate potential recreation demand I (although incorrectly termed) gives rise to the misiden- tification of the function being investigated or the "identification" problem, as it is difficult to tell : Whether observed difference in visitation reflect demand I responses or supply responses. These are but a few of the complications that make it difficult to use tradi tional economic constructs to derive demand forecasts. A variety of approaches have been developed to estimate recreation demand. However, it has only been i within the past few years that economists have become | actively engaged in applying theoretical economic con- j i i structs to quantitative analyses to estimate and project j i recreation demand. Writers in the early 1920s did little j more than isolate problems of allocation, impact and in- j stitutions associated with outdoor recreation in a mar ket-oriented economy. Later works in the 1930s and early 1940s began to provide empiric content to their works, but the use of economic theory was extremely limited. Since the late 1940s, and extending into the 1950s, de mand estimation models were developed, but these were t 372 oriented more toward placing a value on recreation for use in benefit-cost analyses than toward estimates of demand. Since the 1960s (especially the late 1960s), I particular emphasis has been placed on the formulation of j j i econometric models in attempts to overcome the complex : i i t ! I factors of the recreation market and derive statistical estimates of recreation demand. Construction of such models are in their infancy and application of the re sults are extremely limited, especially for statewide ■ outdoor recreation planning. j For example, the market benefit measures attempt I to place values on recreation activity as indicators of : i recreation benefits. These methods, however, offer no j S j I analysis to generate a true demand function, or quantify j i i i I demand, or specify the properties of the demand function. | The "indirect" or "travel cost" approach assumes an in- j verse relationship between recreation demand and distance, j | and assigns an explicit dollar value to distance as a proxy for "price." Although various refinements to the | original formulation have been made, the main idea re- | mains that a market demand schedule for recreation at an existing site can be constructed and used to measure i recreation benefits. This approach, however, is site 373 oriented and attempts to estimate demand for existing sites. It does not include the various socioeconomic j I i characteristics of the population known to affect recrea- | j | tion demand or estimate the demand for new sites. It also relies on past visitation (consumption) data as a i ; ! measure of potential demand. Further, the data available ; for assigning explicit dollar values as price surrogates i I : for distance are inadequate at best. As formulated, this j | t I approach has many deficiencies for statewide outdoor recreation planning. i In contrast, the "direct" approach makes use of questionnaires to obtain "price" information from recrea- ; i tionists based upon their willingness to pay for a par- j I i ticular activity. The problem of obtaining a rational, consistent measure of price also makes this approach inadequate for statewide recreation planning. The socioeconomic method of projecting demand gained prominence with the completion of the two national | recreation surveys which developed participation rates for the population in various recreation activities. This approach rests on the assumption that the current relationship between recreation participation and the socioeconomic characteristics of the population can be 374 applied to the expected future socioeconomic structure of the population. Most statewide recreation plans prepared I | since these surveys have utilized these participation 1 rates to project potential demand. The procedures used ; to apply these participation rates do not, however, in- | elude the effect of travel time or distance on potential ! demand. Further, these plans have typically applied these , rates to determine demand only fox* a segment, and not all, ; of the population. Further, the projections do not ac count for the portion of various activities that do not i occur at parks, which are the major supply components for | which the plan is concerned. Further, these approaches j do not take into account the effect of existing supply on i j recreation behavior or determine the optimum supply re- | quirements, given this effect. Thus, these approaches j | are deficient and can lead to gross over estimates or i under estimates of demand and supply requirements. Although the econometric models developed in the I past few years to project demand have utilized certain | socioeconomic characteristics of the population to derive i ! structural demand functions, these models have also been deficient. These approaches depend on reduced form equa tions which specifically include supply in e . demand model 375 to project demand, or assume that supply always creates ! its own demand. Because of the nature of such models, I J a true measure of potential demand cannot be obtained I and the identification problem still remains. j Due to the complexities of measuring and pro- | j jecting demand, the "visits per acre," recreation stand- j ards and time-income-mobility approaches are commonly | used by recreation planners. None of these approaches, ; however, consider potential demand and are essentially ! I "consumption oriented" (i.e., based on previous visitation data) . Thus, the various approaches that have been de- | veloped to forecast recreation demand and supply require- j ments have been deficient in a number of areas. As j formulated, these approaches have limited application to ! statewide recreation planning. Generally speaking, how ever, the socioeconomic method is the most appropriate. To prepare demand estimates using the various | socioeconomic characteristics that best explain recrea- I [ tion behavior, income is the best single variable. It is a good summary measure of and highly correlated with a number of socioeconomic factors found to influence recre ation demand, and by using income, the effects of these other variables are included. Moreover, by using income 376 as the sole socioeconomic variable, intercorrelations with other socioeconomic factors can be avoided and de mand more easily determined. However, it is also neces sary to include the effect of travel time and distance as these variables also have an important effect on rec reation demand in terms of the location and comparison of facilities that must be provided. These variables are not, however, price surrogates for recreation. Rather, they are parameters in a demand function and use of the distance variable alone will account for the implicit effects of both time and distance. The approach developed in this study to forecast statewide outdoor recreation demand and the supply of facilities needed to meet this demand was designed to help overcome these difficulties and the deficiencies of other methods. It also considers the usual data limita tions faced by recreation planners by utilizing data that are currently available and/or easy to obtain at minimum cost. In this approach, a traditional market demand curve is derived using cross-section data regarding cer tain socioeconomic characteristics of the population known to influence recreation participation. By using cross- 377 sectional data rather than time series data to estimate the demand curve, the function can be held more or less I 7 i i constant as supply variability with respect to any given j i individual would essentially be removed. This helps avoid the identification problem. Further, demand is treated j separately from supply and not combined with supply vari- j ables in a reduced form market clearing equation to pro- I ject demand. This also helps eliminate the identification j problem. In addition, distance is used as the implicit (not explicit) price variable in the demand function. Since this distance variable was derived based upon data re garding recreation travel patterns obtained from a large cross-sectional sample of households residing in dispersed geographic areas, it takes on a wide range of values be- { cause of its dependence on the variation typically found j i in the location of supply. The various distances (or im- ' plicit prices) that people are willing to travel (or pay) i for the recreation "good," combined with the above socio economic characteristics, thus provide a measure of poten- j tial demand at a "price," or location by a travel-time zone, which helps identify the demand curve. Thus, the "zero price" problem is avoided. 378 In addition, a supply function is derived which is "shifted" by the recreation planner to exactly meet I the amount of potential recreation demand at the "price" ! (or distance) people would be willing to pay (or travel) | I i i for recreation. The provision of the proper amount of ! i j supply to exactly meet potential demand at this "price" I i I J thus establishes an equilibrium point between demand and supply or a "market price" for recreation. This procedure ; i \ \ allows the "market" to express itself and not be manipu lated or controlled by the planner. This "shifting" of j the supply curve by the public recreation planner is ac- ; i I : I complished independent of any shifts in the demand curve | since the provision of recreation facilities is subject I to the decisions of public bodies, and not to the usual j market forces. I i The model also, enables determination of the effect of existing supply on recreation behavior. It allows the recreation planner to estimate the relative influence of travel time and certain critical supply variables on the j spatial distribution of visitation, and determine supply ! requirements for new facilities within the framework of existing resources. Further, the model allows determination of the amounts, types and locations of facilities that must be 379 supplied to meet potential demand, given the effect of ! existing supply on recreation behavior. It also permits i the optimum supply requirements to be determined in a j i variety of combinations to enable the planner to provide j j , ; needed facilities within the limits of physical and finan- I ! cial constraints. ! ! i ; i While the approach developed to forecasting state- ; j wide outdoor recreation demand and supply requirements in j ■ this study has its limitations, it can be a valuable tool for recreation planning concerning outdoor recreation. j I Although developed for one state — Arizona — it can be j modified easily to provide outputs regarding future rec- J i j | reation demand and supply requirements for use in pre- j I j ! paring other statewide comprehensive outdoor recreation ] | i plans. I | Conclusions j ■ „ ■ - ■ - - - I As evident in the discussions throughout this ! ! study, one of the major problems in outdoor recreation | planning is the lack of adequate data concerning recrea- i I i tion demand. While the federal agencies concerned with ! | outdoor recreation have probably done the most in the way of data collection and planning, this was done for their 380 own lands, although until a decade ago, much of this work i was on a superficial basis. However, when the Outdoor Resources Review Commission was established in 1958, the federal government entered the planning and research phase of outdoor recreation on a more comprehensive scale. The | work of the Commission led to the establishment of the i | ; Bureau of Outdoor Recreation in 1962. The Bureau's func tions include collecting information, conducting research, ■ doing general recreation planning, and coordinating activ- : ities among federal agencies. Later, in 1965, Congress j passed the Land and Water Conservation Act under which i j states would be granted funds for comprehensive planning. I Research in outdoor recreation at the state level, i | however, has been almost nonexistent. Agencies admin- | istrating outdoor recreation and parks have not been authorized, staffed or financed to do research. The re- | suit is a serious lack of data concerning outdoor recrea tion. Moreover, states have been woefully weak in rec reation planning, and much of what they have done has j been in response to impetus provided by federal agencies. j I During the 1930s, federal grants made park and recreation j I planning possible for the first time in many states. The ! state park agencies have not, until recently, developed long-range plans even for their own programs and certainly I not for all recreation in their state. Within the past i several years, some states have made advances in this re- j gard, and further progress is expected to result from the | ! ! funding provisions of the Land and Water Conservation Act. j I i | However, considerably more must be done in terms j ; j of data collection and research for outdoor recreation, i I | particularly regarding demand, on a state level. While there have been two national recreation surveys which now provide some measure of demand at two points in time \ (1960 and 1965), the information developed is for the na- i ! j tion as a whole and broad census regions, not states. i ' Further, there has been insufficient analysis of the data | in the latest survey regarding the relationship between | various socioeconomic characteristics of the population I and their participation in outdoor recreation, comparable : to that made for the earlier survey. Obviously, it is desirable to correct these deficiencies by relating the j data in future surveys specifically to states and devel oping recreation participation rates, based upon detailed analysis. Further, the concept of participation rates i should be refined to eliminate possible supply constraints i ! to help provide a better measure of potential or ex ante 382 demand. There are other refinements that could be made (such as improved survey techniques) and these will un doubtedly occur in the future. However, due to the costs of such surveys, many states will still have to rely on the national surveys for some time to come. Until that time, however, statewide outdoor plan ning must continue. Several ways in which states could improve their recreation data base for use in forecasting future demand and supply requires are as follows: 1. Maintain a complete inventory of all outdoor recreation facilities provided by all recreation sup pliers, governmental and private. This inventory should be periodically updated to be available for use when demand-supply analyses are made. It would help in deter mining the change of facilities over time for use in the predicting equation, which as indicated in the preceding chapter, would take the form: X , X n X o X 4 x . . = e 1 d .2 P. T * W 5 ly ly iy !y iy where: y = a particular year or time period over which changes in facilities would be observed. 2. Refine estimates of recreation travel patterns. This could be done by sample surveys that would determine these patterns by socioeconomic characteristics of the ! ! people. Obviously, factors such as age and income would I j | have some effect on these patterns. These factors could j then be related to the socioeconomic characteristics af- ' fecting recreation tastes and preferences to provide a i better understanding of recreation demand determinants. j i If such data were gathered over a period of time, changes in travel patterns could be observed and included in pro- ; ’ i I jections of demand. i 3. Determine both the number of visits and length j | of.stay at each park for each type of park-oriented rec- I | I | reation activity. This type of information would help j j i j refine demand estimates and assist in more accurately i ! I determining the type and amount of facilities that must j | be provided. There are, of course, other data improvements that | could be made. In the final analysis, much will depend 1 | upon the time and cost to collect and analyze such data. | ! The ultimate goal for recreation planning would be the collection of sufficient data to allow development of a 384 simulation model of the outdoor recreation market that would be applicable to any state. A simulation model of the recreation market would be a powerful technique for i ' producing conditional forecasts of recreation demand and I i supply requirements in that it would allow the planner to ! j j I i ! simulate the behavior of the market in the "computer lab oratory" under varying conditions. By simulation, the j I planner could observe effects of policy changes (e.g., ' j regarding the provision of facilities), alternate planning ; j strategies, or new planning strategies. Further, with the ! ! addition of other variables, other important effects could i i : be observed, such as how urban agglormerations might be j i affected (say by changes in facilities or increases in I demand), or how the environment might be affected. Deter- i I ! mination of these other effects could be possible. i ‘ ! i | | | t APPENDIX 385 386 TABLE A-l PROPORTION OF OUTDOOR RECREATIONAL ACTIVITIES AT PARKS Activity Percent of Participation Park-Oriented Boating 42 Camping 90 Picnicking 58 Hiking 68 Horseback Riding 34 Nature Walks 65 Driving for Pleasure 52 Sightseeing 51 Swimming 44 Walking for Pleasure 33 Water Skiing 75 Source: Estimated from data in: Eva Mueller and Gerald Gurin, Participation in Outdoor Recreation: Factors Affecting Demand Among American Adults, in ORRRC Study Report No. 20 (Washington, D.C.: Government Print ing Office, 1962), p. 58. 387 TABLE A-2 PARTICIPATION RATES BY INCOME CATEGORY: BOATING 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 1.22 1.76 2.12 2.45 Under $1,500 0.39 0.56 0.68 0.78 $ 1,500 - $ 2,999 2.46 3.55 4.28 4.95 $ 3,000 - $ 4,499 2.04 2.94 3.55 4.10 $ 4,500 - $ 5,999 1.16 1.67 2.02 2.33 $ 6,000 - $ 7,999 1.01 1.46 1.76 2.03 $ 8,000 - $ 9,999 2.35 3.39 4.09 4.73 $10,000 - $14,999 3.83 5.53 6.60 7.70 $15,000 and over 5.22 7.53 9.08 10.49 Sources: Participation rates for 1960 from: Outdoor Recreation Resources Review Commission, National Recreation Survey, ORRRC Study Report No. 19 (Washington, D.C.: Government Printing Office, 1962), pp. 125, 203, 253, 310. Projections based on trends shown in: Outdoor Recreation Resources Review Commission, Prospective Demand for Outdoor Recreation, ORRRC Study Report No. 26 (Wash- i ington, D.C.: Government Printing Office, 1962), p. 26; ; Outdoor Recreation Resources Review Commission, Outdoor Recreation for America (Washington, D.C.: Government ! Printing Office, 1962), p. 221; and U.S. Department of the Interior, Bureau of Outdoor Recreation, Outdoor Rec- ; reatlon Trends (Washington, D.C.: Government Printing ; Office, April, 1967), pp. 20-21. 388 TABLE A-3 PARTICIPATION RATES BY INCOME 1960-1990 CATEGORY: CAMPING Family Income Category 1960 1970 1980 1990 All Categories 0.46 0.78 0.95 1.11 Under $1,500 0.12 0.20 0.25 0.29 $ 1,500 - $ 2,999 0.81 1.37 1.68 1.96 $ 3,000 - $ 4,499 1.02 1.73 2.11 2.46 $ 4,500 - $ 5,999 1.60 2.71 3.31 3.86 $ 6,000 - $ 7,999 2.19 3.71 4.52 5.28 $ 8,000 - $ 9,999 3.81 6.46 7.87 9.19 $10,000 - $14,999 4.27 7.21 8.78 10.25 $15,000 and over 2.35 3.99 4.86 5.67 Sources: National Recreation Survey, pp. 126, 204, 254, 311. Prospective Demand for Outdoor Recreation, p. 26; Outdoor Recreation for America, p. 221y Outdoor Recreation Trends, pp. 20-21. TABLE A-4 389 PARTICIPATION RATES BY INCOME CATEGORY: PICNICKING 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 2.14 3.40 3.66 3.88 Under $1,500 1. 32 2.10 2.26 2.39 $ 1,500 - $ 2,999 2.82 4.48 4.82 5.11 $ 3,000 - $ 4,499 5.65 8.98 9.66 10.24 $ 4,500 - $ 5,999 4.41 7.01 7.54 7.99 $ 6,000 - $ 7,999 4. 56 7.25 7.80 8.27 $ 8,000 - $ 9,999 5.53 8.79 9.46 10.02 $10,000 - $14,999 4.75 7.55 8.12 8.61 $15,000 and over 3.79 6.02 6.48 6.87 Sources: National Recreation Survey, pp. 133, 211, 261, 318 Prospective Demand for Outdoor Recreation 5 P* 26; Outdoor Recreation for America, p. 221 ; Outdoor Rec- reatlon Trends, pp. 20-21. r TABLE A-5 390 PARTICIPATION RATES BY INCOME 1960-1990 CATEGORY: HIKING Family Income Category 1960 1970 1980 1990 All Categories 0. 26 0.40 0.49 0.57 i Under $1,500 0 0 0 0 $ 1,500 - $ 2,999 0.37 0.57 0.70 0.81 $ 3,000 - $ 4,499 0.88 1.35 1.66 1.93 $ 4,500 - $ 5,999 0.63 0.97 1.19 1. 38 $ 6,000 - $ 7,999 0.40 0.62 0.75 0.88 $ 8,000 - $ 9,999 0.87 1.34 1.64 1.91 $10,000 - $14,999 1.62 2.49 3.05 3.55 $15,000 and over 0.57 0.88 1.07 1.25 Sources: National Recreation Survey, pp. 129, 207, 257, 314. Prospective Demand for Outdoor Recreation, p. 26? Outdoor Recreation for America, p. 221; Outdoor Recreation Trends, pp. 20-21. 391 TABLE A-6 ! PARTICIPATION RATES BY INCOME CATEGORY: HORSEBACK RIDING 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 0.42 0.57 0.61 0.67 Under $1,500 0. 54 0.73 0.78 0.86 $ 1,500 - $ 2,999 2.76 3.75 4.01 4.40 $ 3,000 - $ 4,499 1.46 1.98 2.12 2.33 $ 4,500 - $ 5,999 2. 35 3.19 3. 41 3.75 $ 6,000 - $ 7,999 1.93 2.62 2.80 3.08 $ 8,000 - $ 9,999 1.20 1.63 1.74 1.91 $10,000 - $14,999 1.01 1. 37 1.47 1.61 $15,000 and over 3.73 5.06 5.42 5.95 Sources: National Recreation Survey, pp. 130, 208, 258, 315. Prospective Demand for Outdoor Recreation, p. 26; Outdoor Recreation for America, pT 2^21? Outdoor Recreation Trends, pp. 20-21. 392 TABLE A-7 PARTICIPATION RATES BY INCOME CATEGORY: NATURE WALKS 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 0.75 0.87 0.95 1.03 Under $1,500 4.40 5.10 5.57 6.04 $ 1,500 - $ 2,999 2.86 3.32 3.62 3.93 $ 3,000 - $ 4,499 2.86 3.32 3.62 3.93 $ 4,500 - $ 5,999 2.38 2.76 3.02 3.27 $ 6,000 - $ 7,999 2.93 3.40 3.71 4.02 $ 8,000 - $ 9,999 3.00 3.48 3.80 4.12 $10,000 - $14,999 2.70 3.13 3.42 3.71 $15,000 and over 4. 28 4.97 5.42 5.88 Sources: National Recreation Survey, pp. 132, 210, 260, 317. Prospective Demand for Outdoor Recreation, p. 26; Outdoor Recreation for America, p. 221; Outdoor Recreation Trends, pp. 20-21. 393 TABLE A-8 PARTICIPATION RATES BY INCOME CATEGORY: DRIVING FOR PLEASURE 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 6.68 7.00 7. 79 8.30 Under $1,500 8.68 9.10 10.13 10.78 $ 1,500 - $ 2,999 17. 78 18.63 20.74 22.09 $ 3,000 - $ 4,499 22.04 23.10 25.71 27. 38 $ 4,500 - $ 5,999 19.66 20.60 22.93 24.42 $ 6,000 - $ 7,999 22.00 23.06 25.66 27. 33 $ 8,000 - $ 9,999 27.15 28.45 31.67 33.73 $10,000 - $14,999 18.82 19.72 21.95 23.38 $15,000 and over 22. 30 23.37 26.01 27.70 Sources: National Recreation Survey, pp. 127, 205, 255, 312. Prospective Demand for Outdoor Recreation, p. 26; Outdoor Recreation for America, p. 221; Outdoor Recreation Trends, pp. 20-21. 394 TABLE A-9 PARTICIPATION RATES BY INCOME CATEGORY: SIGHTSEEING 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 2.20 3.45 3.86 4.30 Under $1,500 4.03 6.32 7.07 7.88 $ 1,500 - $ 2,999 5.69 8.92 9.98 11.12 $ 3,000 - $ 4,499 7.34 11.51 12.88 14. 35 $ 4,500 - $ 5,999 6.78 10.63 11.90 13.25 $ 6,000 - $ 7,999 8.02 12.58 14.07 15.68 $ 8,000 - $ 9,999 8.91 13.97 15.63 17.42 $10,000 - $14,999 8.86 13.89 15.55 17.32 $15,000 and over 10.91 17.11 19.14 21.33 Sources: National Recreation Survey, pp. 135, 213, 263, 320. Prospective Demand for Outdoor Recreation, p. 26; Outdoor Recreation for America, p. 221; Outdoor Recreation Trends, pp. 20-21. 395 TABLE A-10 PARTICIPATION RATES BY INCOME CATEGORY: SWIMMING 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 5.15 7.74 9.15 10.60 Under $1,500 1.25 1.88 7.41 8.58 $ 1,500 - $ 2,999 4.17 6.27 10.82 12.52 $ 3,000 - $ 4,499 6.82 10.25 12.12 14.03 $ 4,500 - $ 5,999 6.98 10.49 12.40 14. 36 $ 6,000 - $ 7,999 8. 39 12.61 14.91 17.26 $ 8,000 - $ 9,999 10.02 15.06 17.80 20.61 $10,000 - $14,999 13.34 20.05 23.70 27.44 $15,000 and over 13.58 20.41 24.13 27.94 Sources: National Recreation Survey, pp. 136, 214, 264, 321. Prospective Demand for Outdoor Recreation, p. 26? Outdoor Recreation for America, p. 221; Outdoor Recreation Trends, pp. 20-21. 396 TABLE A-11 PARTICIPATION RATES BY INCOME CATEGORY: WALKING FOR PLEASURE 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 4.34 7.70 8.43 9.43 Under $1,500 20.04 35.55 38.93 43.56 $ 1,500 - $ 2,999 22.49 39.90 43.69 48.89 $ 3,000 - $ 4,499 18.82 33.39 36.56 40.91 $ 4,500 - $ 5,999 14.93 26.49 29.00 32.45 $ 6,000 - $ 7,999 15.76 27.96 30.61 34.27 $ 8,000 - $ 9,999 14.51 25.74 28.19 31.54 $10,000 - $14,999 15.71 27.87 30.52 34.15 $15,000 and over 22.76 40. 38 44.21 49.47 Sources: National Recreation Survey, pp. 137, 215, 265, 322. Prospective Demand for Outdoor Recreation, p. 26? Outdoor Recreation for America, p. 221? Outdoor Recreation Trends, pp. 20-21. 397 TABLE A-12 PARTICIPATION RATES BY INCOME CATEGORY: WATER SKIING 1960-1990 Family Income Category 1960 1970 1980 1990 All Categories 0. 30 0. 49 0.68 0.86 Under $1,500 0.14 0.23 0.32 0.40 $ 1,500 - $ 2,999 0.32 0.52 0.73 0.92 $ 3,000 - $ 4,499 0.23 0. 38 0.52 0.66 $ 4,500 - $ 5,999 0. 33 0.54 0.75 0.95 $ 6,000 - $ 7,999 0.52 0.85 1.18 1.49 $ 8,000 - $ 9,999 1.18 1.93 2.68 3.38 $10,000 - $14,999 1.79 2.92 4.06 5.13 $15,000 and over 1.45 2.37 3.29 4.16 Sources: National Recreation Survey, p p . 138 216, 266, 323. Prospective Demand for Outdoor Recreation, p. 26; Outdoor Recreation for America, p. 221; Outdoor Recreation Trends, pp. 20-21. 398 TABLE A-13 ARIZONA POPULATION: 1960, 1970 AND PROJECTED TO 1990 Year Phoenix SMS A Tucson SMSA State of Arizona Phoenix and Tucson SMSA's as Percent of Total State 1960 663,510 265,660 1,302,161 71.3 1970 967,522 351,667 1,772,582 74.4 1980 1,354,530 457,165 2,295,000 76.3 1990 1,896,342 594,313 2,950,000 84.2 Source: Data for the years 1960 and 1970, U.S. Department of Commerce, Bureau of the Census, "Components of Population Change by County: 1960-1970," Current Pop ulation Reports. Series P-25, No. 461 (Washington, D.C.: Government Printing Office, July 28, 1971), pp. 7-8. Data for the years 1980 and 1990, Valley National Bank, Economic Research Department, Phoenix, Arizona, August, 1971 (unpublished data). TABLE A-14 FAMILY INCOME DISTRIBUTIONS 1960-1990 Family Income Category Phoenix SMSA Tucson SMSA 1960 1970 1980 1990 1960 1970 1980 1990 Under $1,500 7.9 5.2 3.0 1.9 7.1 4.4 2.2 1.1 $ 1,500 - $ 2,999 11.2 7.2 4.6 3.2 11.4 7.4 4.8 3.4 $ 3,000 - $ 4,499 14.6 10.4 7.0 5. 3 16.2 12.0 8.6 6.9 $ 4,500 - $ 5,999 17.6 12.9 9.7 8.0 19.4 14.7 11.5 9.8 $ 6,000 - $ 7,999 20.4 20. 2 18.3 15.2 19.6 19.4 17.5 14.4 $ 8,000 - $ 9,999 12.2 14.4 14.2 12.6 11.3 13.5 13.3 11.7 $10,000 - $14,999 10.8 18.9 23.9 25.5 10. 3 18.4 23.4 25.0 $15,000 and over 5.3 10.8 19.3 28.3 4.7 10.2 18.7 27.7 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Sources: Data for 1960 from, U.S. Department of Commerce, Bureau of the Census, Census of Population: 1960, Vol. I, Characteristics of the Population, Part 4, Arizona Table 141, p. 4-309. Projections based on data contained in: National Planning Association, "Economic Projections for the Years 1976 and 2000," in Projections to the Years 1976 and 2000: Economic Growth, Population, Labor Force and Leisure, and Transportation, ORRRC Study Report 23 (Washington, w D.C.: Government Printing Office, 1962) , p. 182. vd TABLE A-15 U.S. NATIONAL PARK SERVICE RECREATION AREAS IN ARIZONA, 1965 Name of Areaa Driving Hours Phoenix Time in From Tucson Acreage Land Water Canyon De Chelly N.M. 6 6 83,800 40 Casa Grande Ruins N.M. 2 2 473 Chiricahua N.M. 6 3 10,696 Coronado N. Mem. 5 2 2,834 Glen Canyon N.R.A. 6 6 90,240 13,440 Grand Canyon N.M. 5 6 197,462 818 Grand Canyon N.P. 6 6 671,025 2,550 Lake Mead N.R.A. 6 6 1,173,316 70,325 Montezuma Castle N.M. 2 5 840 2 Navajo N.M. 6 6 600 Organ Pipe Cactus N.M. 4 3 330,873 1 Petrified Forest N.P. 5 6 94,189 Pipe Spring N.M. 6 6 40 Saguaro N.M. 4 1 78,644 Sunset Crater N.M. 4 6 3,040 Tonto N.M. 3 4 1,140 Tumacacori N.M. 4 1 10 Tuzigoot N.M. 3 5 43 Walnut Canyon N.M. 4 6 1,878 Wapatki N.M. 4 6 35,232 aN.M. = National Monument; N.P. = National Park; N.R.A. = National Recreation Area; N. Mem. = National Memorial. 400 TABLE A-15 — Continued Acreage by BOR Classification Name of Area I II III IV V VI Canyon De Chelly N.M. 300 40,500 40 43,000 Casa Grande Ruins N.M. 473 Chiricahua N.M. 675 200 7,117 2,704 Coronado N. Mem. 70 1,764 1,000 Glen Canyon N.R.A. 1,130 17,520 57,700 8,000 18,710 620 Grand Canyon N.M. 117,280 81,000 Grand Canyon N.P. 300 2,069 173,829 100,360 422,377 75,000 Lake Mead N.R.A. 865 1,083,536 58,880 Montezuma Castle N.M. 842 Navajo N.M. Organ Pipe Cactus N.M. 1,650 283,724 45,000 500 Petrified Forest N.P. 15,200 78,909 80 Pipe Spring N.M. 40 Saguaro N.M. 40 52,704 17,600 8,300 Sunset Crater N.M. 5 3,015 20 Tonto N.M. 860 280 Tumacacori N.M. 10 Tuzigoot N.M. 17 26 Walnut Canyon N.M. 50 1,678 150 Wapatki N.M. 25 35,037 20 150 Sources: Arizona Outdoor Recreation Coordinating Commission, Initial Outdoor Recreation Plan, State of Arizona (Phoenix, Arizona, September, 1965) , p. 31. Arizona Outdoor Recreation Coordinating Commission, A Plan for Outdoor Recreation in Arizona (Phoenix, Arizona, June, 1967), Section 4, p. 7. 401 TABLE A-16 U.S. NATIONAL PARK SERVICE RECREATION FACILITIES IN ARIZONA, 1965 Type, Number and Size of Facility Swimming Boat Tent Beach Picnic Access Camp Name of Areaa No. Acres No. Acres No. Acres No. Acres Canyon De Chelly N.M. 2 6 1 3 Casa Grande N.M. 1 1 Chiricahua N.M. 1 1 1 1 Coronado N. Mem. 1 15 Glen Canyon N.R.A. 2 4 2 11 2 32 Grand Canyon N.M. 2 2 2 2 Grand Canyon N.P. 45 20 12 466 Lake Mead N.R.A. 1 2 3 8 4 20 3 48 Montezuma Castle N.M. 2 2 Navajo N.M. 1 1 3 3 Organ Pipe Cactus N.M. 7 7 1 4 Petrified Forest N.P. 2 6 Pipe Spring N.M. 14 2 Saguro N.M. 6 60 2 8 Sunset Crater N.M. 1 1 Ton to N.M. 1 1 Tumacacori N.M. 2 1 Walnut Canyon N.M. 9 1 Wapatki N.M. 1 1 aN.M. = National Monument? N.P. = National Park? N.R.A. = National Recreation Area? N. Mem. = National Memorial. Sources; Initial Outdoor Recreation Plan, State of Arizona, p. 32. A Plan for Outdoor Recreation in Arizona, p. 9. 402 TABLE A-16 — Continued Type, Number and Size of Facility Trailer Vista Horse Foot Camp Point Marina Trail Trail Name of Area No. Acres No. Acres No. Acres (Miles) (Miles) Canyon De Chelly N.M. 1 3 5 4 5 2 Casa Grande N.M. ChiricaTiua N.M. 1 1 5 1 14 19 Coronado N. Mem. Glen Canyon N.R.A. b b 5 3 2 8 Grand Canyon N.M. 1 1 Grand Canyon N.P. 1 35 18 18 346 Lake Mead N.R.A. b b 1 1 3 6 7 Montezuma Castle N.M. 1 Navajo N.M. 1 1 8 12 Organ Pipe Cactus N.M. 1 1 2 Petrified Forest N.P. 3 3 5 Pipe Spring N.M. Saguaro N.M. 2 2 75 75 Sunset Crater N.M. 3 3 2 Tonto N.M. Tumacacori N.M. Walnut Canyon N.M. 1 1 Wapatki N.M. 1 ^Tent and trailer camps are used interchangeably. 403 TABLE A-17 U.S. NATIONAL PARK SERVICE RECREATION UNITS IN ARIZONA, 1965 ________________Type and Number of Units_______________ Picnic Boat Access Tent Trailer Marina Slips Name of Area3 Tables Parking Spaces Spaces Spaces and Moorings Canyon De Chelly N.M. 20 35 15 Casa Grande N.M. 15 Chiricahua N.M. 15 37 Coronado N. Mem. 32 Glen Canyon N.R.A. 62 400 211 b 100 Grand Canyon N.M. 4 6 Grand Canyon N.P. 191 568 163 Lake Mead N.R.A. 82 1,837 387 b 798 Montezuma Castle N.M. 33 Navajo N.M. 50 Organ Pipe Cactus N.M. 19 208 b Petrified Forest N.P. 42 Pipe Spring N.M. 12 Saguaro N.M. 180 8 Sunset Crater N.M. 5 Tonto N.M. 10 Tumacacori N.M. 2 Walnut Canyon N.M. 9 Wapatki N.M. 5 aN.M. = National Monument; N.P. = National Park; N.R.A. = National Recreation Area; N. Mem. = National Memorial. ^Trailer and tent spaces used interchangeably. Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 34. A Plan for Outdoor Recreation in Arizona, p. 10. 404 TABLE A-18 ANNUAL VISITS TO U.S. NATIONAL PARKS IN ARIZONA 1960-1965 Number of Visits (in thousands) 1960 1961 1962 1963 1964 1965 Name of Area Casa Grande N.M. Canyon De Chelly N.M. Chiricahua N.M. Coronado N. Mem. Grand Canyon N.M. Grand Canyon N.P. Glen Canyon N.R.A. Lake Mead N.R.A.3 Montezuma Castle N.M. Navajo N.M. Organ Pipe Cactus N.M. Petrified Forest N.P. Pipe Spring N.M. Saguaro N.M. Sunset Crater N.M. 76.9 87. 5 83.9 20.5 26.5 25.4 45.4 43.4 51.7 16.9 18.2 24.8 .8 .8 .9 1,186.9 1,252.2 1,446.5 2,254.2 2,220.0 2,688.7 177.7 208.8 221.3 2.5 3.9 4.5 262.1 252.2 294.1 911.5 671.0 705.0 7.3 11.5 18.8 141.0 145.1 146.3 80.5 97.6 112.0 46.1 53.6 56.2 53.8 56.3 55. 2 34.8 34.1 34.6 66. 3 73.8 76.5 56.6 65.2 71.7 105.7 132.4 88.1 30.0 167.5 182.8 51.8 43.7 47.0 28.3 40.0 59.1 .8 .9 1.3 1,538.7 1,575.7 1,689.2 196.4 303.5 3,349.6 3,462.6 3,594.1 231.7 222.0 232.3 7.9 7.9 17.1 329.8 324.7 362.8 786.0 884.0 867.8 24.9 22.4 27.6 177.0 215.9 229.7 145.4 139.7 156.7 68.4 70.4 54.1 54.4 55.6 56.1 38.6 45.4 48.6 80.0 79.4 83.1 83.0 63.0 82.5 Tonto N.M. Tumacacori N.M. Tuzigoot N.M. Walnut Canyon N.M. Wapatki N.M.______ aArizona and Nevada. Source: A Plan for Outdoor Recreation in Arizona, p. 8. 405 TABLE A-19 U.S. FOREST SERVICE RECREATION AREAS IN ARIZONA, 1965 • Driving Time in Hours From Total Acreage Name of Forest Phoenix Tucson Land Water Apache 6 5 1,188,912 1,209 Coconino 3 6 1,796,871 3,767 Coronado 5 2 1,681,720 202 Kaibab 5 6 1,722,981 525 Prescott 3 5 1,247,220 358 Sitgreaves 4 5 779,773 419 Tonto 2 4 2,865,742 29,104 Sources: Initial Outdoor Recreation Plan* State of Arizona, p. 29. A Plan for Outdoor Recreation in Arizona, p., 19. 406 TABLE A-19 — Continued Acreage by BOR Classification Name of Forest I II III IV V VI Apache 3,678 998,838 40 187,565 Coconino 1,859 5,044 1,772,400 21,269 66 Coronado 12,920 19,153 1,528,155 44,533 73,000 4,161 Kaibab 3,778 1,713,945 5,783 Prescott 10,687 1,209,400 27,468 23 Sitgreaves 9,295 766,997 3,900 Tonto 15,040 97,008 2,123,394 360 358,905 139 407 TABLE A-20 U.S. FOREST SERVICE RECREATION FACILITIES IN ARIZONA, 1965 ________ Number and Size of Facility_______ Apache Coconino Coronado Kaibab Type of Facility No. Acres No. Acres No. Acres No. Acres Swimming Beach 2 4 Picnic 4 34 17 35 56 346 2 4 Boat Access 2 4 5 5 1 1 4 6 Tent Camp 13 141 14 77 24 220 6 138 Trailer Camp 14 144 16 106 5 12 Group Camp 4 46 8 224 Vista Point 2 3 1 1 5 7 1 2 Marina 1 1 Horse Traila 204 276 5 34 Foot Traila 7 288 aLength in miles. Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 29. A Plan for Outdoor Recreation in Arizona, p. 20. 408 TABLE A-20 — Continued Number and Size of Facility Prescott Sitgreaves Tonto Totals Type of Facility No. Acres No. Acres No. Acres No. Acres Swimming Beach 1 4 3 8 Picnic 8 34 3 39 18 59 108 551 Boat Access 1 1 2 2 11 23 26 42 Tent Camp 9 76 4 34 24 203 94 889 Trailer Camp 1 3 2 7 38 272 Group Camp 9 126 2 41 2 38 25 475 Vista Point 13 13 2 8 24 34 Marina 4 4 5 5 Horse Trail3 546 2 999 2,066 Foot Traill 295 aLength in miles. 409 TABLE A-21 U.S. FOREST SERVICE RECREATION UNITS IN ARIZONA, 1965 T i m e n-F Number of Units Unit Apache Coconino Coronado Kaibab Prescott Sitgreaves Tonto Totals Picnic Tables 372 477 945 288 334 266 508 3,190 Park Spaces 50 190 20 47 70 1,085 1,462 Tent Spaces 122 222 333 222 191 94 355 1,539 Trailer Spaces 127 297 66 15 41 546 Marina Slips and Moorings 40 130 170 Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 30. A Plan for Outdoor Recreation in Arizona, p. 28. 410 TABLE A-22 ANNUAL VISITS TO U.S. NATIONAL FORESTS IN ARIZONA 1960-1965 National Number of Visits Forest 1960 1961 1962 1963 1964 1965 Apache 278,600 282,100 306,200 335,700 493,400 480,800 Coconino 449,600 584,300 612,900 1,313,700 1,700,100 1,860,000 Coronado 1,156,100 1,274,300 1,386,200 1,567,800 1,694,500 1,463,200 Kaibab 183,100 210,500 199,900 299,400 412,800 680,000 Prescott 479,600 501,100 514,400 506,200 550,900 578,900 Sitgreaves 118,800 128,500 157,900 173,500 236,700 301,900 Tonto 2,031,000 2,019,000 2,011,200 2,103,100 2,388,600 2,361,000 Source: A Plan for Outdoor Recreation in Arizona, p. 21. TABLE A-23 ARIZONA STATE PARK RECREATION AREAS Name of Park Lyman Lake Havasu Lake Buckskin Mt. Picacho Peak Tombstone Court house Painted Rocks Tubac Presidio Jerome Yuma Territorial Prison Driving Time in Hours From Phoenix Tucson 5 5 5 2 4 3 4 3 6 6 6 1 6 3 2 5 Total Acreage Land Water 120 13,000 54 640 1 20 5 2 10 40 11,600 100 Acreage by BOR Classification II III IV 160 11,950 32 1,050 2,600 9,000 22 640 100 VI 1 20 5 2 10 Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 41. A Plan for Outdoor Recreation in Arizona, p. 33. h* to TABLE A-24 ARIZONA STATE PARK RECREATION FACILITIES Number and Size of Facility Buckskin Lake Lyman Painted Mt. Jerome Havasu Lake Rock Type of Facility No. Acres No. Acres No. Acres No. Acres No. Acres Swimming Beach 1 2 1 10 1 1 Picnic 1 5 1 1 1 5 1 1 Boat Access 1 2 1 3 2 3 Tent Camp 1 7 3 30 1 3 Trailer Camp 1 4 Vista Point 1 1 Marina 1 5 Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 42. A Plan for Outdoor Recreation in Arizona, p. 34. TABLE A-24 — Continued Number and Size of Facility Picacho Tombstone Tubac Yuma Peak Courthouse Presidio Prison Totals Type of Facility No. Acres No. Acres No. Acres No. Acres No. Acres Swimming Beach 3 13 Picnic 1 1 1 1 6 14 Boat Access 4 8 Tent Camp 5 40 Trailer Camp 1 4 Vista Point 1 1 2 2 Marina 1 5 415 TABLE A-25 ARIZONA STATE PARK RECREATION UNITS, 1965 Type and Number of Units Name of Park Picnic Tables Boat Access Spaces Tent Spaces Trailer Spaces Marina Buckskin Mt. 15 40 Jerome 4 Lake Havasu 100 100 Lyman Lake 30 30 9 10 Painted Rock 8 Picacho Peak Tombstone Courthouse Tubac Presidio 6 Yuma Prison 3 Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 43. A Plan for Outdoor Recreation in Arizona, p. 36. TABLE A-26 ANNUAL VISITS TO ARIZONA STATE PARKS 1960-1965 Number of Visits Name of Park 1960 1961 1962 1963 1964 1965 Jerome 9,121 Havasu Lake 56,600 Lyman Lake 2,500 11,000 28,600 28,100 24,400 Tombstone Courthouse 17,792 51,200 41,100 57,500 57,500 69,900 Tubac Presidio 269 14,100 24,600 31,500 50,000 44,000 Yuma Territorial Prison 86,400 111,900 135,900 144,600 156,000 Source: A Plan for Outdoor Recreation in Arizona, p. 35. 416 TABLE A-27 RECREATION AREAS IN ARIZONA COUNTIES, 1965 Driving Time in Hours From No. of Recreation Total Acreage Acreage by BOR Classification County Phoenix Tucson Areas Land Water I II III IV V VI Maricopa 1 4 12 93,967 400 575 18,020 7,450 67,162 360 Pima 4 1 5 12,715 415 870 11,339 25 66 Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 47. A Plan for Outdoor Recreation in Arizona, p. 54. TABLE A-28 RECREATION FACILITIES IN ARIZONA COUNTIES, 1965 Number and Size of Facility Maricopa Pima Type No. Acres No. Acrei Swim Beach 1 5 Picnic 7 170 10 134 Boat Access 2 4 Tent Camp 5 17 1 3 Trailer Camp 5 21 1 10 Group Camp 4 19 1 60 Vista Point 5 12 Horse Trail3 42 Foot Trail3 60 aIn miles. Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 47. A Plan for Outdoor Recreation in Arizona, p. 55. 419 TABLE A-29 RECREATION UNITS IN ARIZONA COUNTIES, 1965 Number of Units Type of Unit Maricopa Pima Picnic Tables 196 134 Boat Access Parking Spaces 20 Tent Spaces 61 Trailer Spaces 74 40 Sources: Initial Outdoor Recreation Plan, State of Arizona, p. 48. A Plan for Outdoor Recreation in Arizona, p. 57. TABLE A-30 ANNUAL VISITS TO ARIZONA COUNTIES RECREATION AREAS 1960-1965 Driving Hours Time in From Visits County Phoenix Tucson 1960 1961 1962 1963 1964 1965 Maricopa 1 4 314,322 354,490 367,196 410,288 507,751 563,125 Pima 4 1 13,456 19,972 29,212 43,596 60,618 77,158 Source: A Plan for Outdoor Recreation in Arizona, p. 57. to o i BIBLIOGRAPHY 421 SELECTED BIBLIOGRAPHY Books Alonso, William. Location and Land Use. Cambridge, | Mass.: Harvard University Press, 1964. i Barlowe, Raleigh. Land Resource Economics. Englewood j Cliffs, N.J.: Prentice-Hall, Inc., 1958. | Baumol, William J. Welfare Economics and the Theory j of the State. 2d ed. Cambridge, Mass.: Harvard University Press, 1967. Bilas, Richard A. Intermediate Microeconomic Theory. Preliminary edition. New York: McGraw-Hill Book Comp any, 1965. Bish, Robert L. The Public Economy of Metropolitan Areas. Chicago: Markham Publishing Company, 1971. j | Ciriacy-Wantrup, S. V. Resource Conservation, Economics i and Policies. Berkeley: University of California j Press, 1952. ! Clawson, Marion, and Knetsch, Jack L. Economics of Outdoor Recreation. Resources for the Future, Inc. Baltimore: Johns Hopkins Press, 1966. Deutsch, R. Estimation Theory. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1965. j Ely, Richard T., and Wehrwein, George S. Land Economics. ! New York: The Macmillan Company, 1940. Friedrich, Carl J. Alfred Weber1s Theory of Location. Chicago: University of Chicago Press, 1929. | Goldberger, A. S. Econometric Theory. New York: John Wiley & Sons, Inc., 1964. 422 423 Henderson, James M., and Quandt, Richard E. Micro- j economic Theory. New York: McGraw-Hill Book j Company, 1958. ! 1 Hoover, Edgar M. The Location of Economic Activity. j New York: McGraw-Hill Book Company, 1948. j Isard, Walter. Location and Space Economy. Cambridge, j Mass.: The M.I.T. Press, 1956. j Katona, George. The Powerful Consumer: Psychological Studies of the American Economy. New York: j McGraw-Hill Book Company, 1960. | Kilbridge, Maurice D.; O'Block, Robert P.; and Teplitz, Paul V. Urban Analysis. Boston: Harvard Uni versity Press, Graduate School of Business ! Administration, 1970. Klein, L. R. An Introduction to Econometrics. New York: Prentice-Hall, Inc., 1962. | Landsberg, H. H. y Fischman, L. L.; and Fisher, J. L. Resources in America^ Future: Patterns of Re- j guirements and Availabilities. 1960-2000. Re- j sources for the Future, Inc. Baltimore: Johns j Hopkins Press, 1963. j i I Leftwich, R. H. The Price System and Resource Alloca- j tion. Rev. ed. New York: Holt, Rinehart & j Winston, 1961. j Liebhafsky, H. H. The Nature of Price Theory. Homewood, I 111.: The Dorsey Press, Inc., 1963. ! Losch, August. The Economics of Location. Trans, by j William H. Woglom. 2d rev. ed. New Haven, | Conn.: Yale University Press, 1954. ! Marshall, Alfred. Principles of Economics. 8th ed. ! New York: The Macmillan Company, 1920. j McCarthy, E. J. Basic Marketing: A Managerial Approach. 3rd ed. Homewood, 111.: Richard D. Irwin, Inc., 1968. 424 Musgrave, Richard A. The Theory of Public Finance. New Yorks McGraw-Hill Book Company, 1959. National Recreation and Park Association. Outdoor Rec reation Space Standards. Washington, D.C.: National Recreation and Park Association, 1965. Neumeyer, M. H., and Neumeyer, E. S. Leisure and Rec reation . 3rd ed. New York: A. S. Barnes & Company, 1958. Newman, Phillip C. The Development of Economic Thought. New York: Prentice-Hall, Inc., 1952. Nicosia, F. M. Consumer Decision Processes: Marketing j and Advertising Implications. Englewood Cliffs, j N..J.: Prentice-Hall, Inc., 1966. j I Stonier, A. W. , and Hague, D. C. A Textbook of Economic | Theory. 2d ed. New York: John Wiley & Sons, i Inc., 1961. | I ] Watson, Donald S. Price Theory and Its Uses. Boston: | Houghton Mifflin Company, 1963. i Bator, F. M. "The Anatomy of Market Failure," Quarterly Journal of Economics. Vol. LXXII, No. 3 (August 1958). Bechter, Dan.M. "Outdoor Recreation," Monthly Review. Federal Reserve Bank of Kansas City (November 1970). Becker, Gary S. "A Theory of the Allocation of Time," The Economic Journal, Vol. LXXV, No. 3 (September 1965). Boulding, Kenneth E. "Economic Analysis," Microeconomics. Vol. I. 4th ed. New York: Harper & Row, Pub lishers, 1966. Articles and Periodicals 425 Boyet, Wayne E., and Tolley, George S. "Recreation Pro jection Based on Demand Analysis," Journal of Farm Economics, Vol. XLVIII, No. 4 (November i ' 966y;---------- Carey, Omer L. "The Economics of Recreations Progress and Problems," Western Economic Journal, Vol. Ill, No. 1 (Spring 1965). Chidester, L. W. "The Importance of Recreation as a Land Use in New England," Land Economics, Vol. X, No. 2 (May 1934). Ciricacy-Wantrup, S. V. "Conceptual Problems in Pro jecting the Demand for Land and Water," in Land Economics Institute, Modern Land Policy. Urbanas University of Illinois Press, 1960. Clawson, Marion. "The Crisis in Outdoor Recreation," American Forests (March-April 1959). Crutchfield, James A. "Valuation of a Fishery Resource," Land Economics, Vol. XXXVIII, No. 2 (May 1962). Daiute, R. J. "Methods for Determination of Demand for Outdoor Recreation," Land Economics, Vol. XLII, No. 3 (August 1966). Davidson, Paul; Adams, F. Gerald; and Seneca, Joseph. "The Social Value of Water Recreational Facil ities Resulting from an Improvement in Water Quality: The Delaware Estuary," Water Research. Edited by Allen V. Kneese, and Stephen C. Smith. Baltimore: The Johns Hopkins Press, 1966. Fischer, D. W., and Gates, J. M. "A Comment on ’User Response in Outdoor Recreation: A Production Analysis,1" Journal of Leisure Research. Vol. II, No. 2 (Spring 1970). Gosse, L. E., and Kalter, R. J. "User Response in Out door Recreation: A Comment," Journal of Leisure Research, Vol. II, No. 2 (Spring 1970). 426 Havinghurst, R., and Feigenbaum, K. "Leasure and Life Styles," American Journal of Sociology, Vol. LXIV, No. 4 (January 1959). Hines, Lawrence G. "Measurement of Recreation Benefits — A Reply," Land Economics, Vol. XXXIV, No. 4 (November 1958). Johnson, M. Bruce. "On the Economics of Road Congestion," Econometrica, Vol. XXXII, No. 1 (February 1964). "Travel Time and the Price of Leisure," The Western Economic Journal. Vol. IV, No. 2 (May 1966). Kalter, Robert J., and Gosse, Lois E. "Recreation Demand Functions and the Identification Problem," Journal of Leisure Research. Vol. II, No. 1 (Winter 1970). Knetsch, Jack L. "Assessing the Demand for Outdoor Rec reation," Journal of Leisure Research. Vol. I, No. 1 (Winter 1969). _________. "Outdoor Recreation Demands and Benefits," Land Economics, Vol. XXXIX, No. 4 (November 1963) . Lessinger, Jack. "Measurement of Recreation Benefits — A Reply," Land Economics. Vol. XXXIV, No. 4 (November 1958). Los Angeles Times. "Wilderness-Use Curbs ‘Just Around the Corner*" (September 27, 1971), Part 1, pp. 3, 25. Mead, M. A. "The Patterns of Leisure in Contemporary American Culture," Annals of American Academy of Political and Social Sciences, Vol. 313, No. 14 (September 1957). Merewitz, Leonard. "Recreational Benefits of Water Re source Development," Water Resources Research, Vol. II, No. 4 (1966). 427 Milam, R. L., and Pasour, E. C. "Estimating the Demand for an On-Farm Recreational Service," American Journal of Agricultural Economics. Vol. LII, No. 1 (February 1970). Pearse, Peter H. "A New Approach to the Evaluation of Non-Priced Recreational Resources," Land Eco nomics, Vol. XLIV, No. 1 (February 1968). Renshaw, Edward F. "The Economics of Highway Conges tion," Southern Economic Journal. Vol. XXVIII, No. 4 (November 1962). Robinson, Warren C. "The Simple Economics of Public Outdoor Recreation," Land Economics, Vol. XLIII, No. 1 (February 1967). Samuelson, Paul A. "Aspects of Public Expenditure Theory," Review of Economics and Statistics, Vol. XL, No. 4 (November 1958). ________ . "Diagrammatic Exposition of a Pure Theory of Public Expenditures," Review of Economics and Statistics, Vol. XXXVII, No. 4 (November 1955). ________ . "The Pure Theory of Public Expenditures," Review of Economics and Statistics, Vol. XXXVI, No. 4 (November 1954). Seckler, David W. "On the Uses and Abuses of Economic Science in Evaluating Public Outdoor Recreation," Land Economics. Vol. XLII, No. 4 (November 1966). Seneca, J. J., and Cicchetti, C. J. "User Response in Outdoor Recreation: A Production Analysis," Journal of Leisure Research. Vol. I, No. 3 (Summer 1969). Sinclair, Robert. "Von Thunen and Urban Sprawl," Cultural Geography: Selected Readings. Edited by Fred E. Dohrs and Lawrence M. Sommers. New York: Thomas Y. Crowell Company, 1968. Thompson, Donald L. "New Concept: Subjective Distance on Store Impressions Affect Estimates of Travel Time," Journal of Retailing, Vol. XXXIX, No. 1 (February 1963). 428 Trice, Andrew H., and Wood, Samuel E. "Measurement of Recreation Benefits," Land Economics. Vol. XXXIV, No. 3 (August 1958). Wehrwein, G. S. "Some Problems of Recreation Land," Land Economics, Vol. Ill, No. 2 (May 1927). , and Johnson, H. A. "A Recreation Livelihood Area," Land Economics, Vol. XIX, No. 2 (May 1943). _________, and . "Zoning Land for Recreation," Land Economics, Vol. XVIII, No. 1 (February 1942) . _________, and Spilman, R. G. "Development and Taxation of Private Recreational Land," Land Economics, Vol. IX, No. 4 (November 1933). Wennergren, E. Boyd. "Valuing Non-Market Priced Recrea tional Resources," Land Economics, Vol. XL, No. 3 (August 1964). _________. "Surrogate Pricing of Outdoor Recreation," Land Economics, Vol. XLIII, No. 1 (February 1967). Williams, R. M. "Individual and Group Values," Annals of American Academy of Political and Social Sciences, Vol. 371 (May 1967) . Working, E. J. "What Do Statistical 'Demand Curves' Show?" Quarterly Journal of Economics, Vol. XLI (1927), as reprinted in Readings in Price Theory. Edited by K. Boulding and S. Stigler. Chicago: Richard D. Irwin, Inc., 1952. Reports and Government Publications Arizona Outdoor Recreation Coordinating Commission. Initial Outdoor Recreation Plan — State of Arizona. Phoenix, Arizona, September 1965. 429 A Plan for Outdoor Recreation in Arizona. Phoenix, Arizona, June 1967. Barkley, Paul W. "The Development of Research in the Economics of Recreation," in Cooperative Regional Research Technical Committee for Project No. WM-59, An Economic Study of the Demand for Out door Recreation. Report No. 1. San Francisco, 1968. Brown, William G.; Singh, Ajmer; and Castle, Emery N. An Economic Evaluation of the Oregon Salmon and Steelhead Sport Fishery. Technical Bulletin 78, Agricultural Experiment Station. Corvallis, Oregon: Oregon State University, September 1964. Cicchetti, C. J.; Seneca, J. J.; and Davidson, P. The Demand and Supply of Outdoor Recreation; An Econometric Analysis. New Brunswick, N.J.: Rutgers-The State University, Bureau of Economic Research, 1969. Clawson, Marion. Methods of Measuring the Demand for and Value of Outdoor Recreation. Reprint No. 10. Resources for the Future, Inc. Washington, D.C., 1959. Daane, Kenneth E. The Economic Implications of the Re gional Park System in Maricopa County. Tempe, Arizona: Bureau of Business Services, College of Business Administration, Arizona State Uni versity, March 1964. Guedry, L. J., and Stoevener, Herbert H. "The Role of Selected Population and Site Characteristics in the Demand for Forest Recreation?" in Cooperative Regional Research Technical Committee for Project No. WM-59. An Economic Study of the Demand for Outdoor Recreation. Report No. 2. Reno, Nevada, 1970. Hotelling, H. Letter to Director of the United States National Parks Services (1947) as cited in Robinson, Warren C. "Economic Evaluation of Out door Recreation Benefits," in Outdoor Recreation Resources Review Commission. Economic Studies of Outdoor Recreation. ORRRC Study Report 24. Washington, D.C.: Government Printing Office, 1962. Johnston, W. E., and Pankey, V. S. "Use Prediction Models for Corps of Engineers Reservoirs in California," An Economic Study of the Demand for Outdoor Recreation. Report No. 1. Cooperative Regional Technical Committee for Project WM-59, I 1958. j Lee, Ivan M. "Economic Analysis Bearing Upon Outdoor j Recreation," in ORRRC Study Report 24. Economic ! Studies of Outdoor Recreation. Washington, D.C.: | Government Printing Office, 1962. j Little, A. D., Inc. Tourism and Recreation; A State of j the Art Study. Prepared for the Office of Re- j gional Development Planning. U.S. Department of | Commerce. Washington, D.C.: Government Printing Office, 1967. j McNeely, John G. "Estimation of Demand for Water Based Outdoor Recreation Activities," An Economic Study of the Deanand for Outdoor Recreation. Report No. 1. Cooperative Regional Research Technical Committee for Project No. WM-59, 1968. | Mueller, Eva, and Gurin, Gerald. Participation in Out- j door Recreation; Factors Affecting Demand Among American Adults, in ORRRC Study Report 20. j Washington, D.C. : Government Printing Office, ! 1962. I I National Planning Association. "Economic Projections for the Years 1976 and 2000," Projections to the Years 1976 and 2000: Economic Growth. Population. ] Labor Force and Leisure and Transportation. ORRRC Study Report 23. Washington, D.C.: Gov ernment Printing Office, 1962. Outdoor Recreation Resources Review Commission. National Recreation Survey. ORRRC Study Report j 19. Washington, D.C.: Government Printing Office, 1962. 431 ________ . "Outdoor Recreation and the Megalopolis," | The Future of Outdoor Recreation in Metropolitan | Regions of the United States. ORRRC Study Re port 21. Washington, D.C.: Government Printing i Office, 1962. j ________ . Outdoor Recreation for America. Washington, j D.C.: Government Printing Office, 1962. | ________ . Prospective Demand for Outdoor Recreation. ORRRC Study Report 26. Washington, D.C.s j Government Printing Office, 1962. j Outdoor Recreation Resources Review Commission Staff, j National Planning Association, Bureau of Labor i Statistics, U.S. Department of Labor, and Golden- thal, A. J. Projections to the Years 1976 and 2000: Economic Growth, Population. Labor Force and Leisure and Transportation. ORRRC Study Report 23. Washington, D.C.: Government Printing j Office, 1962. ! Perloff, Harvey S., and Wingo, Lowdon. "Urban Growth and j the Planning of Outdoor Recreation," Trends in American Living and Outdoor Recreation. ORRRC I Study Report 22. Washington, D.C.: Government | Printing Office, 1962. Robinson, Warren C. "Economic Evaluation of Outdoor | Recreation Benefits," in Outdoor Recreation Re sources Review Commission. Economic Studies of Outdoor Recreation. ORRRC Study Report 24. Washington, D.C.: Government Printing Office, 1962. | J Scott, Anthony. "The Valuation of Game Resources: Some Theoretical Aspects," Canadian Fisheries Reports. Report No. 4. Ottawa: Department of Fisheries, Queens Printer, 1965. j j Stanford Research Institute. California Recreation and ' Parks Study: An Element of the State Resources j Development Program. Prepared for the State of j California Department of Parks and Recreation. j Stanford Research Institute, South Pasadena, J California, December 1965. 432 Stoevener, Herbert H., and Guedry, L. J. "Sociological Characteristics of the Demand for Outdoor Rec reation," in Cooperative Regional Research Tech nical Committee for Project No. WM-59. An Eco nomic Study of the Demand for Outdoor Recreation. Report No. 1. San Francisco, California, 1968. U.S. Department of Commerce, Bureau of the Census. "Components of Population Change by County: 1960- 1970," Current Population Reports. Series P-25, No. 461. Washington, D.C.: Government Printing Office, July 28, 1971. ________ . Census of Population: 1960. Vol. I. Char acteristics of the Population, Part 4, Arizona Table 141. ________ . Statistical Abstract of the United States: 1970. 91st ed. Washington, D.C.: Government Printing Office, 1970. U.S. Department of the Interior, Bureau of Outdoor Rec reation. Outdoor Recreation Grants-in-Aid Manual. Washington, D.C.: Government Printing Office, September 1965. ________ . Outdoor Recreation Trends. Washington, D.C.: Government Printing Office, April 1967. ________ . The 1965 Survey of Outdoor Recreation Activ ities. Washington, D.C.: Government Printing Office, 1968. Wilson, Robert R. "Consumer Behavior Models: Time Allo cation, Consumer Assembly and Outdoor Recrea tion," in Cooperative Regional Research Technical Committee for Project No. WM-59. An Economic Study of the Demand for Outdoor Recreation. Report No. 2. Reno, Nevada, 1970. Other Sources Davis, Robert K. "The Value of Big Game Hunting in a Private Forest," Transactions of the Twenty- Ninth North American Wildlife and Natural Re sources Conference. 1964. 433 Gray, James R., and Anderson, L. Wayne. "Recreation Economics in South Central New Mexico." Bul letin 488. Agricultural Experiment Station, University Park. New Mexico: New Mexico State University, 1964. Knetsch, Jack L. "Some Topics of Interest in Outdoor Recreation Economics," from the Proceedings of a Seminar, April 5-6, 1965. Outdoor Recreation Research. Texas: Texas A & M University. Lerner, L. J. "Quantitative Indices of Recreational Value," Proceedings of the Committee on the Economics of Water Resources Development. Western Agricultural Economics Research Council and Western Farm Economics Association. Reno, Nevada, 1962. Meramec Basin Research Project. "Recreation," The Meramec Basin, Vol. Ill (Water Needs and Prob lems) . St. Louis: Washington University, 1961. Morrison, C. C. "A National Survey of Outdoor Recreation Participation and Preferences." Speech delivered at the Annual Meeting of the Association of American Geologists. Toronto, Canada, August 30, 1966. Myles, George A. "Competitive Aspects of Recreational Areas," in Proceedings 1967, Western Farm Eco nomics Association. Las Cruces, New Mexico, 1967. Volk, Donald J. "Factors Affecting Recreational Use of National Parks." Paper given at the Annual Convention of the Association of American Geographers. Columbus, Ohio, 1965. 434 Unpublished Materials Barker, Edward H. "An Economic Analysis of the Changes in Retail Store Sales and Location in the Cen tral Business Districts of One Hundred and Nine Central Cities in the United States, 1948-1958." Unpublished Ph.D. dissertation, University of Southern California, 1963. Valley National Bank, Economic Research Department. Phoenix, Arizona. Unpublished population pro jections for the state of Arizona.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Participation Of Rank-And-File Member In The Decision Making Process And His Access To The Grievance Machinery Of His Union: Including A Survey And Case Study Of Uaw Local 1531, Amarillo, Texas

PDF

An Economic Analysis Of Nurse Mobility Patterns

PDF

The Evolution Of The Role Of Government In The Economic Development Of Mexico

PDF

Collective Bargaining, Disputes And Dispute Settlement Procedures In Public Employment

PDF

The computation of military materiel requirements

PDF

American Political Economics: An Emerging Consensus

PDF

The Anomaly Of External Effects: A History Of Concept Development

PDF

Grade Labeling And The Demand Curve: Theoretical Implications And Empirical Evidence

PDF

Arbitration Of Technological Change

PDF

The Political Economy Of The Islamic State; A Comparative Study

PDF

Oskar Lange'S Contributions To The Theory Of Socialism, With Applicationsto Yugoslavia And Czechoslovakia

PDF

The economics of air cargo transportation

PDF

A Comparative Appraisal Of Developmental Planning In India And Communist China Under The Second Five Year Plans

PDF

The Size Distribution Of Income And Its Relationship To Economic And Social Well-Being In The State Of California

PDF

Competition And Concentration In The Brewing Industry

PDF

Economic Aspects In The Evolution Of The Great Lakes Freighter

PDF

Effects Of Swedish Fiscal Policy And Planning On Economic Growth And Stability, 1930-1964

PDF

Interpersonal Relations In Ethnically Mixed Small Work Groups

PDF

Analysis Of Small Manufacturers: Pricing Policies In Southern California

PDF

A Study Of Debt Adjustment In Michigan

Asset Metadata

Creator Mcmahon, Peter Joseph (author)

Core Title Forecasting Selected Statewide Recreation Requirements

Degree Doctor of Philosophy

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag economics, general,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Phillips, E. Bryant (committee chair), Backer, Edward H. (committee member), Pollard, Spencer D. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-474717

Unique identifier UC11362110

Identifier 7217483.pdf (filename),usctheses-c18-474717 (legacy record id)

Legacy Identifier 7217483.pdf

Dmrecord 474717

Document Type Dissertation

Rights Mcmahon, Peter Joseph

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA