University of Southern California Dissertations and Theses

Talent migration: does urban density matter?

(USC Thesis Other)

Talent migration: does urban density matter?

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

TALENT MIGRATION: DOES URBAN DENSITY MATTER?

By

Cheng-Yi Lin

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(POLICY, PLANNING, AND DEVELOPMENT)

August 2013

Copyright 2013 Cheng-Yi Lin
DEDICATION

To my beloved parents and those who have supported me along my academic journey
ii

TABLE OF CONTENTS

LIST OF TABLES iv

ABSTRACT vii

CHAPTER 1. INTRODUCTION 1
1.1. Background and Motivation 1
1.2. Research Questions, Approach, and Significance 3
1.3. Organization of the Study 5

CHAPTER 2. REVIEW OF THE LITERATURE 7
2.1 Population Density and Talent Migration 7
2.1.1. Theoretical Debate 7
2.1.2. Business Cycles, Life Courses, and Density-migration
Nexus 12
2.2 Determinants of Internal Migration 15

CHAPTER 3. DATA AND METHODOLOGY 19
3.1. Data 19
3.2. Background on Talent Migration 27
3.3. Model Specifications 30

CHAPTER 4. DENSITY, TALENT MIGRATION, AND THE
BUSINESS CYCLE 32
4.1. Descriptive Analysis 32
4.2. Estimation Results 36
4.2.1. Results for MSA Population Density 36
4.2.2. Results for PUMA Population Density 42
4.2.3. Results for PUMA “Talented” Population Density 46
4.2.4. Results for “Boom” Year vs. “Bust” Year 50
4.3. Discussion 58

CHAPTER 5. DENSITY, TALENT MIGRATION, AND LIFE
CYCLE 61
5.1. Descriptive Analysis 61
5.2. Estimation Results 62
5.2.1. Results for Metropolitan Area Population Density 62
5.2.2. Results for PUMA Population Density 68
5.2.3. Results for PUMA “Talented” People Population Density 73
5.3. Discussion 78

CHAPTER 6. CONCLUSION 80
6.1. Summary of Findings 80
6.2. Contributions and Policy Implications 82
6.3. Directions for Future Research 85
iii

REFERENCES 87

APPENDICES 99
Appendix A. Occupational Types of the Creative Class 99
Appendix B. Regression Results for the Migration Models – MSA
Population Density (without amenity variables) 102
Appendix C. Regression Results for the Migration Models – MSA
Population Density (with amenity variables) 108
Appendix D. Regression Results for the Migration Models – PUMA
Population Density (without amenity variables) 114
Appendix E. Regression Results for the Migration Models – PUMA
Population Density (with amenity variables) 120
Appendix F. Regression Results for the Migration Models – PUMA
“Talented” People Population Density (w/o amenity variables) 126
Appendix G. Regression Results for the Migration Models – PUMA
“Talented” People Population Density (w/ amenity variables) 132

iv

LIST OF TABLES

Table 3.1. Variable Definitions and Data Sources 25

Table 3.2. Migrants in the U.S. (MSA Non-GQ Population) 28

Table 3.3. BAPLUS Numbers of Migrants and Their Shares of Group Population
(MSA Non-GQ Population) 29

Table 3.4. MAPLUS Numbers of Migrants and Their Shares of Group Population
(MSA Non-GQ Population) 29

Table 3.5. SCC Numbers of Migrants and Their Shares of Group Population (MSA
Non-GQ Population) 29

Table 3.6. BOH Numbers of Migrants and Their Shares of Group Population (MSA
Non-GQ Population) 29

Table 4.1. Descriptive Statistics for MSA PUMAs Characteristics 33

Table 4.2. Top 25 PUMAs Receiving BAPLUS In-migrants, 2009 35

Table 4.3. Correlations between PUMA Population Density and the Number of In-
Migrants (Non-GQ & MSA) 36

Table 4.4. Regression Results for the Migration Models – MSA Population Density
& All (2006) (without amenity variables) 40

Table 4.5. Regression Results for the Migration Models – MSA Population Density
& All (2006) (with amenity variables) 41

Table 4.6. Regression Results for the Migration Models – PUMA Population
Density & All (2006) (without amenity variables) 44

Table 4.7. Regression Results for the Migration Models – PUMA Population
Density & All (2006) (with amenity variables) 45

Table 4.8. Regression Results for the Migration Models – “Talented” Population
Density & All (2006) (without amenity variables) 48

Table 4.9. Regression Results for the Migration Models – “Talented” Population
Density & All (2006) (with amenity variables) 49

Table 4.10. Regression Results for the Migration Models – MSA Population Density
& All (2009) (without amenity variables) 52

v

Table 4.11. Regression Results for the Migration Models – MSA Population Density
& All (2009) (with amenity variables) 53

Table 4.12. Regression Results for the Migration Models – PUMA Population
Density & All (2009) (without amenity variables) 54

Table 4.13. Regression Results for the Migration Models – PUMA Population
Density & All (2009) (with amenity variables) 55

Table 4.14. Regression Results for the Migration Models – “Talented” Population
Density & All (2009) (without amenity variables) 56

Table 4.15. Regression Results for the Migration Models – “Talented” Population
Density & All (2009) (with amenity variables) 57

Table 4.16. Density Effects on Talented Groups 58

Table 5.1. Correlations between PUMA Population Density and the Number of In-
Migrants (Non-GQ & MSA)

62

Table 5.2. Parameter Estimates for 2005 MSA Population Density (without amenity
variables) 64

Table 5.3. Parameter Estimates for 2005 MSA Population Density (with amenity
variables) 65

Table 5.4. Parameter Estimates for 2008 MSA Population Density (without amenity
variables) 66

Table 5.5. Parameter Estimates for 2008 MSA Population Density (with amenity
variables) 67

Table 5.6. Parameter Estimates for 2005 PUMA Population Density (without
amenity variables) 69

Table 5.7. Parameter Estimates for 2005 PUMA Population Density (with amenity
variables) 70

Table 5.8. Parameter Estimates for 2008 PUMA Population Density (without
amenity variables) 71

Table 5.9. Parameter Estimates for 2008 PUMA Population Density (with amenity
variables) 72

Table 5.10. Parameter Estimates for 2005 PUMA “Talented” People Population
Density (without amenity variables) 74
vi

Table 5.11. Parameter Estimates for 2005 PUMA “Talented” People Population
Density (with amenity variables) 75

Table 5.12. Parameter Estimates for 2008 PUMA “Talented” People Population
Density (without amenity variables) 76

Table 5.13. Parameter Estimates for 2008 PUMA “Talented” People Population
Density (with amenity variables) 77

vii

ABSTRACT

This dissertation sought to examine whether urban population density matters for talented
migrants. Most researchers in the fields of urban planning and urban economics utilize
population or employment density based on large geographic units. However these
density measures may mask considerable variation across localities and thus just plain
“density” is much too general and much too vague to yield useful findings. To overcome
this issue, this study constructs various measures of urban population density – MSA
population density, PUMA population density, and PUMA “talented” people density –
and tests whether they are statistically significant in explaining the in-migration of
talented individuals in the continental United States. The dissertation addresses three
research questions. First, are denser places especially attractive to talented people?
Second, does density become more or less important for the talented migrants in the face
of the recent economic downturn? Third, does the density-talent migration nexus vary
with age, educational or occupational groups?
For the first research question, estimation results from multivariate tests show that,
although the effects of various density measures are small when PUMA densities are
utilized, talented individuals do prefer denser places. This finding corroborates what the
theories of the agglomeration economies suggest in terms of density preferences for
talented individuals.
Turning to the second question, correlation coefficients do show that PUMA
population density is relatively stronger in the “bust” year (2009). Multivariate tests using
MSA population density show that although the density variable exerts mixed effects on
viii

various talented groups in 2006, it has positive effects on the talented in 2009. This
suggests that density has become more important for talented migrants in the “bust” year
(2009). When PUMA population density and PUMA “talented” people population density
are tested however, the difference between the “boom” and “bust” years of density effects
is not obvious. Talented movers do not systematically change their density preferences
over the recent business cycle.
As for the third research question, although theories suggest that younger people
may prefer denser places whereas older people may prefer places with lower densities,
the estimation results of this study do not support this prediction for the older talented
groups. Various talented groups all prefer denser places. Nonetheless, both the correlation
coefficients and multivariate tests do show that the effect of population densities wanes as
we move up the age stratum. Density is more important for the youngest (25-34) migrant
group. Another point worth mentioning is that by and large PUMA population densities
do not matter for the SCC and BOH migrants belonging to the 45-54 age group. This may
suggest that the more senior members of the creative class may be idiosyncratic when
making their migration decisions.
1

CHAPTER 1
INTRODUCTION

1.1. Background and Motivation
The ability of regions and cities to attract talented people is thought to be a critical
factor facilitating their growth prospects. Various theories including new growth theory
(also known as the endogenous growth theory; Romer, 1986, 1990; Lucas, 1988, 1993),
theory of entrepreneurship (Audretsch and Keilbach, 2005; Acs, 2006), human capital
theory (Glaeser, 1994; Glaeser et al., 1995; Simon and Nardinelli, 1996, 2002; Glaeser
and Shapiro, 2003; Abel and Gabe, 2011), and creative capital theory (Florida, 2002,
2003) have been proposed to explain the linkage between talented people and regional
and urban growth. The essential argument of these theories is that places that attract
talented people are most likely to prosper.
One way to attract talented individuals is to develop industrial clusters. Suggested
by Porter (1990), establishing clusters for various industries has been deemed a
promising way to bolster urban and regional economic development (Bröcker et al.,
2003; Cumbers and MacKinnon, 2004). With the concept of the creative class
popularized by Richard Florida, attracting creative people to localities has been viewed
as a way to strengthen the local economy. In particular creating cultural and creative
2

clusters that concentrate creative professionals has become a popular strategy for
politicians, policymakers, planners, and economists to promote physical and economic
revitalization of cities (Mommaas, 2004).
1

Indeed, inspired by Alfred Marshall, urban and regional economists have
emphasized the importance of spatial proximity or concentration in promoting
production and consumption functions of cities and contributing to their economic
growth. Urban planners have also emphasized the importance of concentrating people
in certain places. The proponents of Compact City, New Urbanism, and Smart Growth
policies have promoted a high-density planning paradigm, which has recently been
associated with “sustainability” policies (e.g., fewer energy consumption, less
automobile use; Barrett, 1996) and the possible revitalization of central cities (Dawkins
and Nelson, 2003). All of these suggest that places with higher densities offering a
variety of production and consumption amenities could be especially attractive to
talented individuals.
On the other hand, the literature on residential location choice suggests that most
Americans favor low-density suburbs. Several studies also point out that many talented
individuals prefer to live at low densities (Glaeser, 2005; David Sawicki in Lang and
Danielsen, 2005, p. 216; Markusen, 2006; Karlsson, 2011).

1
A number of studies (for example, van der Linde, 2003; Wolfe and Gertler, 2004) however find that
innovative clusters are not easily replicable by government intervention policies.
3

This dissertation seeks to test the aforementioned competing views on the
density-migration nexus for talented individuals, notably following Gordon and Ikeda’s
(2011) approach by using finer (sub-metropolitan) population density measures at the
Public Use Microdata Area (PUMA) level. This study extends their work in four ways.
First, in addition to PUMA population density used by them, two additional density
measures are constructed to examine density effects on talented migration. Second, this
study adds two more groups of creative class arrivals with reference to Richard
Florida’s definitions. Third, this study analyzes data for both 2006 and 2009 which
allows me to examine whether the recent economic downturn affected the
density-talent migration nexus. Fourth, various migrant groups are further classified by
three age strata. By so doing, this study provides a more nuanced view of the effects of
population densities on talent migration.

1.2. Research Questions, Approach, and Significance
Noting two fundamentally competing views – i.e., the attractiveness of high-density
places versus that of low-density places for talented individuals, this dissertation
develops and addresses three research questions. First, are denser places especially
attractive to talented people, where alternate definitions of the group are tested? Second,
does density become more or less important for the talented in the face of the recent
4

economic downturn? Third, does the density-talent migration nexus vary with age,
educational or occupational groups?
To investigate these questions, this study develops more fine-grained population
density measures and conducts multivariate tests. Gordon and Ikeda (2011) and Gordon
(2013) argue that the standard density measures used by most researchers involve
averages of the number of population or employment over large geographic areas such
as the city (Wheeler, 2004), county (Ciccone and Hall, 1996; Gabe et al., 2007;
McGranahan and Wojan, 2007), metropolitan area (Knudsen et al., 2008; Scott, 2010;
Abel and Gabe, 2011) or even state (Glaeser and Resseger, 2010). Given the features of
the contemporary urban spatial structure, many urban areas in the world include
multiple sub-centers or various sizes (Anas et al., 1998). Average densities based on
these large and complex geographic units include “significant noise” (Gordon, 2013, p.
672) and thus mask considerable sub-metropolitan variations associated with the
density preferences of talented people. These accounts may explain why previous
studies on the density-talent migration linkage have produced mixed results. By using
finer density measurements, this dissertation elaborates the role of population density
on talented migration.
This study also tests the foundations of the linkage between density and talent
migration and carries importance underlying planning practice. Although theories on
5

internal migration suggest that both business cycles and life cycle of individuals affect
people’s migration decisions, little research has tested the density-talent migration
nexus by taking into account both of two factors. This study hence suggests new
insights into the density-talent migration study by investigating whether the business
cycle as well as migrants’ life cycle stage affect talented migration associated with
population density.
There are also practical implications from this study. As mentioned earlier, many
cities attempt to attract and concentrate talented individuals. Is there an optimal density
for places that policymakers and planners can designate? This study suggests that
answers to this question are much more complex. This work aspires to inform planners
as they craft policies on urban and regional planning.

1.3. Organization of the Study
This dissertation is organized as follows. Chapter 2 reviews theories on density and
talented migrants and the determinants of internal migration. Two competing theories
explain the density-migration nexus for talented individuals. Agglomeration theories
posit that the talented should prefer to live or work at high densities whereas theories of
residential location choice suggest that they favor lower-density suburbs. These
theories provide the foundation for econometric tests in the subsequent empirical
6

chapters. Determinants of internal migration offer the rationale for control variables
included in the empirical models. Chapter 3 describes data, analysis of the recent U.S.
migration patterns, and model specifications. Chapter 4 presents the estimation results
for migration models examining the association between various population density
measures and migration behaviors of talented people. It also examines the role of
density for talented individuals in the face of the recent economic downturn. Chapter 5
reports the regression results on whether the talented at various phases of their life
course have different density preferences with respect to their educational and
occupational backgrounds. Chapter 6 develops the conclusions in the way of major
findings, contributions, and policy implications. It further describes the limitations of
this study and offers directions for future research.

7

CHAPTER 2
REVIEW OF THE LITERATURE

“The future of the city depends on the demand for density. If cities are going to survive
and flourish, then people must continue to want to live close to one another.”
Glaeser et al. (2001, p. 27)

2.1. Population Density and Talent Migration
2.1.1. Theoretical Debate
The existence of industrial clusters demonstrates the importance of physical proximity or
density for individuals of various talents. For many individuals, their jobs involve
innovation activities so it is important for especially talented and specialized
professionals to develop learning opportunities through access to new ideas, information,
and knowledge. “Dense” concentration promotes frequent face-to-face interaction
opportunities, which in turn facilitate learning activities essential to innovations by well-
educated and skilled workers. But there are many types of clusters – at various densities.
This leaves open the question of what the proper density is in each situation. There are
always trade-offs and there can be no such thing as the unconstrained maximization of
“density.”
This is why one can ask, with the advent of information technologies, does
proximity still matter for the productivity of talented individuals? Theories of
agglomeration economies offer various explanations. First, their jobs often involve the
fast flow of information (such as with financial professionals on Wall Street) and
8

complex processes of idea exchange. For instance, industrial languages or jargon used by
semiconductor production engineers in Silicon Valley cannot always be understood by
their counterparts in Boston’s Route 128 (Saxenian, 1990). In addition, many economic
activities involve the transmission of tacit knowledge or complex non-codifiable know-
how. To fully grasp these messages requires mutual understanding and trust among
economic agents (Leamer and Storper, 2001). Achieving mutual understanding and trust
is more challenging without repeated face-to-face interactions. Sometimes even the
transmission of simple, codifiable information or messages requires trust and mutual
understanding facilitated by face-to-face interactions (Kiriakos, 2011). Geographic
proximity that facilitates frequent face-to-face interactions therefore speeds the flow of
ideas (Glaeser et al., 1992; Glaeser, 1998) and promotes effective idea exchange. Second,
studies show that knowledge diffusion attenuates as distance increases (Krugman, 1991;
Jaffe et al., 1993). Physical proximity thus allows talented individuals to better capture
the benefits of knowledge spillovers that are conducive to their productive activities.
Third, being in dense places can provide talented individuals with formal and informal
networking opportunities that stimulate creative processes (Karlsson, 2011). Moretti
(2012) points out that his best ideas often come from informal interactions. Saxenian
(1994) describes the story about the informal interactions among workers in Silicon
Valley after working hours. A similar story appeared in the New York Times, which shows
how non-business-related activities — for example, cricket games — facilitate the
circulation of job information among workers in Silicon Valley.
1
In addition, sharing a
locality, talented workers may obtain unofficial and serendipitous information (Kiriakos,

1
See Lohr, S. “Silicon Valley Shaped by Technology and Traffic,” The New York Times, December 20,
2007, accessed June 15, 2013,
http://www.nytimes.com/2007/12/20/technology/20cluster.html?pagewanted=all for further information.
9

2011), which might not be able to obtained via formal meetings or electronic
communications.
Except for offering learning opportunities fostering innovations, geographic
proximity also generates other benefits for workers. Dense concentrations of workers can
facilitate quality matches between firms and workers (Wheeler, 2001; Andersson et al.,
2007) and can increase labor productivity (Ciccone and Hall, 1996; Ciccone, 2002;
Baptista, 2003; Puga, 2010). In addition, various theories predict that workers in dense
urban areas receive higher wages for their higher productivity. Recent studies corroborate
this prediction (Andersson et al., 2007; Di Addario and Patacchini, 2008).
The cited discussions focus on the production-side agglomeration benefits available
in high-density places. Glaeser et al. (2001) however argue that in the field of urban
economics the function of dense cities as consumption enhances has been inadequately
studied. Glaeser and his colleagues (2001, 2006) argue that density plays a key role in
facilitating the consumption functions of cities. Dense places provide people with a
variety of amenities such as museums, theaters, professional sports games, and
restaurants. Scale economies and specialization in dense urban areas make these diverse
choices possible and beneficial to consumers. High density also reduces transport costs,
specifically the time cost, which is highly valued in the modern era, and facilitates social
interactions. Consumption amenities make denser urban area more attractive to people
even when their wage level is the same as the level available in less dense places.
2

Agglomeration theories accordingly predict a variety of ways by which denser places can

2
They also argue that high density does not guarantee urban prosperity. To succeed, dense urban areas need
to offer amenities that are attractive to the well-educated. Cities such as New York, San Francisco, London,
Paris, and Barcelona are all dense cities that provide a rich set of consumption amenities and they are all
high human capital cities.
10

be attractive to talented people.
Yet high density is not always an unmixed blessing for residents. Diseconomies of
scale also arise with the benefits of agglomeration economies. Many dense urban areas
have been associated with negative externalities such as traffic congestion and more
crime. Moving to suburban communities with lower densities has been the choice of
many Americans to avoid these externalities. Massive residential suburbanization in the
United States had taken off at least since the mid-1900s. At the beginning of the twenty-
first century, many Americans work and live in the suburbs (Glaeser and Kahn, 2001).
3

The majority of them prefer to live in low-density suburban communities (Downs, 1994;
Gordon and Richardson, 1997). Why do most Americans choose to live at low densities?
Two sets of theories have been offered to explain this phenomenon (Mieszkowski and
Mills, 1993). One set of theories is called the natural evolution theory. It postulates that
when people get richer, they demand larger housing and more open space. In addition,
since people’s opportunity costs rises with higher incomes, they prefer to avoid traffic
congestion. Low-density suburban communities, coupled with the diffusion of the
automobile (Kopecky and Suen, 2010) and highways (Baum-Snow, 2007), provide
people with the possibility to satisfy these desires.
The second set of theories is called the fiscal-social problems approach. This
approach attributes suburbanization to high taxes, low-quality public services, racial
tensions, high crime rates, congestion, and inferior environmental quality in high-density
central cities. Recent studies have however challenged studies supporting the second
theory by citing the endogeneity issue between various factors such as crime (Jargowsky
and Park, 2009) and suburbanization.

3
This trend does not change even after the recent economic downturn (Kolko, 2012).
11

Does the low-density residential pattern preferred by many Americans also apply to
the especially talented individuals? Glaeser (2005) suggests that places with a large
number of skilled people are not necessarily of high density. David Sawicki points out
that creative class people actually live in suburbs near Atlanta (in Lang and Danielsen,
2005, p. 216). Markusen (2006) also argues that highly educated workers, including some
of the “super-creative core” defined by Richard Florida, work and live in low-density
suburbs. Similarly, Karlsson (2011) also stresses that many artists such as painters are
more creative when residing in artistic communities in rural areas.
Recent empirical studies have produced mixed findings on the relationship between
population density and residential choice for talented people. McGranahan and Wojan
(2007) examine residential preferences of the creative class with respect to 1990 county-
level population density in metropolitan and non-metropolitan areas. They find that
higher density is more attractive to the members of the creative class in non-metropolitan
areas given that a reasonable level of services requires a certain level of density. On the
contrary, in search of higher quality of life, the members of the creative class within
metropolitan areas prefer places with lower density. In studying artists’ locations, Wojan
et al. (2007) find that arts employment is concentrated in various levels of geography
from both large and small metro areas to non-metro and rural counties.
4
In one of their
regression models, they delineate arts employment share in 1990 as a function of
observable county-level characteristics that include population density. They find that the
parameter estimates of population density, despite not statistically significant for non-
metro counties, are statistically significant and positive for both all counties and metro

4
The 2012 population density of San Juan County, Colorado, one of the arts magnets mentioned, is less
than one person per square kilometer.
12

counties. The arts employment share is higher in moderate-density metro counties. Scott
(2010) studies the in-migration decisions of U.S. engineers between 1994 and 1999. He
finds that 1990 metropolitan population density has a positive effect on engineers in the
Chemical and Petroleum categories but has a negative impact on Electrical and
Electronics and Industrial types. Gordon and Ikeda (2011) utilize Public Use Microdata
Area (PUMA) population density to test whether density matters for the in-migration of
highly-educated people and people belonging to “Arts, design, entertainment, sports, and
media occupations.” They show that PUMA population density has positive effects on
these movers.
In sum, agglomeration theories suggest that denser places should be more attractive
to talented people. On the other hand, theories of residential suburbanization posit that
low-density has been an attractor for both the general public and talented individuals.
Recent empirical studies exploring the relationship between population density and
residential choice of talented people are however mixed.
2.1.2. Business Cycles, Life Courses, and Density-migration Nexus
Business cycles can affect people’s migration decisions. Studies show that the rates of
internal migration are higher in the “boom” years but lower in the “bust” years of
business cycles (Makower et al., 1939; Saks and Wozniak, 2011). A recent study by Frey
(2009) corroborates this trend. He finds that the U.S. had the lowest overall migration
rate during the 2007 to 2008 recession period since World War II. How does the recent
economic downturn affect the density-migration nexus? Frey (2009) finds that throughout
the 2000-2008 period, emerging suburban and exurban counties gained more in-migrants
than mature suburban counties. For core and high-density counties, some had fewer in-
13

migrants while others had more movers over this period. In a similar vein, Kolko (2012)
finds that although after the recent economic downturn some people were still drawn to
the highest-density areas, most Americans still prefer low-density suburbs. During the
boom period, hot housing market seems to drive people to low-density suburbs and
exurbs in search of cheaper housing (Frey, 2009). The finding by Kolko that most people
still prefer low-density suburbs may be due to the fact that the recent economic crisis
limits people’s mobility and discourages long-range migration (Frey, 2011; Stoll, 2013).
On the other hand, as shown in aforementioned studies, during both the boom and bust
years there are always people who prefer high-density neighborhoods. This group of
people may value more the production and consumption benefits that high-density places
offer. Are talented people sensitive to the recent economic crisis when making their
moving decisions with respect to density? It seems that no theories or empirical work
address this question. Some studies however give some clues. Although young adults and
college graduates are thought to be more mobile, the recent economic downturn
discourages their migration (Frey, 2011). Saks and Wozniak (2011) find that business
cycles did not affect the migration decisions of older adults (36-65-year-old). In addition,
they find that more educated workers are less procyclical. Among the highly educated,
older people are especially acyclical. These findings suggest that business cycles may not
play a role in shaping migration decisions for the well-educated.
Since Rossi (1955), life cycle has been recognized as an important determinant for
individual and household migration. How does the density-migration nexus vary across
life cycles? Life-course perspectives on migration postulate that younger adults are drawn
to denser places given that these destinations offer them more opportunities such as more
14

job options. Middle-aged people aged 45-54 (those with children in particular) may
demand larger housing and thus prefer lower density suburbs. Older people (the retired
elderly in particular) prefer less dense localities in search of amenities and better quality
of life.
5
Bures (1997) examines the relationship between density and migration for five
age groups (25 and above). The density preferences of most age groups (25-44 and 65+)
are consistent with the predictions of the life-course framework. Since people who are in
their mid-life years (45-54) have more resources to move, they tend to leave dense areas
for moderate density environments. Two recent studies specifically examine the density-
migration nexus for talented people while taking life cycle into account. Whisler et al.
(2008) use MSA population density to study the migration of different life-course groups
of college-educated people. They find that young and about-to-retire people are more
likely to leave from high-density metropolitan areas. The findings with respect to young
people contradict the prediction of the life cycle perspective. In studying U.S. migrant
engineers, Scott (2010) stratifies them into two age groups – 60 and below as well as 61
and above. He finds that for engineers aged 60 and below metropolitan population
density has a positive impact on engineers in Chemical and Petroleum fields but has a
negative impact on Electrical and Electronics and Industrial workers. For engineers of
age 61 and above, density decreases the number of Electrical and Electronics engineers.
In addition to densities of localities, talented individuals also take other things into
account when making their moving decisions. The next section reviews determinants of
internal migration that provide the rationale of choosing proper control variables for the
empirical chapters.

5
Millington (2000)’s study supports the hypothesis that the importance of labor force considerations
declines over the life course while housing and amenities become more important along the lifecycle.
15

2.2. Determinants of Internal Migration
Starting with Ravenstein’s (1885, 1889) work on the laws of migration, there have been
extensive empirical studies on human migration. Stillwell and Congdon (1991) identify
two approaches for migration modeling – micro and macro approaches that provide a
classification framework. Micro theory focuses on individuals and households. It studies
people’s decisions on whether to move, choices among destinations,
6
reasons for
migration, and the relationship between personal characteristics (e.g., age, race, marital
status, educational attainment, income, occupation) or household characteristics (e.g.,
size, composition) and migratory behavior. Macro theory on the other hand focuses on
the relationship between macro variables and migration. It studies how broader context
variables such as unemployment rate affect migratory behaviors. Although this
conceptual framework dichotomizes the determinants of migration, when making their
moving decisions, individuals and households often take both micro and macro factors
into account.
Two representative theories for the micro approach are the human capital theory of
migration
7
and the life cycle theory. Human capital theory views migratory behavior of
people as an investment decision (Schultz, 1961; Sjaastad, 1962). While making their
moving decisions, people take into account both pecuniary and non-pecuniary costs and
benefits associated with their origin and destination places. Once they have decided to
migrate, movers are assumed to maximize their utility and expect to reap net benefits in
the long run.

6
The decisions on whether to move and where to move are deemed joint decisions (Linneman and Graves,
1983).
7
Human capital theory here has to do with migration, which is different from the one mentioned earlier on
economic growth.
16

Life cycle theory recognizes the different impulses that impact the migration
decisions of different age groups. The motivations of moving for individuals and
households are likely to change as they experience various stages of their lives. Different
life events including marriage, divorce, separation, birth and growth of children, new
household formation of children, and retirement may lead to a change in household
composition and size and in turn trigger housing adjustment and migration.
Existing studies find that once people are getting older, their mobility rates drop
(Long, 1988; Haurin and Haurin, 1991; Molloy et al., 2011). However, mobility rates rise
again at the beginning of retirement period (Miller, 1977; Plane, 1993). When older
people do migrate, they seem to have different considerations compared with younger
people. For instance, when making migration decisions, older couples close to retirement
may value more for amenities (Chen and Rosenthal, 2008). Job opportunities, wages, and
quality of education for children may be less important for the elderly. After getting
retired, the elderly tend to move closer to their children in order to get health assistance
(Silverstein, 1995) and lower their living costs (Walters, 2000). In addition, they tend to
move to places with nice weather (Rappaport, 2007). Life cycle theory also applies to
talented individuals. Hansen and Niedomysl (2009) study the migration of creative
people in Sweden and find that this group of people is more likely to move right after
they finished university studies.
Macro theory, instead of focusing on individual level, examines the relation between
migration and macro variables associated with origin and/or destination characteristics. A
widely-citied theory is the pull-push theory (Lee, 1966). It posits that “pull factors” can
attract people to places whereas “push factors” deflect people from places. Since this
17

study focuses on in-migrants, this section only discusses “pull factors” of the destination.
Economic factors are frequently mentioned “pull factors.” Since Ravenstein’s (1885)
work, a number of studies (Shaw, 1985; Greenwood and Hunt, 1989) have identified
employment and wage opportunities as major determinants of migration. In addition,
business cycles can affect people’s migration decisions. Studies across countries show
that the rates of internal migration are higher in the “boom” years but lower in the “bust”
years of business cycles (Makower et al., 1939; Molho, 1984; Gordon, 1985; Greenwood
et al., 1986; Greenwood, 1997; Saks and Wozniak, 2011). A recent study by Frey (2009)
corroborates this trend. He finds that the U.S. has the lowest overall migration rate during
the 2007 to 2008 recession period since World War II.
However, many recent studies (e.g., Rappaport, 2008; 2009) have found that
amenities of localities play a key role in attracting people. Glaeser et al. (2001) propose
the “Consumer City” thesis. They argue that once people get richer, they place more
value on the quality of life. Cities that can offer a variety of goods and services (e.g., nice
restaurants, live performance venues), good public services (e.g., good schools and less
crime), nice physical settings (nice weather, proximity to the coast), and reduced costs of
travel time can attract more people, highly-educated people in particular.
In recent years, a wealth of research has specifically examined the migration of
highly-educated and creative workers. The major focus of this strand of research is to
determine whether job or amenity related factors play a greater role in attracting talented
people. Creative capital theory proposed by Richard Florida (2002) has stimulated many
studies on the importance and behavior of the “creative class.” The underlying argument
of the creative capital theory is that a “creative class” of individuals is the driving force
18

for regional growth; places that attract and retain this group of people are most likely to
prosper. Florida argues that the 3 T’s – Technology, Talent, and Tolerance – determines
the success of places. In addition, he stresses that the members of the creative class are
drawn to places that are open, tolerant, and diverse. He argues that it is the life-style
available at places, not jobs, that attracts the members of the creative class. Empirical
evidence on whether 3 T’s attract or retain talented people is still rare however (King,
2011).
Some recent studies suggest that job-related factors play a more important role in
drawing the talented movers. Hansen and Niedomysl (2009) and Niedomysl and Hansen
(2010) sent a questionnaire to survey Swedish migrants in 2007. They find that compared
with movers with lower education, highly educated migrants consider job as an important
factor in their moving decisions. Martin-Brelot et al. (2010) use the data from a
questionnaire survey responded by more than 2,300 creative professionals in 11 European
cities to test Richard Florida’s hypothesis that the members of the creative class are
mainly attracted to places with ‘soft’ factors such as diversity and openness. They find
that ‘soft’ factors mainly play a role in retaining creative workers. What really attracts
creative movers to certain places is the job-related factors including better job
opportunities and higher wages. In studying the destination choices of engineers in the
U.S., Scott (2010) finds that the major determinant of migration for them is not amenities
but job opportunities. Aharonovitz (2011) also suggests that highly educated and talented
people are more likely to be involved in job-related migration. In sum, despite the
conflicting views on whether amenities or jobs play a decisive role in attracting talented
movers, recent empirical studies largely support jobs as the key in drawing them.
19

CHAPTER 3
DATA AND METHODOLOGY

3.1. Data
The major data source for this dissertation comes from the American Community
Survey (ACS) one-year estimates. This source offers various benefits for examining
the association between population density and the number of talented people
migrating to different places. First, the U.S. Census Bureau surveys approximately
one percent of the U.S. total population in each year, which provides a large sample
size. Second, the ACS data include information on the education levels and the
occupations of in-migrants and their demographic and socio-economic characteristics
for the years 2005 to 2011.
1
The seven-year time span provides the opportunity to
examine the potential effects of the recent economic recession. Third, data available
for sub-metropolitan geographic units in the Public Use Microdata Sample (PUMS)
files is known as Public Use Microdata Areas (PUMAs). Delineated and defined by
the U.S. Census Bureau, a PUMA is an area with population of 100,000 or more
people which can be considered proximate to a large urban community or

1
Although ACS one-year PUMS data are available from 2000 to 2011, this data series only covers the
whole U.S. since 2005. See U.S. Census Bureau. 2009. Design and Methodology: American
Community Survey. Washington, DC: U.S. Government Printing Office for more information. The
study area for this investigation includes all metropolitan/micropolitan statistical area (MSA) PUMAs
in the continental US, excluding Puerto Rico. Since Simonton (2011) suggests that highly creative
people tend to move to metropolitan areas, this study focuses on the MSA PUMAs.
20

neighborhood (DiPasquale and Kahn, 1999). Numbers of people at PUMA level,
together with land area information from the TIGER database
2
of the U.S. Census
Bureau, allow me to study the linkage between local area population density and
talent migration by constructing finer area density measurements than have typically
been investigated.
In addition to the ACS data, the MSA population and income data from the U.S.
Bureau of Economic Analysis (BEA) allow me to control for various MSA context
factors that may influence migration decisions. Since previous studies suggest that
local amenities also play a critical role in shaping the migration decisions of talented
people, this study incorporates two types of amenity variables – i.e., climate and
crime. City climate data come from the 2000 County and City Data Book. MSA crime
data are adopted from the 2005 and 2008 editions of Crime in the United States
published by the Federal Bureau of Investigation (FBI).
Dependent Variables: The outcome variable in this analysis is the number and
nature of in-migrants per ten thousand persons arriving at the MSA PUMAs in the
study year(s).
3
Several sets of variables are constructed to measure the outcome for
various groups including all in-migrants (ALL) and various talented groups in 2006

2
TIGER stands for the Topologically Integrated Geographic Encoding and Referencing system.
3
These migrants are people who changed their residences within the past one year. People who
migrated within the same PUMA are not considered in-migrants and thus are excluded. Due to the
nature of the data, this study cannot differentiate whether migrants made single or multiple moves.
21

and 2009, the years chosen for analysis and comparison. This study defines talented
people as individuals who are either highly educated or especially creative. “Highly
educated” individuals include individuals with bachelor ‐and ‐above degrees
(BAPLUS); another variable accounts for masters ‐and ‐above degrees (MAPLUS).
Following Richard Florida (2002) and others, this study also identifies two groups of
“creative” people – the “super-creative core” (SCC) and the “bohemians” (BOH). The
SCC group includes individuals in the occupations “architecture and engineering,”
“computers and math,” “life, physical, and social sciences,” “education, training, and
library,” and “arts, design, entertainment, sports, and media” and with the positions of
analyst and engineer. The BOH group includes authors, designers, musicians, actors,
directors, photographers, dancers, artists, performers, editors, technical writers, and
other media workers.
4
This research focuses on these four talent groups in light of the
hypothesis that destination area density is thought to be especially important to these
groups (Glaeser, 1999; Glaeser et al., 2001; Christensen et al., 2004; Storper and
Venables, 2004; Florida, 2009; Glaeser and Resseger, 2010). This analysis further
studies three age groups – age 25 to 34, 35 to 44, and 45 to 54 – for each of the
previous five groups for two reasons. First, previous studies show that these three age

4
There have been various debates on how to define “creative class.” Major critiques of this Richard
Florida concept include the idea that members of the “creative class” are not homogeneous and the
group includes some workers whose jobs do not involve much creativity. Although this research
defines SCC and BOH groups based on Florida’s (2002; 2005) definitions, it also takes into account
these critiques and data availability and thus the occupations included in this study may be somewhat
different from the ones suggested by Florida. For detailed types of the talented people defined in this
study, see Appendix A.
22

groups are most likely to move.
5
Second, because migratory behaviors vary across
different age groups, this classification recognizes the different impulses that impact
the migration decisions of different age groups. Data for these variables are calculated
based on the ACS Public Use Microdata Sample (PUMS) one-year estimates. Because
ACS PUMS data prior to 2006 do not include people living in group quarters (GQs),
for comparability I have excluded persons who reside in group quarters.
Density Variables: Although this study develops and tests three sets of
population density measures – MSA density, PUMA density, and the PUMA
“talented” people density, the explanatory variable of greatest interest is PUMA
population density. MSA and PUMA density variables for both 2005 and 2008 (to
account for lagged effects) are measured, using the reported area total population
divided by the total land area in square kilometers. Similarly, PUMA “talented”
people population density is constructed for both 2005 and 2008 by dividing the total
resident number of each talent group by the total land area in square kilometers.
Comparing the estimation results between MSA and PUMA densities allows me to
see whether finer density measures suggest new insights on the density-talent nexus.
Using lagged values can reduce potential endogeneity bias between number of
in-migrants and destination area density.

5
Long (1988) shows that younger people have higher mobility rates than older people. In addition,
Hansen and Niedomysl (2009) find creative class individuals are more likely to move after finishing
their college education.
23

The ACS data also allow me to construct PUMA employment density measures.
However the correlation coefficients between PUMA population density and
employment density are over 0.95 across the various groups and years. In addition,
the one-way average commute distance to/from work reported in the 2009 National
Household Travel Survey (NHTS) was 19 kilometers (11.8 miles).
6
These
magnitudes suggest that most commutes occur within a PUMA, making the choice of
density measures moot. Moreover, PUMA employment density is not available for
some PUMAs. Lastly, exchange of information and ideas as well as innovation are
not confined to just job settings (Knudsen et al., 2008; Abel et al., 2012). The
regression tests in this study therefore focus on PUMA population densities.
7

Control Variables: These tests control for several job-related and
amenity-related variables both in PUMAs and in the surrounding metropolitan areas.
Because destination area per capita income is a rough measure of prosperity and
opportunity, MSA per capita income is incorporated in various models and should
exert a positive effect on in-migrants. The percent of manufacturing employers is used
to reflect the industrial composition of a given PUMA. Some types of manufacturing
jobs such as high-tech industries in Silicon Valley and Seattle require high-skilled
workers whereas others like transportation equipment manufacturing in Detroit

6
See Table 5 in Santos et al. (2011, p. 13).
7
I have tested various regressions using PUMA employment densities. The results are consistent with
those with PUMA population densities.
24

involve relatively less-skilled workers (Chapple et al., 2004). Thereby the effect of
this variable depends on the composition of manufacturing in each of the PUMAs and
the type of talented movers migrating to these PUMAs. The PUMA
employment-population ratio represents job opportunities. More job opportunities
may attract more in-migrants. PUMA self-employed population ratio is a proxy for
local area entrepreneurship that may be particularly crucial for the talented
in-migrants. PUMA mean travel time to work is a measure for level of service of local
infrastructure. Existing studies show that people tend to move to places with shorter
commuting time (Levine, 1998; Clark et al., 2003). Longer travel time is thereby
expected to deflect in-migrants.
This study also controls for the effects of amenity variables, including January
average temperature, the number of violent crimes (per 100,000 inhabitants), and the
number of property crimes (per 100,000 inhabitants). Higher January average
temperature reflects warmer winters that should be an attracter to talented migrants. It
is expected that higher numbers of violent and property crimes should deflect talented
individuals.
When choosing a destination, migrants can be thought to consider the broader
context for the destination PUMAs. This study includes MSA size variables, which
may reflect both job and amenity opportunities. The tests also incorporate Census
25

Division and MSA dummy variables to pick up the effects of omitted context
variables. Table 3.1 lists all the variables used in the regression models and their
labels.

Table 3.1. Variable Definitions and Data Sources
Variable Description
Dependent variables
Six_ALL_Inmig_Pc Number of ALL in-migrants (per ten thousand persons) (All) in 2006
Six_BAPLUS_Inmig_Pc Number of BA+ in-migrants (per ten thousand persons) (All) in 2006
Six_MAPLUS_Inmig_Pc Number of MA+ in-migrants (per ten thousand persons) (All) in 2006
Six_SCC_Inmig_Pc Number of SCC in-migrants (per ten thousand persons) (All) in 2006
Six_BOH_Inmig_Pc Number of BOH in-migrants (per ten thousand persons) (All) in 2006
Nine_All_Inmig_Pc Number of ALL in-migrants (per ten thousand persons) (All) in 2009
Nine_BAPLUS_Inmig_Pc Number of BA+ in-migrants (per ten thousand persons) (All) in 2009
Nine_MAPLUS_Inmig_Pc Number of MA+ in-migrants (per ten thousand persons) (All) in 2009
Nine_SCC_Inmig_Pc Number of SCC in-migrants (per ten thousand persons) (All) in 2009
Nine_BOH_Inmig_Pc Number of BOH in-migrants (per ten thousand persons) (All) in 2009
Measures of density
Five_MSA_Pop_Den_BEA MSA Population Density (persons/km
2
) in 2005
Eight_MSA_Pop_Den_BEA MSA Population Density (persons/km
2
) in 2008
Five_PUMA_Pop_Den PUMA Population density (persons/km
2
) in 2005
Eight_PUMA_Pop_Den PUMA Population density (persons/km
2
) in 2008
Five_PUMA_BAPLUS_Pop_Den PUMA BA+ Population density (persons/km
2
) in 2005
Eight_PUMA_BAPLUS_Pop_Den PUMA BA+ Population density (persons/km
2
) in 2008
Five_PUMA_MAPLUS_Pop_Den PUMA MA+ Population density (persons/km
2
) in 2005
Eight_PUMA_MAPLUS_Pop_Den PUMA MA+ Population density (persons/km
2
) in 2008
Five_PUMA_SCC_Pop_Den PUMA SCC Population density (persons/km
2
) in 2005
Eight_PUMA_SCC_Pop_Den PUMA SCC Population density (persons/km
2
) in 2008
Five_PUMA_BOH_Pop_Den PUMA BOH Population density (persons/km
2
) in 2005
Eight_PUMA_BOH_Pop_Den PUMA BOH Population density (persons/km
2
) in 2008
Control variables
Five_MSA_Pop_BEA_M Dummy for medium-size MSAs (2005) (1,000,000 persons ≤ population <
5,000,000 persons)
Five_MSA_Pop_BEA_L Dummy for large-size MSAs (2005) (5,000,000 persons ≤ population)
26

Five_MSA_Inc_BEA_Ln Log of MSA Per capita personal income (dollars) (2005)
Five_Pct_M_Vo PUMA Percent of Civilian Employed Population 16 Years and Over in the
Manufacturing Industry in 2005
Five_Emp_Pop_Ratio Employment/Population Ratio for the Population 16 to 64 Years Old in 2005
Five_PUMA_Self_Emp_Pop_Ratio (PUMA Self Employed Population/PUMA Employed Population)*100 in 2005
Five_PUMA_Mttw PUMA Mean travel time to work (minutes) in 2005
Five_Violent_Crime Cases of violent crime per 100,000 inhabitants in 2005
Five_Property_Crime Cases of property crime per 100,000 inhabitants in 2005
Eight_MSA_Pop_BEA_M Dummy for medium-size MSAs (2008) (1,000,000 persons ≤ population<5,000,000
persons)
Eight_MSA_Pop_BEA_L Dummy for large-size MSAs (2008) (5,000,000 persons ≤ population)
Eight_MSA_Inc_BEA_Ln Log of MSA Per capita personal income (dollars) (2008)
Eight_Pct_M_Vo PUMA Percent of Civilian Employed Population 16 Years and Over in the
Manufacturing Industry in 2008
Eight_Emp_Pop_Ratio Employment/Population Ratio for the Population 16 to 64 Years Old in 2008
Eight_PUMA_Self_Emp_Pop_Ratio (PUMA Self Employed Population/PUMA Employed Population)*100 in 2008
Eight_PUMA_Mttw PUMA Mean travel time to work (minutes) in 2008
Eight_Violent_Crime Cases of violent crime per 100,000 inhabitants in 2008
Eight_Property_Crime Cases of property crime per 100,000 inhabitants in 2008
January_Average_Temperature Average daily temperature (degrees Fahrenheit) (for cities) in January 2000
Census Division dummies Pacific Census Division as the reference
MSA dummies Atlanta-Sandy Springs-Marietta, GA Metropolitan Statistical Area as the reference
Notes: 1. MSA population and income data are from the U.S. Bureau of Economic Analysis (BEA); data for January average
temperature at the city level are from the County and City Data Book (2000 Edition); MSA crime data are adopted
from the 2005 and 2008 editions of Crime in the United States published by the Federal Bureau of Investigation (FBI);
all other data are from the American Community Survey (ACS) published by the U.S. Census Bureau.
2. For brevity, the dependent variables for sub age-groups are not listed here. Three age groups are denoted as groups A
(25-34), B (35-44), and C (45-54). For example, Six_ALL_Inmig_Pc_A stands for the number of all in-migrants per
ten thousand persons in the age group A (25-34 years old) in 2006.

27

3.2. Background on Talent Migration
Many Americans change residences each year. Table 3.2 reports the numbers and
percentages of domestic, international, and total migrants
8
living in metropolitan
statistical area (MSA) PUMAs in non-group quarters (non-GQ) for the U.S. from
2005 to 2010. It also shows annual migrants as shares of the U.S. total population.
During this period, 42.5 to 46.8 million people one year old and over moved within
the U.S. or moved into the U.S. from other countries. The majority of these migrants
were domestic, ranging from 40.8 to 45.1 million people moving within the U.S. each
year, amounting to more than 96 percent of the total number of migrants. This also
means that each year approximately fourteen to fifteen percent of the total domestic
non-GQ population for MSA PUMAs moved to new homes within the U.S.


8
Migrants here include both in- and out-migrants. In-migrants could either move from a particular
U.S. PUMA or from another country to a particular U.S. PUMA. Out-migrants do not include movers
to other countries.
28

Table 3.2. Migrants in the U.S. (MSA Non-GQ Population)
Year 2005 2006 2007 2008 2009 2010
Total Population 286,843,924 289,681,902 291,824,061 294,074,458 296,989,516 299,581,632
Migrants
(Counts)
Domestic 43,854,715 43,730,852 41,554,662 40,844,215 40,979,619 45,095,021
Abroad 1,763,974 1,734,829 1,675,613 1,676,199 1,498,853 1,683,790
Total 45,618,689 45,465,681 43,230,275 42,520,414 42,478,472 46,778,811
Migrant
(Percentages)
Domestic 96.13% 96.18% 96.12% 96.06% 96.47% 96.40%
Abroad 3.87% 3.82% 3.88% 3.94% 3.53% 3.60%
Total 100.00% 100.00% 100.00% 100.00% 100.00% 100.00%
Migrants as
Share of Total
Population
Domestic 15.29% 15.10% 14.24% 13.89% 13.80% 15.05%
Abroad 0.61% 0.60% 0.57% 0.57% 0.50% 0.56%
Total 15.90% 15.70% 14.81% 14.46% 14.30% 15.61%
Notes: 1. Migrants include all people who changed residences, but do not include people less than 1 year old.
2. All estimates include MSA PUMAs only and do not include people in group quarters (GQs).
Source: Author’s calculations based on American Community Survey (ACS) 1-Year PUMS data.

Tables 3.3 through 3.6 report the numbers of migrants moving within and into the U.S.
for different educational attainments and selected occupational groups. Tables 3.3 and
3.4 are categorized based on educational attainment while Tables 3.5 and 3.6 are
classified by occupation. The BOH group has the highest migration rate, ranging from
sixteen to nineteen percent, followed by the groups designated as SCC, BAPLUS, and
MAPLUS. The creative class groups apparently have higher migration rates, above
the rate for the U.S. as a whole as reported in Table 3.2 and greater than the rates of
the highly educated groups. By and large, each group’s mobility had dropped after
2005 and bounced back after the onset of the slow recovery in 2009.

29

Table 3.3. BA+ Numbers of Migrants and Their Shares of Group Population (MSA Non-GQ Population)
2005 2006 2007 2008 2009 2010
Total Population 53,675,812 54,811,151 56,406,746 57,436,263 58,481,294 60,292,224
Migrants 8,311,023 8,292,125 7,987,449 7,861,202 7,903,943 8,367,075
Migrants as Share of
Group Population
15.48% 15.13% 14.16% 13.69% 13.52% 13.88%
Table 3.4. MA+ Numbers of Migrants and Their Shares of Group Population (MSA Non-GQ Population)
2005 2006 2007 2008 2009 2010
Total Population 19,012,978 19,368,275 19,960,946 20,271,970 20,809,877 21,492,231
Migrants 2,546,166 2,555,528 2,457,697 2,411,412 2,417,365 2,588,060
Migrants as Share of
Group Population
13.39% 13.19% 12.31% 11.90% 11.62% 12.04%
Table 3.5. SCC Numbers of Migrants and Their Shares of Group Population (MSA Non-GQ Population)
2005 2006 2007 2008 2009 2010
Total Population 11,540,048 11,424,317 11,630,863 11,931,325 11,938,927 11,448,108
Migrants 2,049,662 2,005,259 1,937,264 1,939,830 1,924,533 1,958,933
Migrants as Share of
Group Population
17.76% 17.55% 16.66% 16.26% 16.12% 17.11%
Table 3.6. BOH Numbers of Migrants and Their Shares of Group Population (MSA Non-GQ Population)
2005 2006 2007 2008 2009 2010
Total Population 2,686,233 2,615,619 2,677,806 2,698,998 2,693,856 2,718,948
Migrants 495,384 460,836 459,684 439,481 436,334 479,334
Migrants as Share of
Group Population
18.44% 17.62% 17.17% 16.28% 16.20% 17.63%
Notes: 1. Migrants do not include people less than 1 year old.
2. All estimates include MSA PUMAs only and do not include people in group quarters (GQs).
Source: Author’s calculations based on American Community Survey (ACS) 1-Year PUMS data.

30

3.3. Model Specifications
I examined the association between destination population densities and number of
in-migrants. The primary estimating equation is presented in equation (1). The
dependent variables are the number of in-migrants per ten thousand persons for five
groups (g) – ALL, BAPLUS, MAPLUS, SCC, and BOH – to place p in metropolitan
statistical area (MSA) m in a given calendar year t. The units of analysis in the
regressions are the metropolitan PUMAs.

In-migrants
gpt
= β
0
+ β
1
D + β
2
J + β
3
A + β
4
C + U
pmt
(1)

where D represents the explanatory variable of interest – i.e., year 2005 or 2008
population density of different types (c) for geographic area g, J is a vector of
job-related variables (i.e., MSA per capita income, percent of manufacturing jobs, job
opportunities, proportion of self-employed population, and mean travel time to work),
A controls for some consumption amenities (i.e., January average temperature and
crime),
9
C is the set of variables (i.e., MSA size, and Census Division or MSA
dummies) that reflect the broader contextual settings.

9
Except for MSA size, MSA income, and January average temperature, all of the explanatory
variables are measured at the PUMA level.
31

I test whether β
1
is different from zero. If zero cannot be rejected, this would
suggest that density matters for a given group. To address the concern of endogeneity
bias between the outcome variables and explanatory variables, except for January
average temperature and the dummy variables, all of the explanatory variables are
specified with one-year lags – i.e., all explanatory variables are either in the year2005
or 2008.
32

CHAPTER 4
DENSITY, TALENT MIGRATION, AND THE BUSINESS CYCLE

This chapter addresses the two research questions of this study. First, are denser places
especially attractive to talented people? If so, are there dense area settings that are most
attractive? Second, does density become more or less important for the talented in the
face of the recent economic downturn?
1
To investigate these questions, this chapter first
presents descriptive analyses to examine the association between PUMA population
density and talent in-migration. The chapter proceeds by developing three measures of
destination density – MSA population density, PUMA population density, and PUMA
“talented” people density – and utilizes multivariate regression tests to examine whether
the patterns suggested in the descriptive section are corroborated. Finally, the chapter
concludes with a discussion on the main findings.

4.1. Descriptive Analysis
Table 4.1 provides descriptive statistics for some major characteristics of PUMAs located
in MSAs, which are used in the following analyses.
2
MSA PUMA average land area for
the year 2005 is approximately 1,877 square kilometers. The average population in MSA
PUMAs in 2005 was 140,471.
3
The major focus of the current analysis involves the
number of in-migrants and PUMA population density. The average number of in-

1
There is a related question of whether the age of the movers matters in each case. The next chapter
investigates this question.
2
PUMAs that straddle several MSAs are excluded from this analysis.
3
Since people in non-MSA PUMAs and group quarters are excluded from the sample, the minimum
number of PUMA population in Table 4.1 is below 100,000 people.
33

migrants to MSA PUMAs in 2006 was 8,505.
4
The average PUMA residential density in
2005 was 1,459 persons per square kilometer. The coefficient of variation
5
for PUMA
residential density however indicates that there exists considerable variation across
PUMAs with respect to population density.

Table 4.1. Descriptive Statistics for MSA PUMAs Characteristics
Variable N Mean Standard
Deviation
Coefficient
of Variation
Maximum Minimum
All In-migrants (2006) 1,655 8,505 5,223 61.405 42,988 468
PUMA Population (2005) 1,655 140,471 38,072 27.103 366,830 79,040
PUMA Land Area (2005, sq km) 1,655 1,877 7,136 380.146 178,994 4
PUMA Population Density
(2005, per sq km)
1,655 1,459 3,246 222.514 39,924 0.679

The accepted wisdom is that talented people tend to concentrate in certain places such as
financial professionals in Manhattan, film producers in Hollywood, and high-tech
workers in San Jose and Boston (Route 128 high-tech corridor). Concentration is
however not a synonym for high density. While Manhattan is a relatively dense area, San
Jose is a fairly low-density place. Table 4.2 lists the top-25 receiving PUMAs ranked by
the number of 2009 BAPLUS in-migrants. PUMA population densities vary dramatically
across these PUMAs.
6
The densest receiving PUMA (in Manhattan, New York) was 427
times as dense as the least dense PUMA (in Warren County, Clarke County, Loudoun
County, and Fauquier County, Virginia), yet each one succeeded in drawing well-
educated people. These data suggest that high population density may not be strongly

4
In the following regression analyses, the number of in-migrants per 10,000 persons is used as the
dependent variable. Since migrants observed past characteristics of destination PUMAs and made their
moving decisions, this study uses one year lags for all explanatory variables except for January average
temperature and dummy variables.
5
A coefficient of variation (CV) is a normalized measure of dispersion of the variable and it is not affected
by the variable's measurement unit. CV is typically the ratio of the standard deviation to the mean but it is
sometimes represented by using the previous ratio multiplied by 100. This study uses the later approach to
compute CV .
6
Although not provided in this chapter, tables for other relocating groups across years show similar results.
34

associated with more in-migrants. The simple idea that higher densities dominate is
challenged.
Table 4.3 reports the correlations between PUMA population density and the
number of total as well as sub-group in-migrants for MSA PUMAs in the United States.
Four patterns emerge. First PUMA population density is negatively associated with the
number of total in-migrants but is positively associated with sub-groups of in-migrants.
Second the correlations between these two variables for all in-migrants are fairly weak
and are only statistically significant in 2006. The correlations between these two
variables, despite being weak, are stronger for various sub-groups. Third, within the sub-
groups, in general the coefficients for the “bohemian” group are the strongest. Fourth, the
correlations are stronger for the “bust” year 2009 than the “boom” year 2006.
In sum, data presented so far show a weak link between local area densities and the
number of in-migrants, especially for the selected groups studied. The next section
describes multivariate tests that address the question whether the associations shown here
stand up once the proper control variables are included.
35

Table 4.2. Top 25 PUMAs Receiving BAPLUS In-migrants, 2009
Rank PUMA Code Division State County
Number of
BAPLUS In-
migrants
PUMA Population
Density
1 1703502 East North Central Illinois Cook County 36,601 10,243
2 1703510 East North Central Illinois Cook County 26,297 5,876
3 1703509 East North Central Illinois Cook County 22,119 8,063
4 5100100 South Atlantic Virginia Arlington County 19,788 3,237
5 1703501 East North Central Illinois Cook County 16,460 12,045
6 5100200 South Atlantic Virginia Alexandria city 13,455 3,855
7 3603805 Middle Atlantic New York New York County 13,080 44,799
8 5100600 South Atlantic Virginia Warren County, Clarke County, Loudoun County & Fauquier County 13,075 105
9 1703503 East North Central Illinois Cook County 12,958 7,302
10 2503200 New England Massachusetts Middlesex County 12,581 6,582
11 1100105 South Atlantic District of Columbia District of Columbia 12,581 5,294
12 5100305 South Atlantic Virginia Fairfax County 12,069 962
13 4802202 West South Central Texas Denton County 11,897 178
14 3603807 Middle Atlantic New York New York County 11,107 19,590
15 2503100 New England Massachusetts Middlesex County 11,047 5,870
16 2503400 New England Massachusetts Middlesex County & Norfolk County 10,819 2,210
17 2503302 New England Massachusetts Suffolk County 10,550 6,571
18 3400601 Middle Atlantic New Jersey Hudson County 10,477 7,993
19 5100301 South Atlantic Virginia Fairfax County & Falls Church city 10,220 1,719
20 2401004 South Atlantic Maryland Montgomery County 9,933 1,186
21 4204109 Middle Atlantic Pennsylvania Philadelphia County 9,500 8,041
22 2503301 New England Massachusetts Suffolk County 9,200 8,114
23 4805304 West South Central Texas Travis County 8,719 1,044
24 1301103 South Atlantic Georgia Fulton County 8,219 1,531
25 1301201 South Atlantic Georgia DeKalb County 8,174 1,558
36

Table 4.3. Correlations between PUMA Population Density and the Number of In-Migrants (Non-GQ & MSA)
Panel A. 2006 Number of In-migrants with 2005 PUMA Population Density
2006 Number of In-migrants
ALL BAPLUS MAPLUS SCC BOH
2005 PUMA
Population Density
All -0.08270***
(N=1,655)
0.11767***
(N=1,642)
0.18664***
(N=1,538)
0.15150***
(N=1,468)
0.27881***
(N=866)
2006 Number of In-migrants / 10,000 Persons
2005 PUMA
Population Density
All -0.08009***
(N=1,655)
0.13084***
(N=1,642)
0.18936***
(N=1,538)
0.16204***
(N=1,468)
0.28625***
(N=866)
Panel B. 2009 Number of In-migrants with 2008 PUMA Population Density
2009 Number of In-migrants
ALL BAPLUS MAPLUS SCC BOH
2008 PUMA
Population Density
All -0.01202
(N=1,715)
0.18360***
(N=1,699)
0.21001***
(N=1,571)
0.23459***
(N=1,473)
0.33249***
(N=806)
2009 Number of In-migrants / 10,000 Persons
2008 PUMA
Population Density
All 0.00857
(N=1,715)
0.19952***
(N=1,699)
0.21456***
(N=1,571)
0.24472***
(N=1,473)
0.33452***
(N=806)
Notes: ***significant at 1% level; ** significant at 5% level.

4.2. Estimation Results
This study tests the explanatory power of three measures of destination density – MSA
population density, PUMA population density, and the PUMA “talented” people density.
This section reports results of multivariate regression tests to examine whether the
density-talent migration patterns suggested in the descriptive section are corroborated.
4.2.1. Results for MSA Population Density
The first tests described involve the population density of the destination metropolitan
area. Table 4.4 presents the estimates of OLS models of migration, which examine the
effects of 2005 MSA population density on all in-migrants and four talented in-migrant
groups in 2006. There are two sets of models for each of these groups that differ in the
inclusion of geographic controls – i.e., Census Division dummies versus MSA dummies.
In the Census Division fixed effects models (Pacific Division as the reference area),
MSA population density is only statistically significant for all in-migrants and for the
Bohemian group. Although “ALL” in-migrants are not especially attracted to denser
places, Bohemian movers prefer denser places. Most control variables are statistically
significant with expected signs. Talented movers prefer places with higher MSA per
37

capita incomes, lower proportion of manufacturing workers, more employment
opportunities, a higher ratio of self-employers (except the Super-Creative Core group),
and shorter commuting time to work.
When MSA dummy variables are included, although the effects of MSA population
density are statistically significant for “ALL” in-migrants and four designated talented
groups, the density effects vary across these sub-groups. Movers with master’s degrees
and those belonging to the Super-Creative Core group prefer denser places whereas
migrants with bachelor’s degrees and those belonging to the Bohemian group dislike
denser places. The mixed results are different from what most agglomeration theories
suggest but are similar to Scott’s (2010) findings – i.e., various types of talented groups
have different density preferences.
The results for control variables for MSA fixed effects models are largely consistent
with those in the Census Division fixed effects models. The only exception is that MSA
per capita income has a negative impact on the arrival of MAPLUS migrants. There are
three possible explanations for the unexpected sign for the income variable. One
possibility is that the correlation coefficient between the MSA per capita income and
MSA population density is 0.618 and a high correlation between these two variables
raises the concern of collinearity. Another possibility is that, when MSA dummy
variables are included, some unobserved MSA characteristics may be highly correlated
with MSA per capita income. Another explanation could be, as Greenwood (1997) put it,
“if consumption amenities are not included in the estimated model, the error term will
pick up their effects and be correlated with [wages or income]” (p. 677). All these
possibilities may bias the parameter estimates of the income variable.
38

Since recent studies have stressed the importance of consumption amenities for
migration decisions of talented movers, this study also tests models that incorporate
amenity variables (Table 4.5). This can also (at least partially) address the
aforementioned omitted variable bias raised by Greenwood (1997).
In the Census Division fixed effects models (Pacific Division as the reference area),
although most control variables possess expected signs, none of the density variables are
statistically significant for any talented groups. Various groups of talented people prefer
PUMAs located in medium-size
7
and high-income MSAs. They also moved to places
with more job opportunities and a higher ratio of self-employment population. They
dislike places with a higher proportion of manufacturing workers, longer commuting time
to work, and warmer winters. The results for the weather variable contradict the recent
literature that predicts people tend to move to places with nice weather (Rappaport,
2007). This perplexing finding may be due to the inclusion of Census Division dummies
that may be correlated with the weather variable (Rickman and Rickman, 2011).
Contrary to the Census Division specifications, the density variables are statistically
significant for all in-migrants and four designated talented groups in the MSA fixed
effects models. Nonetheless the effects of MSA population density vary across these sub-
groups. Migrants with bachelor or master’s degrees prefer places with a lower density
whereas movers belonging to the Super Creative Core and the Bohemian groups prefer
places with a higher density. The finding on the highly educated groups contradicts what
agglomeration theories suggest while the overall result is again similar to Scott’s (2010)

7
In this study, I define a large-size MSA as one with a population size greater than 5,000,000 persons; a
medium-size MSA as one with a population size ranging from 1,000,000 to 5,000,000 persons; a small-size
MSA as one with a population size less than 1,000,000 persons.
39

research, which found that MSA population density has mixed effects on various types of
U.S. migrant engineers.
Turning to control variables in the MSA fixed effects models, consistent with
previous migration literature, talented in-migrants prefer places with higher income,
fewer manufacturing workers, more job opportunities, more entrepreneurial
opportunities, shorter commuting time to work, and warmer winters. Several control
variables however possess unexpected signs. While both groups of the creative class
prefer safer places with lower violent and property crimes, it is perplexing that violent
crime (property crime) exerts a positive effect on migrants with bachelor’s (master’s)
degrees. Another finding contrast to the one in the Census Division fixed effects models
is that talented migrants except for people with bachelor’s degrees prefer small MSAs.
The confounding findings of these control variables may be due to the MSA dummy
variables, which might have already taken crime rates and MSA size into account.

40

Table 4.4. Regression Results for the Migration Models – MSA Population Density & All (2006) (without amenity variables)
Dependent variable
Independent variable Six_ALL_Inmig_Pc Six_BAPLUS_Inmig_Pc Six_MAPLUS_Inmig_Pc Six_SCC_Inmig_Pc Six_BOH_Inmig_Pc
Constant -983.683
(682.160)
-7513.357***
(1014.726)
-1995.761***
(357.671)
-3159.402***
(338.319)
-820.013***
(195.704)
412.964***
(155.958)
-660.659***
(148.174)
-594.592***
(115.834)
-51.778*
(27.983)
-320.142***
(26.147)
Five_MSA_Pop_Den_BEA -0.230***
(0.062)
-0.831***
(0.037)
-0.014
(0.019)
-0.154***
(0.024)
0.004
(0.010)
0.275***
(0.046)
-0.002
(0.007)
0.091**
(0.039)
0.006***
(0.002)
-0.045***
(0.010)
Five_MSA_Inc_BEA_Ln 95.915
(69.363)
786.934***
(111.772)
173.342***
(37.128)
306.416***
(41.076)
74.992***
(19.403)
-41.895***
(14.969)
62.422***
(15.036)
59.146***
(10.957)
5.248*
(2.769)
33.418***
(3.047)
Five_Pct_M_Vo -11.695***
(1.982)
-10.664**
(4.706)
-6.381***
(1.148)
-9.499***
(2.611)
-2.235***
(0.496)
-3.532***
(1.030)
-1.402***
(0.394)
-2.612***
(0.822)
-0.381***
(0.126)
-1.025***
(0.235)
Five_Emp_Pop_Ratio 9.087***
(2.356)
7.989**
(3.184)
6.297***
(0.777)
6.218***
(1.017)
1.865***
(0.337)
1.822***
(0.406)
1.395***
(0.290)
1.325***
(0.381)
0.195**
(0.078)
0.187
(0.127)
Five_PUMA_Self_Emp_Pop_Ratio -6.356***
(2.397)
-3.065
(2.121)
3.337***
(1.046)
6.881***
(1.304)
1.605***
(0.469)
3.282***
(0.626)
0.025
(0.385)
1.147**
(0.534)
0.328*
(0.180)
0.222
(0.275)
Five_PUMA_Mttw 6.994
(4.955)
-0.598
(5.697)
-3.076**
(1.247)
-7.488***
(2.163)
-1.572***
(0.436)
-3.370***
(0.966)
-1.236***
(0.383)
-2.570***
(0.824)
-0.175*
(0.093)
-0.379**
(0.178)
New_England 9.386
(31.687)
24.098
(30.813)
11.233
(15.211)
-4.672
(10.464)
0.579
(2.153)

Middle_Atlantic -141.240***
(35.085)
-8.292
(13.476)
-1.290
(6.196)
-7.820
(5.018)
1.646
(1.735)

East_North_Central 8.052
(82.960)
28.012
(23.088)
8.113
(8.597)
-0.132
(6.427)
1.093
(2.011)

West_North_Central -11.378
(39.516)
-19.799
(15.260)
-13.875**
(6.704)
-14.442**
(6.253)
1.179
(2.069)

South_Atlantic 78.809*
(40.594)
23.142
(17.198)
9.150
(8.120)
-2.468
(5.899)
-0.012
(1.606)

East_South_Central 93.468
(71.959)
34.175**
(14.862)
11.190
(7.260)
-2.537
(6.248)
1.402
(2.330)

West_South_Central 103.488***
(34.256)
15.031
(12.864)
5.401
(5.492)
-2.977
(5.033)
2.037
(1.905)

Mountain 8.454
(60.543)
-25.348
(19.378)
-8.190
(7.594)
-8.866
(7.252)
1.305
(2.004)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 402 402 402 402 391 391 374 374 255 255
N 1,655 1,655 1,642 1,642 1,538 1,538 1,468 1,468 866 866
R
2
0.2225 0.4755 0.2729 0.4415 0.2283 0.4379 0.1548 0.3625 0.0842 0.2869
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

41

Table 4.5. Regression Results for the Migration Models – MSA Population Density & All (2006) (with amenity variables)
Dependent variable
Independent variable Six_ALL_Inmig_Pc Six_BAPLUS_Inmig_Pc Six_MAPLUS_Inmig_Pc Six_SCC_Inmig_Pc Six_BOH_Inmig_Pc
Constant -696.684
(849.647)
-49371.720***
(6259.868)
-1577.429***
(383.912)
-6818.530*
(3744.668)
-697.570***
(198.246)
-12460.730***
(1923.392)
-522.253***
(157.237)
-19139.140***
(961.780)
-0.162
(33.888)
-5921.125***
(275.108)
Five_MSA_Pop_Den_BEA -0.116
(0.073)
-0.363***
(0.026)
0.004
(0.030)
-0.523***
(0.123)
0.012
(0.014)
-0.151***
(0.012)
8.740×10
-5

(0.009)
0.220***
(0.065)
0.004
(0.003)
0.300***
(0.013)
Five_MSA_Pop_BEA_M 13.728
(30.291)
-956.238***
(180.945)
31.169***
(11.189)
10.268
(87.285)
10.984**
(4.950)
-257.659***
(53.550)
10.152**
(4.335)
-250.214***
(27.939)
3.668***
(1.393)
-114.496***
(6.851)
Five_MSA_Pop_BEA_L -64.384
(40.544)
-1980.312***
(288.065)
17.067
(25.547)
-31.299
(148.818)
6.945
(13.311)
-409.464***
(89.000)
12.295
(8.628)
-707.290***
(57.068)
6.268***
(1.859)
-308.114***
(16.116)
Five_MSA_Inc_BEA_Ln 64.624
(83.607)
4532.014***
(564.400)
145.460***
(39.255)
646.411*
(350.057)
67.613***
(20.528)
1162.718***
(176.846)
52.946***
(15.924)
1793.680***
(86.116)
1.211
(3.029)
554.121***
(25.324)
Five_Pct_M_Vo -9.741***
(1.310)
-6.839***
(2.245)
-6.190***
(0.965)
-7.219***
(1.682)
-2.318***
(0.526)
-2.732***
(0.830)
-1.389***
(0.399)
-1.955***
(0.617)
-0.379***
(0.126)
-0.898***
(0.209)
Five_Emp_Pop_Ratio 10.756***
(1.872)
10.361***
(2.373)
5.878***
(0.674)
5.663***
(0.873)
1.883***
(0.351)
1.849***
(0.476)
1.359***
(0.294)
1.318***
(0.376)
0.190**
(0.087)
0.243*
(0.135)
Five_PUMA_Self_Emp_Pop_Ratio -4.893**
(2.157)
-2.395
(2.294)
4.345***
(1.000)
7.412***
(1.297)
1.967***
(0.475)
3.612***
(0.637)
0.333
(0.417)
1.354**
(0.571)
0.412**
(0.206)
0.302
(0.290)
Five_PUMA_Mttw 1.929
(2.837)
-5.685**
(2.339)
-5.077***
(1.077)
-8.975***
(2.242)
-2.232***
(0.552)
-3.850***
(1.116)
-1.896***
(0.459)
-3.050***
(0.889)
-0.360***
(0.110)
-0.444**
(0.199)
January_Average_Temperature -1.201
(1.604)
55.824***
(7.618)
-1.157*
(0.698)
-0.792
(4.283)
-0.540*
(0.320)
11.932***
(2.401)
-0.473**
(0.223)
32.484***
(1.895)
-0.086
(0.067)
10.818***
(0.509)
Five_Violent_Crime 0.012
(0.095)
0.360
(0.234)
-0.024
(0.024)
0.537**
(0.220)
-0.016
(0.012)
-0.150
(0.103)
-0.012
(0.010)
-0.632***
(0.054)
-0.006*
(0.003)
-0.149***
(0.006)
Five_Property_Crime 0.011
(0.015)
0.213***
(0.031)
-0.003
(0.006)
-0.002
(0.014)
-0.002
(0.003)
0.042***
(0.009)
3.371×10
-4

(0.002)
-0.111***
(0.005)
-2.550×10
-5

(0.001)
-0.050***
(0.002)
New_England -14.131
(55.197)
7.099
(40.791)
-0.572
(20.328)
-12.031
(13.570)
0.244
(2.501)

Middle_Atlantic -129.950**
(57.949)
-30.205
(26.546)
-15.190
(12.441)
-17.826**
(8.702)
0.189
(2.297)

East_North_Central -120.637**
(48.223)
-32.350
(20.574)
-14.383
(9.486)
-18.539**
(7.596)
-3.291
(2.232)

West_North_Central -74.349
(48.939)
-52.433**
(24.200)
-27.325**
(11.715)
-27.599***
(9.253)
-2.864
(2.447)

South_Atlantic 103.574**
(45.517)
29.229*
(14.934)
12.149*
(6.943)
-1.493
(5.220)
0.080
(1.675)

East_South_Central 77.308
(72.885)
16.910
(17.106)
3.138
(7.460)
-9.434
(7.139)
0.650
(3.044)

West_South_Central 118.334**
(50.339)
16.603
(18.350)
7.205
(8.098)
-5.498
(7.039)
0.447
(2.387)

Mountain -28.157
(61.555)
-39.491
(24.082)
-13.770
(9.830)
-14.858
(9.097)
0.041
(2.222)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 277 277 277 277 272 272 259 259 200 200
N 1,348 1,348 1,337 1,337 1,253 1,253 1,204 1,204 740 740
R
2
0.2845 0.5274 0.3158 0.4788 0.2669 0.4598 0.1778 0.3729 0.1077 0.2873
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
42

4.2.2. Results for PUMA Population Density
As mentioned in Chapter 1, most MSAs are large and their average densities mask
considerable within-MSA variation. This is why sub-metropolitan areas are the focus of
this study. Table 4.6 reports the estimation results for a migration model for 2006 in-
migrants using the finer measure of population density – at PUMA level. In the Census
Division fixed effects models testing PUMA population density, talented individuals
prefer places with higher density, higher income, lower proportion of manufacturing
workers, more job opportunities, more entrepreneurial activities (except the Super
Creative group), and shorter commuting time to work. The estimation results for MSA
fixed effects models are largely consistent with the ones found in the Census Division
fixed effects models.
Table 4.7 shows the regression results when amenity variables are added to the
migration models. Similar to the results in Table 4.6, most variables in both the Census
Division and MSA fixed effects models in Table 4.7 have the expected signs. In both the
Census Division and MSA fixed effects models, talented migrants prefer denser PUMAs.
This finding is consistent with that of Gordon and Ikeda (2011). The effects of density are
similar in both types of fixed effects models. The density effects are stronger for the
BAPLUS group compared with those of other talented groups. The effects of PUMA
population density on the in-migration of various talented people are however quite
small.
Most job-related control variables possess expected signs. Although almost no
amenity variables are statistically significant in the Census Division fixed effects models,
most of them are however significant in the MSA fixed effects models. Talented
43

individuals prefer places with warmer winters. The results for crime variables are
however mixed. Violent crime has a positive effect on movers with bachelor’s degrees
and belonging to the Bohemian group. Property crime is positively associated with
migrants with bachelor’s and master’s degrees. The unexpected signs of the crime
variables for some of the talented groups may have to do with the fact that MSA
dummies may have picked up the safety aspect of destination places. Table 4.7 also
shows the estimation results on MSA size. In the Census Division fixed effects models,
talented migrants prefer medium-size MSAs. In the MSA fixed effects models however,
talented movers favor small MSAs.

44

Table 4.6. Regression Results for the Migration Models – PUMA Population Density & All (2006) (without amenity variables)
Dependent variable
Independent variable Six_ALL_Inmig_Pc Six_BAPLUS_Inmig_Pc Six_MAPLUS_Inmig_Pc Six_SCC_Inmig_Pc Six_BOH_Inmig_Pc
Constant 716.830
(657.893)
-4014.832***
(882.412)
-1654.191***
(340.173)
-2027.970***
(287.941)
-739.863***
(170.109)
-377.122***
(126.451)
-566.679***
(131.938)
-740.321***
(111.701)
-61.216**
(25.473)
-183.292***
(40.294)
Five_PUMA_Pop_Den 0.002
(0.004)
0.008
(0.007)
0.008***
(0.003)
0.010**
(0.004)
0.004***
(0.001)
0.004**
(0.002)
0.003***
(0.001)
0.003**
(0.001)
10.939×10
-4
***
(2.396×10
-4
)
10.253×10
-4
***
(2.605×10
-4
)
Five_MSA_Inc_BEA_Ln -74.619
(68.923)
430.194***
(98.522)
135.363***
(36.127)
188.470***
(36.272)
65.329***
(16.714)
36.644**
(14.789)
51.662***
(13.539)
72.615***
(13.677)
5.513**
(2.567)
18.443***
(4.770)
Five_Pct_M_Vo -12.185***
(2.284)
-9.420**
(4.090)
-5.888***
(1.068)
-7.803***
(2.280)
-1.975***
(0.457)
-2.829***
(0.927)
-1.222***
(0.373)
-2.053***
(0.722)
-0.273**
(0.108)
-0.800***
(0.207)
Five_Emp_Pop_Ratio 11.066***
(2.167)
8.827***
(2.918)
7.376***
(0.916)
7.282***
(1.123)
2.244***
(0.370)
2.217***
(0.429)
1.728***
(0.322)
1.663***
(0.402)
0.289***
(0.082)
0.306**
(0.131)
Five_PUMA_Self_Emp_Pop_Ratio -9.135***
(2.554)
-2.701
(2.122)
2.972***
(1.022)
7.346***
(1.430)
1.530***
(0.447)
3.403***
(0.656)
-0.079
(0.381)
1.294**
(0.649)
0.354*
(0.210)
0.253
(0.316)
Five_PUMA_Mttw 3.173
(4.581)
-1.367
(5.579)
-4.395***
(1.262)
-8.489***
(2.040)
-1.943***
(0.445)
-3.681***
(0.907)
-1.585***
(0.400)
-2.818***
(0.790)
-0.203**
(0.087)
-0.432**
(0.181)
New_England 27.614
(56.927)
26.073
(32.490)
11.059
(15.408)
-4.018
(10.746)
0.631
(2.146)

Middle_Atlantic -184.892***
(61.763)
-26.904
(18.579)
-7.883
(5.891)
-13.019**
(6.398)
0.518
(1.909)

East_North_Central 10.035
(96.866)
22.756
(25.011)
5.213
(8.490)
-1.941
(7.022)
0.248
(1.869)

West_North_Central 15.747
(61.200)
-25.439
(18.701)
-17.749**
(7.209)
-16.495**
(6.963)
-0.235
(1.956)

South_Atlantic 107.518*
(61.979)
29.143
(19.910)
10.292
(8.650)
-0.721
(6.495)
0.161
(1.525)

East_South_Central 126.160
(87.338)
35.874**
(17.778)
10.134
(7.548)
-2.459
(6.802)
0.602
(2.219)

West_South_Central 131.245**
(57.795)
18.074
(15.752)
5.203
(5.890)
-2.204
(5.551)
1.741
(1.795)

Mountain 51.121
(71.751)
-24.641
(21.216)
-9.792
(7.952)
-8.759
(7.462)
0.454
(1.894)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 402 402 402 402 391 391 374 374 255 255
N 1,655 1,655 1,642 1,642 1,538 1,538 1,468 1,468 866 866
R
2
0.1984 0.4802 0.3010 0.4793 0.2619 0.4730 0.1825 0.3945 0.1348 0.3295
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

45

Table 4.7. Regression Results for the Migration Models – PUMA Population Density & All (2006) (with amenity variables)
Dependent variable
Independent variable Six_ALL_Inmig_Pc Six_BAPLUS_Inmig_Pc Six_MAPLUS_Inmig_Pc Six_SCC_Inmig_Pc Six_BOH_Inmig_Pc
Constant -329.335
(945.183)
-52807.390***
(3457.052)
-1492.693***
(360.150)
-8399.911***
(1628.047)
-689.776***
(175.415)
-9650.554***
(902.822)
-494.604***
(144.051)
-19585.460***
(1001.368)
-2.464
(31.610)
-7243.407***
(346.549)
Five_PUMA_Pop_Den 0.002
(0.002)
0.004***
(0.001)
0.007***
(0.001)
0.008***
(0.002)
0.003***
(0.001)
0.004***
(0.001)
0.002***
(4.144×10
-4
)
0.003***
(0.001)
0.001***
(1.043×10
-4
)
0.001***
(1.446×10
-4
)
Five_MSA_Pop_BEA_M 9.645
(30.739)
-1187.823***
(68.000)
30.852***
(11.211)
-178.682***
(35.976)
11.012**
(4.882)
-170.511***
(18.119)
9.811**
(4.270)
-225.687***
(19.662)
3.563***
(1.350)
-89.412***
(5.826)
Five_MSA_Pop_BEA_L -107.383***
(39.136)
-2324.862***
(127.713)
9.450
(21.534)
-296.032***
(69.909)
7.084
(11.223)
-261.707***
(34.685)
9.046
(7.664)
-674.940***
(44.604)
6.087***
(1.760)
-352.404***
(17.647)
Five_MSA_Inc_BEA_Ln 24.545
(93.409)
4827.220***
(310.093)
131.716***
(36.720)
758.776***
(151.616)
64.642***
(17.957)
893.286***
(82.899)
48.281***
(14.580)
1838.957***
(91.707)
0.620
(2.882)
689.952***
(32.469)
Five_Pct_M_Vo -9.734***
(1.486)
-6.078***
(2.309)
-5.613***
(0.953)
-5.658***
(1.349)
-2.040***
(0.503)
-2.059***
(0.712)
-1.193***
(0.390)
-1.441***
(0.514)
-0.274**
(0.107)
-0.682***
(0.163)
Five_Emp_Pop_Ratio 11.523***
(1.934)
10.903***
(2.365)
6.806***
(0.781)
6.690***
(0.914)
2.243***
(0.407)
2.259***
(0.528)
1.673***
(0.336)
1.653***
(0.411)
0.303***
(0.095)
0.368***
(0.140)
Five_PUMA_Self_Emp_Pop_Ratio -5.531***
(2.053)
-2.206
(2.296)
4.286***
(0.969)
7.771***
(1.402)
1.949***
(0.444)
3.692***
(0.669)
0.290
(0.447)
1.462**
(0.674)
0.405*
(0.236)
0.323
(0.332)
Five_PUMA_Mttw 1.065
(3.000)
-6.180**
(2.475)
-5.954***
(1.214)
-9.947***
(1.897)
-2.514***
(0.575)
-4.173***
(0.977)
-2.146***
(0.477)
-3.314***
(0.786)
-0.420***
(0.102)
-0.517***
(0.174)
January_Average_Temperature -1.768
(1.413)
55.918***
(7.893)
-1.146
(0.700)
-0.831
(3.380)
-0.482
(0.345)
11.834***
(1.984)
-0.474**
(0.229)
32.454***
(1.693)
-0.068
(0.063)
10.718***
(0.543)
Five_Violent_Crime -0.001
(0.098)
1.093
(0.711)
-0.018
(0.026)
1.192***
(0.371)
-0.013
(0.013)
-0.376*
(0.201)
-0.011
(0.010)
-0.631***
(0.053)
-0.004
(0.003)
0.246***
(0.015)
Five_Property_Crime 0.018
(0.017)
0.234***
(0.022)
-0.003
(0.006)
0.019**
(0.008)
-0.002
(0.003)
0.037***
(0.005)
1.010×10
-5

(0.002)
-0.117***
(0.005)
-8.750×10
-5

(0.001)
-0.113***
(0.005)
New_England -19.837
(51.729)
7.239
(41.031)
-0.156
(20.685)
-11.927
(13.647)
0.696
(2.579)

Middle_Atlantic -145.737**
(58.845)
-39.702
(28.251)
-18.484
(12.937)
-20.680**
(9.508)
-0.814
(2.377)

East_North_Central -129.518***
(47.246)
-39.275*
(20.428)
-16.736*
(9.406)
-20.675***
(7.624)
-3.943*
(2.184)

West_North_Central -74.174
(49.621)
-59.092**
(25.313)
-30.381**
(11.891)
-29.666***
(9.590)
-3.596
(2.441)

South_Atlantic 120.515***
(45.965)
32.504*
(16.798)
12.138
(7.769)
-0.106
(5.680)
0.257
(1.580)

East_South_Central 86.155
(73.156)
14.400
(17.603)
1.288
(7.507)
-10.054
(7.195)
0.039
(2.952)

West_South_Central 143.162***
(52.888)
19.250
(16.850)
6.145
(7.282)
-4.198
(6.689)
0.252
(2.196)

Mountain -19.255
(61.567)
-43.181*
(24.108)
-16.077*
(9.637)
-15.582*
(9.066)
-0.319
(2.143)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 277 277 277 277 272 272 259 259 200 200
N 1,348 1,348 1,337 1,337 1,253 1,253 1,204 1,204 740 740
R
2
0.2806 0.5293 0.3421 0.5125 0.2960 0.4921 0.2015 0.4014 0.1505 0.3272
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
46

4.2.3. Results for PUMA “Talented” Population Density
There are other ways to describe destination population densities. This study further tests
the effects of various PUMA “talented” population densities. Existing studies show that
talented people are attracted to places where talented individuals are already abundant
(Ritsilä and Ovaskainen, 2001; Florida, 2003; Simonton, 2011; Moretti, 2012).
8

“Talented” population density may be an attractor to talented individuals because of the
effect of knowledge spillovers that in turn leads to higher productivity and wages for
them.
The “talented” people density can be computed by taking the ratio between the
number of a particular type of talent group and the land area of a destination PUMA. For
instance, the PUMA BAPLUS population density can be obtained by using the total
number of BAPLUS population divided by the land area of a particular PUMA. Table 4.8
shows that talented migrants did prefer denser places with more talented individuals as
previous literature suggests. The effects of “talented” densities are similar between the
Census Division and MSA fixed effects models.
For the control variables that are statistically significant, all possess the expected
signs.
9
Although the share of self-employment population represents an entrepreneurial
milieu that may be conducive to creative activities, this variable is not statistically
significant for the SCC and BOH in-migrants.
When amenity variables are included in both types of fixed effects models, various
talented groups still prefer denser places where talented people cluster (Table 4.9). Most

8
A recent study by King (2011) however finds that in Canada places with more stock of talent (measured
by the proportion of individuals holding a university degree) repel the members of the creative class.
9
The correlation coefficients between four PUMA “talented” people population densities and MSA per
capita income are all below 0.25. There is no indication of multicollinearity here.
47

of them favor places with higher incomes, lower ratio of manufacturing workers, more
job opportunities, and shorter commuting time to work. Self-employment opportunities
only matter for the highly educated. As for the MSA size preferences, similar to previous
results with different density measures, when Census Division dummies are used,
talented movers prefer medium-size MSAs. For models including MSA dummies,
talented migrants are attracted to small-size MSAs. The results for amenity variables are
still perplexing as shown in the previous regressions. This may be attributed to the
inclusion of Census Division and MSA dummies.

48

Table 4.8. Regression Results for the Migration Models – “Talented” People Population Density & All (2006) (without amenity variables)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc Six_MAPLUS_Inmig_Pc Six_SCC_Inmig_Pc Six_BOH_Inmig_Pc
Constant -1612.442***
(324.020)
-2402.814***
(278.905)
-730.662***
(157.268)
-539.923***
(131.169)
-547.997***
(122.431)
-875.904***
(114.607)
-75.225***
(25.824)
-191.268***
(40.475)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den
/ Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.030**
(0.013)
0.028*
(0.015)
0.030**
(0.012)
0.026**
(0.012)
0.036**
(0.016)
0.032*
(0.017)
0.019***
(0.003)
0.017***
(0.003)
Five_MSA_Inc_BEA_Ln 137.498***
(34.279)
231.676***
(34.186)
67.357***
(15.563)
55.417***
(15.240)
51.867***
(12.564)
87.967***
(13.816)
7.762***
(2.534)
20.226***
(4.793)
Five_Pct_M_Vo -5.543***
(0.947)
-7.643***
(2.114)
-1.833***
(0.412)
-2.751***
(0.859)
-1.157***
(0.343)
-2.075***
(0.703)
-0.282***
(0.107)
-0.856***
(0.206)
Five_Emp_Pop_Ratio 6.342***
(0.713)
5.785***
(0.915)
1.758***
(0.298)
1.592***
(0.367)
1.384***
(0.274)
1.193***
(0.364)
0.151*
(0.078)
0.153
(0.124)
Five_PUMA_Self_Emp_Pop_Ratio 1.898**
(0.960)
5.967***
(1.365)
1.060**
(0.415)
2.812***
(0.622)
-0.417
(0.376)
0.847
(0.635)
0.223
(0.226)
0.068
(0.339)
Five_PUMA_Mttw -3.727***
(1.188)
-7.107***
(2.069)
-1.627***
(0.412)
-3.141***
(0.898)
-1.317***
(0.374)
-2.366***
(0.782)
-0.098
(0.085)
-0.319*
(0.174)
New_England 28.417
(30.663)
11.652
(14.497)
-3.127
(10.089)
0.624
(2.178)

Middle_Atlantic -28.882*
(16.481)
-8.543*
(5.143)
-13.389**
(5.637)
0.714
(1.959)

East_North_Central 20.566
(22.637)
4.398
(7.590)
-1.976
(6.394)
0.339
(1.956)

West_North_Central -19.416
(16.437)
-14.737**
(6.335)
-14.100**
(6.257)
0.207
(1.920)

South_Atlantic 29.102
(18.300)
10.001
(7.858)
-0.848
(5.950)
-0.356
(1.599)

East_South_Central 32.565**
(15.931)
8.508
(6.775)
-3.217
(6.217)
0.237
(2.270)

West_South_Central 16.675
(14.670)
4.733
(5.380)
-2.424
(5.259)
1.440
(1.927)

Mountain -19.055
(19.846)
-7.335
(7.375)
-7.065
(7.033)
0.661
(1.961)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 402 402 391 391 374 374 255 255
N 1,642 1,642 1,538 1,538 1,468 1,468 866 866
R
2
0.3384 0.4948 0.3120 0.4987 0.2179 0.4083 0.1397 0.3364
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
49

Table 4.9. Regression Results for the Migration Models – “Talented” People Population Density & All (2006) (with amenity variables)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc Six_MAPLUS_Inmig_Pc Six_SCC_Inmig_Pc Six_BOH_Inmig_Pc
Constant -1456.793***
(328.308)
-6228.061***
(1659.912)
-679.697***
(159.606)
-8580.045***
(760.907)
-472.175***
(132.908)
-19044.490***
(975.537)
-6.423
(34.902)
-7050.572***
(347.899)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den
/ Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.024***
(0.007)
0.022***
(0.007)
0.025***
(0.008)
0.023***
(0.007)
0.030***
(0.011)
0.027**
(0.011)
0.017***
(0.002)
0.016***
(0.002)
Five_MSA_Pop_BEA_M 29.723***
(10.607)
-128.834***
(36.212)
10.566**
(4.609)
-146.817***
(16.270)
9.359**
(4.088)
-221.966***
(19.876)
3.621**
(1.416)
-87.724***
(5.722)
Five_MSA_Pop_BEA_L 9.304
(20.023)
-220.879***
(69.973)
6.769
(10.329)
-225.823***
(31.446)
9.026
(7.242)
-659.078***
(43.988)
6.491***
(1.818)
-347.953***
(17.077)
Five_MSA_Inc_BEA_Ln 133.916***
(33.819)
567.816***
(154.739)
66.365***
(16.394)
798.631***
(70.531)
47.963***
(13.446)
1791.270***
(89.208)
1.933
(3.120)
673.312***
(32.357)
Five_Pct_M_Vo -5.355***
(0.890)
-5.716***
(1.390)
-1.911***
(0.461)
-2.056***
(0.683)
-1.131***
(0.361)
-1.490***
(0.523)
-0.276***
(0.102)
-0.728***
(0.157)
Five_Emp_Pop_Ratio 5.797***
(0.655)
5.300***
(0.853)
1.749***
(0.309)
1.631***
(0.433)
1.330***
(0.285)
1.195***
(0.360)
0.166*
(0.088)
0.202
(0.135)
Five_PUMA_Self_Emp_Pop_Ratio 3.303***
(0.959)
6.610***
(1.397)
1.489***
(0.425)
3.161***
(0.648)
-0.041
(0.457)
1.062
(0.683)
0.270
(0.256)
0.148
(0.360)
Five_PUMA_Mttw -5.196***
(1.036)
-8.647***
(2.058)
-2.150***
(0.523)
-3.626***
(1.030)
-1.852***
(0.437)
-2.846***
(0.837)
-0.324***
(0.121)
-0.371*
(0.212)
January_Average_Temperature -1.091*
(0.655)
-3.859
(3.467)
-0.447
(0.320)
10.271***
(1.783)
-0.451**
(0.210)
31.352***
(1.659)
-0.067
(0.063)
10.263***
(0.575)
Five_Violent_Crime -0.018
(0.025)
1.249***
(0.385)
-0.013
(0.012)
-0.339*
(0.199)
-0.011
(0.010)
-0.631***
(0.053)
-0.005
(0.003)
0.259***
(0.016)
Five_Property_Crime -0.003
(0.006)
0.007
(0.009)
-0.002
(0.003)
0.032***
(0.004)
4.560×10
-5

(0.002)
-0.111***
(0.005)
-1.598×10
-4

(0.001)
-0.112***
(0.005)
New_England 10.270
(39.170)
1.174
(19.661)
-10.516
(12.950)
0.750
(2.554)

Middle_Atlantic -40.205
(26.278)
-18.270
(11.922)
-20.677**
(8.731)
-1.067
(2.492)

East_North_Central -36.838*
(18.892)
-15.536*
(8.724)
-19.612***
(6.999)
-3.862*
(2.166)

West_North_Central -51.807**
(23.110)
-26.585**
(10.892)
-26.694***
(8.825)
-3.154
(2.381)

South_Atlantic 32.735**
(15.468)
12.074*
(7.068)
-0.104
(5.219)
-0.138
(1.623)

East_South_Central 13.571
(16.227)
0.700
(6.819)
-10.077
(6.730)
-0.158
(2.992)

West_South_Central 18.160
(15.935)
5.732
(6.693)
-4.429
(6.377)
-0.049
(2.276)

Mountain -37.160
(22.907)
-13.342
(8.992)
-13.653
(8.558)
0.009
(2.167)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 277 277 272 272 259 259 200 200
N 1,337 1,337 1,253 1,253 1,204 1,204 740 740
R
2
0.3700 0.5223 0.3384 0.5133 0.2310 0.4124 0.1624 0.3370
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
50

4.2.4. Results for “Boom” Year vs. “Bust” Year
Are the previous results sensitive to the phase of the recent business cycle? Table 4.10
reports the regression results for a migration model for the year 2009 using MSA
population density. In the Census Division fixed effects models, MSA population density
is only statistically significant for ALL and BOH in-migrants. ALL in-migrants prefer
places with lower densities whereas BOH in-migrants favor places with higher densities.
In the MSA fixed effects models, various talented groups all prefer denser places. Among
talented groups, MSA population density has the largest effect on BAPLUS group,
followed by MAPLUS, SCC, and BOH groups. Compared with the 2006 (“boom” year)
migration models (Table 4.4), the magnitude of the parameter estimates for MSA
population density is larger for migration models for the “bust” year (2009).
When amenity variables are included in the Census Division fixed effects models
(Table 4.11), MSA population density only matters for the BOH group. Although in the
MSA fixed effects models the density variable is statistically significant for BAPLUS,
MAPLUS, and BOH groups, it exerts different effects. Both BAPLUS and MAPLUS
migrants move toward places with lower densities while BOH movers migrate to denser
places. This migration pattern is largely in line with the one found for the year 2006 (see
Table 4.5). Nonetheless the magnitude of the parameter estimate of MSA population
density is larger for the highly educated but smaller for BOH movers in 2009. This
suggests that denser places become less attractive to these talented groups.
Table 4.12 presents the estimation results for the 2009 migration models with
PUMA population density. In both the Census Division and MSA fixed effects models,
talented migrants all prefer denser PUMAs. The effects of PUMA population density are
51

similar between two sets of fixed effects models. Unlike models with MSA population
density, there is no noticeable difference in terms of the effects of PUMA population
density between the “boom” year (2006) (see Table 4.6) and the “bust” year (2009).
Table 4.13 shows the estimation results for the 2009 migration models with PUMA
population density and additional amenity variables. Similar to the results shown in Table
4.12, in different models talented movers also prefer denser places. Other locational
characteristics talented people favor are also similar to the results shown in Table 4.12.
Compared with Table 4.7, results in Table 4.13 suggest that there is no substantial
difference in terms of density preferences over the recent business cycle for talented
individuals.
Table 4.14 displays the regression results for the 2009 migration models using
PUMA “talented” population density. The effects of PUMA “talented” population
densities are similar between Census Division and MSA fixed effects models. Talented
in-migrants prefer places where their counterparts concentrate. Like PUMA population
density models, there is no noticeable difference for the effects of PUMA “talented”
population density between the “boom” year (2006) (Table 4.8) and the “bust” year
(2009). Table 4.15 shows the estimation results with amenity variables. The results for
the density variables are similar to those shown in Table 4.14. Also there is no noticeable
difference between the “boom” year (2006) (Table 4.9) and the “bust” year (2009).
52

Table 4.10. Regression Results for the Migration Models – MSA Population Density & All (2009) (without amenity variables)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc Nine_BAPLUS_Inmig_Pc Nine_MAPLUS_Inmig_Pc Nine_SCC_Inmig_Pc Nine_BOH_Inmig_Pc
Constant -1181.459**
(561.207)
1215.880
(834.202)
-1851.289***
(319.518)
1572.569***
(208.568)
-820.732***
(183.456)
478.399***
(79.423)
-599.898***
(148.440)
522.847***
(81.408)
-30.115
(39.770)
453.706***
(50.358)
Eight_MSA_Pop_Den_BEA -0.166***
(0.041)
1.947***
(0.146)
-0.008
(0.013)
1.300***
(0.102)
-0.003
(0.007)
0.525***
(0.051)
0.006
(0.006)
0.468***
(0.029)
0.012***
(0.002)
0.060***
(0.007)
Eight_MSA_Inc_BEA_Ln 163.188***
(57.285)
-68.234
(86.128)
179.692***
(33.087)
-150.386***
(22.568)
82.231***
(18.397)
-43.873***
(7.734)
62.182***
(14.542)
-45.795***
(8.164)
5.038
(4.309)
-39.615***
(3.718)
Eight_Pct_M_Vo -14.854***
(2.706)
-16.753**
(6.456)
-7.894***
(1.578)
-12.243***
(3.457)
-2.808***
(0.666)
-4.506***
(1.329)
-2.216***
(0.553)
-3.930***
(1.074)
-0.695***
(0.228)
-1.554***
(0.573)
Eight_Emp_Pop_Ratio 3.096
(1.902)
2.623
(2.305)
3.388***
(0.785)
3.384***
(1.212)
0.854***
(0.302)
0.891**
(0.423)
0.607**
(0.253)
0.507
(0.331)
-0.032
(0.141)
0.049
(0.292)
Eight_PUMA_Self_Emp_Pop_Ratio -7.298**
(2.830)
-2.663
(2.819)
3.037**
(1.343)
6.086***
(1.406)
1.091**
(0.540)
2.482***
(0.612)
-0.134
(0.515)
0.913
(0.590)
0.196
(0.245)
0.075
(0.455)
Eight_PUMA_Mttw -0.355
(3.979)
-6.361
(4.278)
-4.341***
(1.219)
-9.278***
(1.885)
-1.955***
(0.480)
-4.073***
(0.876)
-1.675***
(0.363)
-3.478***
(0.610)
-0.254
(0.167)
-0.631**
(0.277)
New_England 92.003*
(47.436)
41.119
(33.140)
10.961
(14.147)
6.371
(11.467)
4.148
(3.069)

Middle_Atlantic -121.903***
(27.110)
-14.131
(10.666)
0.172
(4.687)
-8.663**
(3.615)
-1.257
(1.582)

East_North_Central 55.330
(72.760)
47.080*
(25.695)
18.345*
(10.154)
12.507
(7.921)
7.298**
(2.996)

West_North_Central 18.420
(37.487)
-14.906
(15.827)
-12.415*
(7.278)
-6.788
(5.788)
4.136*
(2.375)

South_Atlantic 72.226*
(37.373)
14.929
(18.931)
9.798
(8.218)
-1.256
(6.341)
-1.411
(1.936)

East_South_Central 94.432
(61.695)
30.298**
(12.096)
11.247*
(5.755)
-1.115
(4.603)
2.864
(2.034)

West_South_Central 43.507
(30.421)
-7.926
(12.109)
0.208
(5.788)
-6.160
(3.767)
0.272
(2.082)

Mountain -9.154
(43.297)
-31.031**
(15.457)
-10.635
(6.555)
-8.312
(5.648)
2.843
(2.533)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 403 403 400 400 383 383 364 364 251 251
N 1,715 1,715 1,699 1,699 1,571 1,571 1,473 1,473 806 806
R
2
0.1808 0.4289 0.2174 0.3852 0.1770 0.3589 0.1578 0.3747 0.1161 0.2974
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

53

Table 4.11. Regression Results for the Migration Models – MSA Population Density & All (2009) (with amenity variables)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc Nine_BAPLUS_Inmig_Pc Nine_MAPLUS_Inmig_Pc Nine_SCC_Inmig_Pc Nine_BOH_Inmig_Pc
Constant 365.651
(897.150)
53936.250***
(3333.689)
-1127.887***
(389.857)
-306558.500***
(81299.590)
-596.650***
(182.843)
-120794.800***
(36425.140)
-378.179**
(182.068)
-2137.972
(4985.972)
48.063
(49.655)
-346.981
(397.980)
Eight_MSA_Pop_Den_BEA -0.115*
(0.060)
-1.763***
(0.111)
0.005
(0.023)
-14.515***
(3.868)
0.002
(0.011)
-5.933***
(1.740)
0.007
(0.009)
-0.051
(0.322)
0.015***
(0.003)
0.091***
(0.026)
Eight_MSA_Pop_BEA_M 19.117
(28.105)
1429.713***
(69.405)
37.300***
(11.327)
-2979.288***
(802.694)
14.203***
(4.975)
-1139.870***
(358.162)
11.667***
(4.374)
-24.192
(37.350)
4.778***
(1.565)
-22.086**
(11.140)
Eight_MSA_Pop_BEA_L 29.407
(51.425)
1380.115***
(72.614)
45.193**
(22.760)
-2529.517***
(693.021)
15.655
(9.517)
-966.405***
(308.115)
16.202*
(8.344)
32.220
(27.656)
4.357
(2.903)
-2.159
(9.931)
Eight_MSA_Inc_BEA_Ln 37.041
(83.918)
-4858.973***
(307.951)
132.477***
(37.471)
28935.310***
(7671.692)
68.820***
(18.976)
11407.860***
(3438.195)
48.236***
(17.274)
222.962
(472.857)
-0.056
(4.573)
36.088
(38.543)
Eight_Pct_M_Vo -13.640***
(1.727)
-11.580***
(2.808)
-7.644***
(1.136)
-9.693***
(1.757)
-2.846***
(0.605)
-3.554***
(0.735)
-2.359***
(0.491)
-3.277***
(0.623)
-0.499***
(0.148)
-1.050***
(0.309)
Eight_Emp_Pop_Ratio 3.835**
(1.669)
4.374**
(1.783)
2.292***
(0.566)
2.415***
(0.640)
0.540*
(0.318)
0.663*
(0.372)
0.333
(0.232)
0.290
(0.242)
-0.237
(0.159)
-0.203
(0.242)
Eight_PUMA_Self_Emp_Pop_Ratio -4.145
(2.671)
-1.626
(2.726)
4.990***
(1.122)
6.709***
(1.093)
1.722***
(0.437)
2.657***
(0.535)
0.562
(0.461)
1.292***
(0.429)
0.559**
(0.255)
0.395
(0.369)
Eight_PUMA_Mttw -4.599**
(1.883)
-9.196***
(2.578)
-6.530***
(1.190)
-9.977***
(2.110)
-2.734***
(0.579)
-4.214***
(1.020)
-2.345***
(0.398)
-3.681***
(0.654)
-0.500***
(0.140)
-0.834***
(0.198)
January_Average_Temperature -4.026***
(1.482)
-88.646***
(5.440)
-2.182***
(6.682)
312.972***
(84.158)
-0.859***
(0.307)
122.869***
(37.685)
-0.733***
(0.240)
-0.283
(4.772)
-0.260***
(0.078)
0.481
(0.469)
Eight_Violent_Crime -0.004
(0.094)
5.444***
(0.346)
-0.052**
(0.025)
-17.025***
(4.621)
-0.031**
(0.013)
-6.689***
(2.069)
-0.023**
(0.009)
0.010
(0.253)
-0.007**
(0.004)
-0.042
(0.035)
Eight_Property_Crime -0.006
(0.018)
-0.405***
(0.027)
-0.003
(0.007)
0.467***
(0.135)
3.049×10
-4
(0.003)
0.177***
(0.060)
-0.001
(0.003)
-0.019***
(0.005)
0.001
(0.001)
0.003
(0.002)
New_England 33.231
(72.134)
15.832
(45.509)
-5.394
(20.335)
-6.139
(16.628)
2.113
(3.233)

Middle_Atlantic -198.451***
(45.705)
-57.407**
(22.571)
-17.740*
(10.593)
-24.813***
(7.677)
-5.003**
(2.253)

East_North_Central -113.911**
(42.952)
-25.802
(22.233)
-13.265
(9.812)
-10.643
(8.661)
-2.160
(2.549)

West_North_Central -73.646
(45.025)
-60.188*
(32.847)
-30.545**
(12.288)
-20.957*
(11.906)
-1.865
(3.535)

South_Atlantic 75.046**
(37.771)
19.877
(13.762)
10.414
(6.586)
-0.134
(5.367)
-0.836
(1.987)

East_South_Central 81.434
(72.422)
16.050
(17.102)
3.381
(7.249)
-5.840
(6.500)
0.516
(2.825)

West_South_Central 38.674
(48.698)
-9.702
(16.629)
-2.613
(6.979)
-8.170
(6.508)
-1.281
(2.557)

Mountain -56.089
(49.369)
-55.337**
(22.009)
-22.540**
(9.312)
-18.580**
(7.568)
0.771
(2.627)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 247 247 241 241 234 234 179 179
N 1,314 1,314 1,304 1,304 1,211 1,211 1,142 1,142 634 634
R
2
0.2550 0.5062 0.2814 0.4372 0.2238 0.3872 0.2015 0.4001 0.1449 0.2752
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
54

Table 4.12. Regression Results for the Migration Models – PUMA Population Density & All (2009) (without amenity variables)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc Nine_BAPLUS_Inmig_Pc Nine_MAPLUS_Inmig_Pc Nine_SCC_Inmig_Pc Nine_BOH_Inmig_Pc
Constant 244.092
(591.256)
-11079.000***
(473.559)
-1498.066***
(326.856)
-5334.988***
(372.774)
-680.863***
(167.066)
-1627.132***
(168.504)
-526.094***
(126.900)
-1862.909***
(120.189)
-54.595
(40.689)
-171.687***
(31.909)
Eight_PUMA_Pop_Den 0.008**
(0.003)
0.012**
(0.006)
0.009***
(0.003)
0.010**
(0.005)
0.003***
(0.001)
0.004**
(0.002)
0.003***
(0.001)
0.003***
(0.001)
11.923×10
-4
***
(3.412)
10.547×10
-4
***
(3.625×10
-4
)
Eight_MSA_Inc_BEA_Ln 23.479
(60.425)
1128.029***
(54.015)
142.411***
(34.513)
524.317***
(37.043)
67.371***
(16.866)
163.896***
(17.603)
53.674***
(12.569)
187.535***
(11.808)
6.693
(4.523)
19.862***
(4.502)
Eight_Pct_M_Vo -14.292***
(2.823)
-14.396**
(5.846)
-7.109***
(1.413)
-10.272***
(2.949)
-2.467***
(0.611)
-3.684***
(1.169)
-1.881***
(0.494)
-3.137***
(0.940)
-0.537***
(0.189)
-1.213**
(0.506)
Eight_Emp_Pop_Ratio 4.568**
(1.783)
3.799*
(2.117)
4.264***
(0.942)
4.357***
(1.390)
1.204***
(0.366)
1.269**
(0.504)
0.892***
(0.280)
0.818**
(0.351)
0.058
(0.124)
0.159
(0.254)
Eight_PUMA_Self_Emp_Pop_Ratio -9.268***
(2.728)
-2.493
(2.850)
2.639**
(1.337)
6.219***
(1.649)
0.907*
(0.541)
2.496***
(0.698)
-0.199
(0.462)
0.960*
(0.534)
0.262
(0.199)
0.070
(0.347)
Eight_PUMA_Mttw -4.335
(3.587)
-7.586*
(4.028)
-5.740***
(1.137)
-10.270***
(1.626)
-2.469***
(0.483)
-4.376***
(0.787)
-1.982***
(0.390)
-3.751***
(0.556)
-0.218
(0.146)
-0.708***
(0.268)
New_England 114.810*
(59.128)
44.813
(33.722)
12.141
(14.200)
6.859
(11.310)
3.602
(2.911)

Middle_Atlantic -157.185***
(53.142)
-31.862*
(17.346)
-7.234
(7.182)
-13.889***
(5.049)
-3.077
(1.933)

East_North_Central 54.884
(84.118)
40.876
(26.828)
15.226
(10.416)
9.679
(7.831)
5.737**
(2.778)

West_North_Central 40.110
(52.254)
-20.813
(18.602)
-14.816*
(8.051)
-10.185*
(5.931)
1.310
(2.120)

South_Atlantic 99.391*
(51.510)
21.432
(20.342)
12.140
(8.985)
0.113
(6.552)
-1.940
(1.852)

East_South_Central 118.940
(72.665)
30.629**
(14.436)
11.065*
(6.412)
-2.127
(4.599)
0.911
(2.016)

West_South_Central 71.994
(46.152)
-4.794
(13.102)
1.047
(6.032)
-6.340*
(3.660)
-1.282
(2.051)

Mountain 25.359
(51.367)
-30.291*
(16.011)
-10.399
(6.852)
-9.444*
(5.484)
0.489
(2.514)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 403 403 400 400 383 383 364 364 251 251
N 1,715 1,715 1,699 1,699 1,571 1,571 1,473 1,473 806 806
R
2
0.1735 0.4438 0.2524 0.4263 0.2096 0.3951 0.2010 0.4168 0.1677 0.3426
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

55

Table 4.13. Regression Results for the Migration Models – PUMA Population Density & All (2009) (with amenity variables)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc Nine_BAPLUS_Inmig_Pc Nine_MAPLUS_Inmig_Pc Nine_SCC_Inmig_Pc Nine_BOH_Inmig_Pc
Constant 948.794
(1003.080)
-983621.900***
(112097.400)
-998.434***
(381.764)
-353146.300***
(78177.480)
-546.459***
(167.273)
-10478.700***
(897.109)
-350.063**
(168.522)
-1203.251***
(195.635)
15.603
(54.994)
-1329.830**
(526.584)
Eight_PUMA_Pop_Den 0.006***
(0.002)
0.009***
(0.001)
0.007***
(0.001)
0.008***
(0.002)
0.003***
(4.114×10
-4
)
0.003***
(0.001)
0.002***
(3.980×10
-4
)
0.003***
(0.001)
0.001***
(1.767×10
-4
)
0.001***
(1.808×10
-4
)
Eight_MSA_Pop_BEA_M 16.032
(28.547)
-17918.970***
(2073.065)
36.035***
(11.251)
-6484.188***
(1445.288)
13.464***
(4.899)
-201.598***
(21.696)
11.030**
(4.308)
-21.121***
(7.029)
4.848***
(1.558)
-24.619***
(9.190)
Eight_MSA_Pop_BEA_L -12.792
(49.955)
-18848.870***
(2142.017)
39.188*
(20.115)
-6677.895***
(1493.053)
12.833
(8.102)
-192.566***
(17.382)
14.906**
(7.215)
29.291***
(4.392)
7.291***
(2.575)
-4.213
(8.141)
Eight_MSA_Inc_BEA_Ln -22.669
(94.831)
90668.570***
(10345.320)
115.586***
(36.802)
32584.920***
(7215.357)
62.014***
(17.578)
977.921***
(87.874)
43.662***
(15.975)
129.391***
(21.646)
1.922
(5.227)
128.472**
(50.237)
Eight_Pct_M_Vo -12.926***
(2.090)
-9.687***
(3.157)
-6.878***
(1.127)
-7.994***
(1.541)
-2.520***
(0.596)
-2.810***
(0.613)
-2.032***
(0.467)
-2.536***
(0.516)
-0.339***
(0.112)
-0.745***
(0.219)
Eight_Emp_Pop_Ratio 4.759***
(1.718)
5.467***
(1.726)
3.105***
(0.743)
3.369***
(0.830)
0.872**
(0.418)
1.039**
(0.488)
0.626**
(0.256)
0.608**
(0.254)
-0.103
(0.122)
-0.079
(0.190)
Eight_PUMA_Self_Emp_Pop_Ratio -4.470*
(2.622)
-1.645
(2.837)
4.783***
(1.180)
6.691***
(1.406)
1.613***
(0.429)
2.615***
(0.630)
0.472
(0.400)
1.281***
(0.402)
0.532***
(0.187)
0.365
(0.276)
Eight_PUMA_Mttw -6.085***
(2.038)
-10.179***
(2.123)
-7.369***
(1.291)
-10.833***
(1.828)
-3.010***
(0.621)
-4.479***
(0.924)
-2.564***
(0.464)
-3.937***
(0.580)
-0.496***
(0.150)
-0.892***
(1.198)
January_Average_Temperature -4.971***
(1.453)
1216.482***
(138.209)
-2.222***
(0.678)
432.168***
(96.477)
-0.871***
(0.323)
9.902***
(0.694)
-0.712***
(0.240)
-1.003***
(0.168)
-0.161**
(0.074)
1.434**
(0.577)
Eight_Violent_Crime -0.016
(0.100)
-47.112***
(5.587)
-0.044
(0.028)
-16.972***
(3.882)
-0.027**
(0.014)
-0.063**
(0.028)
-0.020*
(0.010)
0.119***
(0.038)
-0.003
(0.004)
-0.088**
(0.037)
Eight_Property_Crime 0.003
(0.020)
3.965***
(0.450)
-0.002
(0.007)
1.370***
(0.312)
0.001
(0.003)
0.013***
(0.002)
-0.001
(0.003)
-0.020***
(0.003)
1.962×10
-4

(0.001)
0.003*
(0.002)
New_England 22.961
(68.352)
15.856
(45.383)
-5.479
(20.468)
-5.602
(16.734)
3.545
(3.678)

Middle_Atlantic -228.276***
(49.524)
-68.447***
(24.129)
-22.287*
(11.871)
-28.059***
(8.488)
-4.855**
(2.316)

East_North_Central -138.782***
(48.803)
-35.806
(22.266)
-17.204*
(10.037)
-13.844
(8.774)
-2.286
(2.659)

West_North_Central -81.356*
(46.355)
-69.281**
(35.058)
-34.078
(13.170)
-24.651*
(12.544)
-3.645
(3.479)

South_Atlantic 95.403**
(41.023)
22.880
(15.412)
11.578
(7.192)
0.487
(5.825)
-2.468
(2.041)

East_South_Central 81.064
(73.113)
10.673
(17.895)
1.269
(7.511)
-7.894
(6.658)
-1.419
(2.899)

West_South_Central 61.397
(49.661)
-9.727
(15.916)
-2.688
(6.482)
-9.039
(6.287)
-4.233
(2.626)

Mountain -47.851
(48.394)
-59.413***
(22.388)
-23.974**
(9.298)
-20.153***
(7.555)
-1.376
(2.996)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 247 247 241 241 234 234 179 179
N 1,314 1,314 1,304 1,304 1,211 1,211 1,142 1,142 634 634
R
2
0.2558 0.5171 0.3115 0.4734 0.2500 0.4193 0.2363 0.4384 0.1831 0.3190
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.
56

Table 4.14. Regression Results for the Migration Models – “Talented” People Population Density & All (2009) (without amenity variables)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc Nine_MAPLUS_Inmig_Pc Nine_SCC_Inmig_Pc Nine_BOH_Inmig_Pc
Constant -1424.358***
(315.364)
-5073.918***
(326.709)
-653.431***
(152.486)
-1536.814***
(153.730)
-492.906***
(114.962)
-1780.654***
(105.884)
-74.158**
(35.428)
-132.894***
(28.424)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
0.033**
(0.016)
0.031*
(0.018)
0.029**
(0.013)
0.027**
(0.014)
0.044**
(0.017)
0.039**
(0.018)
0.021***
(0.004)
0.018***
(0.004)
Eight_MSA_Inc_BEA_Ln 139.178***
(33.019)
503.625***
(34.617)
66.311***
(15.371)
157.124***
(16.747)
51.480***
(11.373)
180.596***
(10.511)
9.101**
(3.936)
16.631***
(3.980)
Eight_Pct_M_Vo -6.654***
(1.222)
-9.994***
(2.671)
-2.292***
(0.549)
-3.584***
(1.087)
-1.773***
(0.455)
-3.103***
(0.907)
-0.578***
(0.202)
-1.350**
(0.531)
Eight_Emp_Pop_Ratio 3.512***
(0.745)
3.273***
(1.106)
0.895***
(0.285)
0.822**
(0.396)
0.663***
(0.240)
0.516*
(0.295)
-0.041
(0.131)
0.069
(0.266)
Eight_PUMA_Self_Emp_Pop_Ratio 1.330
(1.330)
4.553**
(1.810)
0.344
(0.507)
1.772**
(0.722)
-0.653
(0.425)
0.340
(0.602)
0.082
(0.186)
-0.175
(0.345)
Eight_PUMA_Mttw -4.815***
(1.077)
-8.554***
(1.765)
-2.046***
(0.446)
-3.668***
(0.817)
-1.577***
(0.335)
-3.086***
(0.587)
-0.044
(0.141)
-0.489*
(0.267)
New_England 45.443
(31.751)
11.875
(13.287)
7.294
(10.520)
3.376
(2.966)

Middle_Atlantic -36.567**
(16.200)
-9.638
(6.568)
-15.079***
(4.748)
-3.014*
(1.766)

East_North_Central 36.501
(23.894)
13.378
(9.384)
9.106
(7.237)
5.812**
(2.826)

West_North_Central -15.736
(15.962)
-12.780*
(6.952)
-7.863
(5.155)
1.683
(2.179)

South_Atlantic 20.392
(18.343)
11.208
(7.987)
0.071
(5.829)
-2.528
(1.934)

East_South_Central 26.638**
(12.554)
9.124*
(5.513)
-2.853
(4.126)
0.555
(2.096)

West_South_Central -4.619
(12.080)
0.962
(5.547)
-5.957*
(3.348)
-1.672
(2.208)

Mountain -25.662*
(14.536)
-8.775
(6.064)
-7.759
(5.161)
0.767
(2.646)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 400 400 383 383 364 364 251 251
N 1,699 1,699 1,571 1,571 1,473 1,473 806 806
R
2
0.2993 0.4511 0.2652 0.4269 0.2567 0.4458 0.1692 0.3454
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

57

Table 4.15. Regression Results for the Migration Models – “Talented” People Population Density & All (2009) (with amenity variables)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc Nine_MAPLUS_Inmig_Pc Nine_SCC_Inmig_Pc Nine_BOH_Inmig_Pc
Constant -956.081***
(344.458)
-315943.300***
(79194.320)
-524.515***
(147.669)
-9706.549***
(1213.200)
-324.014**
(151.639)
-1175.343***
(189.558)
0.661
(57.022)
-1698.437***
(603.407)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
0.024***
(0.007)
0.023***
(0.008)
0.023***
(0.007)
0.022***
(0.007)
0.037***
(0.011)
0.033***
(0.011)
0.018***
(0.002)
0.016***
(0.002)
Eight_MSA_Pop_BEA_M 34.440***
(10.592)
-5793.888***
(1463.772)
12.709***
(4.592)
-183.898***
(27.050)
10.116**
(3.957)
-17.303**
(7.161)
4.781***
(1.499)
-30.773***
(10.487)
Eight_MSA_Pop_BEA_L 38.084**
(18.744)
-5957.383***
(1512.502)
12.403*
(7.483)
-173.176***
(22.791)
14.144**
(6.557)
33.024***
(4.891)
7.536***
(2.429)
-10.163
(9.648)
Eight_MSA_Inc_BEA_Ln 114.041***
(33.175)
29161.200***
(7309.166)
60.832***
(15.503)
908.636***
(11.845)
41.517***
(14.402)
127.559***
(20.535)
3.802
(5.400)
163.285***
(57.443)
Eight_Pct_M_Vo -6.492***
(1.034)
-7.907***
(1.487)
-2.323***
(0.546)
-2.723***
(0.579)
-1.884***
(0.422)
-2.492***
(0.483)
-0.360***
(0.112)
-0.848***
(0.224)
Eight_Emp_Pop_Ratio 2.447***
(0.565)
2.425***
(0.631)
0.603*
(0.309)
0.634*
(0.362)
0.437*
(0.225)
0.335
(0.224)
-0.199
(0.147)
-0.170
(0.219)
Eight_PUMA_Self_Emp_Pop_Ratio 3.612***
(1.241)
5.384***
(1.506)
1.052**
(0.455)
1.991***
(0.665)
-0.024
(0.387)
0.707
(0.458)
0.342**
(0.163)
0.138
(0.246)
Eight_PUMA_Mttw -6.434***
(1.130)
-9.357***
(1.972)
-2.600***
(0.552)
-3.829***
(0.971)
-2.147***
(0.365)
-3.293***
(0.644)
-0.339**
(0.134)
-0.672***
(0.226)
January_Average_Temperature -2.099***
(0.630)
385.905***
(97.747)
-0.800***
(0.299)
9.059***
(1.100)
-0.646***
(0.217)
-0.888***
(0.153)
-0.139*
(0.073)
1.794***
(0.642)
Eight_Violent_Crime -0.044*
(0.026)
-15.178***
(3.929)
-0.027**
(0.013)
-0.067**
(0.031)
-0.020**
(0.009)
0.077*
(0.041)
-0.004
(0.004)
-0.113***
(0.043)
Eight_Property_Crime -0.002
(0.007)
1.221***
(0.316)
0.001
(0.003)
0.012***
(0.002)
-0.001
(0.002)
-0.017***
(0.003)
2.567×10
-4

(0.001)
0.005**
(0.002)
New_England 18.842
(43.399)
-4.076
(19.515)
-3.957
(15.726)
3.879
(3.690)

Middle_Atlantic -69.695***
(22.336)
-23.253**
(11.073)
-28.334***
(7.724)
-4.760**
(2.170)

East_North_Central -34.373*
(20.503)
-16.773*
(9.251)
-12.925
(7.926)
-1.762
(2.583)

West_North_Central -62.221*
(31.665)
-30.811***
(11.638)
-21.252*
(11.320)
-2.706
(3.376)

South_Atlantic 22.432
(14.091)
11.024*
(6.604)
0.553
(5.268)
-2.901
(2.059)

East_South_Central 9.817
(16.517)
0.757
(6.843)
-8.030
(5.902)
-1.472
(2.819)

West_South_Central -9.237
(14.980)
-2.684
(6.006)
-8.740
(5.805)
-4.583*
(2.685)

Mountain -53.800
(20.668)
-21.526**
(8.357)
-17.964***
(6.783)
-0.752
(3.139)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 247 247 241 241 234 234 179 179
N 1,304 1,304 1,211 1,211 1,142 1,142 634 634
R
2
0.3469 0.4918 0.3007 0.4507 0.2909 0.4697 0.1941 0.3278
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.
58

4.3. Discussion
This chapter reports results of tests that address the two research questions of this study.
First, are denser places especially attractive to talented people? The correlation
coefficients between PUMA population density and talented in-migrants are positive but
weak. This suggests that density, even at the smaller area (PUMA) level, plays only a
moderate role in drawing talented individuals.
This study also describes results of tests that utilize three different density measures
– MSA population density, PUMA population density, and PUMA “talented” people
population density – in multivariate tests while controlling for job- and amenity-related
variables to answer this question. In general, the estimation results show that talented
individuals do prefer denser places (Table 4.16). This finding is consistent with the
predictions of urban agglomeration theory. The effects of various density measures are
however weak, consistent with the results from the simple correlation coefficients. This is
particularly true when density measures at PUMA level are utilized.

Table 4.16. Density Effects on Talented Groups
Year Density
Measure
Without amenity variables With amenity variables
Census Division MSA Census Division MSA
2006 MSA BOH(+) MAPLUS&SCC(+);
BAPLUS&BOH(-)
SCC&BOH(+);
BAPLUS&MAPLUS(-)
PUMA All talent groups(+) All talent groups(+) All talent groups(+) All talent groups(+)
Talented People All talent groups(+) All talent groups(+) All talent groups(+) All talent groups(+)
2009 MSA BOH(+) All talent groups(+) BOH(+) BOH(+);
BAPLUS&MAPLUS(-)
PUMA All talent groups(+) All talent groups(+) All talent groups(+) All talent groups(+)
Talented People All talent groups(+) All talent groups(+) All talent groups(+) All talent groups(+)

There are three noticeable differences between the models with MSA population density
and PUMA population density. First, compared with the estimation results using MSA
population density, the results from PUMA population density are consistent across
models and sub-groups – i.e., talented in-migrants favor denser places, which is in line
59

with the findings of Gordon and Ikeda (2011). Second, the magnitude of parameter
estimates for MSA population density is larger than the ones for PUMA population
density. Third, the explanatory power for models using PUMA population density is
higher than models using MSA population density.
As mentioned earlier, in the fields of urban planning and urban economics, most
researchers use density measures that are calculated based on large geographic areas. By
comparing the results using MSA and PUMA densities, this study demonstrates that the
commonly-used density measures such as MSA population density are much too general
and much too vague. Density measures based on large geography may contain
“significant noise” (Gordon, 2013, p. 672) and thus fail to reveal the role of population
density really played in determining the migration patterns of various talented groups.
Although we do know that talented migrants prefer to live at higher densities (when
it is measured at the PUMA level), are there dense area settings that are most attractive to
them? Or put differently, can we find “optimal” densities for various talented
individuals? Table 4.2, which lists the top-25 PUMAs receiving BAPLUS in-migrants in
2009, shows that both high- and low-density PUMAs successfully attract BAPLUS
migrants. In other words, various talented groups migrate to different places with various
levels of densities that meet their needs and preferences. This suggests that there are no
simply formulas for densities that talented individuals particularly prefer.
The second question this chapter tries to address is: Does density become more or
less important for the talented in the face of the recent economic downturn? Correlation
coefficients do show that PUMA population density is relatively stronger in the “bust”
year (2009). This is in line with the regression results using MSA population density
60

(without amenities variables included): Compared with the 2006 (“boom” year) migration
models, the effects of MSA population density are larger for migration models for the
“bust” year (2009). However, when PUMA population density and PUMA “talented”
people density are used, the difference between the “boom” and “bust” years is not
salient. Regardless of the economic situation, various talented movers seemingly prefer
denser places, other things equal. This suggests that the migration preferences of the
talented movers do not vary significantly between the “boom” year (2006) and the “bust”
year (2009).

61

CHAPTER 5
DENSITY, TALENT MIGRATION, AND LIFE CYCLE

This chapter focuses on in-migrants by the three age groups mentioned previously, 25-34,
35-44, and 45-54. In addition to addressing the two research questions posed in the
previous chapter, this chapter addresses the following question: does the density-talent
migration nexus vary with age, educational or occupational groups?

5.1. Descriptive Analysis
Table 5.1 presents correlations between PUMA population density and the number of
total and talented in-migrants per ten thousand persons for three age groups for the MSA
PUMAs. Several patterns emerge. First, the correlations between PUMA population
density and in-migrants are weak and sometimes not statistically significant. Second, the
coefficients are a decreasing function of age – i.e., the older the group, the lower the
coefficients. Third, for the youngest age group (25-34) the correlations are stronger for
the “bust” year of 2009 than the “boom” year of 2006. This is however not always the
case for the 35-44 and 45-54 age groups.

62

Table 5.1. Correlations between PUMA Population Density and the Number of In-Migrants (Non-GQ & MSA)
Panel A. 2006 Number of In-migrants with 2005 PUMA Population Density
2006 Number of In-migrants / 10,000 Persons
ALL BAPLUS MAPLUS SCC BOH
2005 PUMA
Population
Density
25-34 0.04317*
(N=1,647)
0.17524***
(N=1,557)
0.21968***
(N=1,226)
0.21094***
(N=1,120)
0.27586***
(N=465)
35-44 -0.00812
(N=1,632)
0.12198***
(N=1,412)
0.17657***
(N=1,056)
0.10017***
(N=797)
0.23248***
(N=253)
45-54 -0.07319***
(N=1,610)
0.09449***
(N=1,284)
0.17881***
(N=887)
0.04695
(N=615)
0.11506
(N=172)
Panel B. 2009 Number of In-migrants with 2008 PUMA Population Density
2009 Number of In-migrants / 10,000 Persons
ALL BAPLUS MAPLUS SCC BOH
2008 PUMA
Population
Density
25-34 0.16252***
(N= 1,708)
0.24297***
(N=1,614)
0.26840***
(N=1,251)
0.22773***
(N=1,129)
0.29464***
(N=424)
35-44 -0.00761
(N=1,679)
0.11536***
(N=1,395)
0.17591***
(N=1,014)
0.16213***
(N=805)
0.13115**
(N=231)
45-54 -0.02684
(N=1,648)
0.09097***
(N=1,236)
0.14120***
(N=799)
0.08054*
(N=546)
0.03653
(N=139)
Notes: ***significant at 1% level; **significant at 5% level; *significant at 10% level.

5.2. Estimation Results
5.2.1. Results for Metropolitan Area Population Density
Table 5.2 and Table 5.3 (with amenity variables) report parameter estimates for 2005
MSA population density from migration regression models for 2006 ALL and talented
in-migrants for the three age groups. In the Census Division fixed effects models, only a
few parameter estimates for the density variable are statistically significant. Among
talented groups, the density variable is statistically significant for MAPLUS and BOH
63

in-migrants belonging to the 45-54 age group and they favor denser places. In the MSA
fixed effects models, the effects of the density variable are mixed across various age,
educational, and occupational groups.
Table 5.4 and Table 5.5 (with amenity variables) also report the parameter estimates
for 2008 MSA population density from migration regression models for 2009 ALL and
talented in-migrants for three age groups. In the Census Division fixed effects models,
only a few parameter estimates are statistically significant. Among the talented groups,
BAPLUS and BOH groups belonging to the 25-34 age group prefer to live in denser
places while MAPLUS in-migrants belonging to the 45-54 age group dislike denser
places. In the MSA fixed effects models, the effects of the density variable is also
sensitive to the model specification – i.e., whether amenity variables are included or not.
The volatility of estimation results may be due to the fact that MSA population density
may have captured some qualities reflected in the amenity variables, and Census Division
and MSA dummies.
1


1
For complete regression results, see Appendices B and C.
64

Table 5.2. Parameter Estimates for 2005 MSA Population Density (without amenity variables)
2006 ALL 2006 BAPLUS 2006 MAPLUS 2006 SCC 2006 BOH
Group A (Age 25-34) -0.041**
(0.017)
-0.165***
(0.011)
-0.001
(0.010)
-0.048***
(0.011)
0.005
(0.005)
0.273***
(0.046)
0.001
(0.003)
-0.007
(0.005)
0.003
(0.002)
0.007***
(0.002)
N 1,647 1,647 1,557 1,557 1,226 1,226 1,120 1,120 465 465
R
2
0.1657 0.3885 0.1991 0.3761 0.1591 0.3437 0.1031 0.2839 0.0655 0.2483
Group B (Age 35-44) -0.025**
(0.011)
-0.145***
(0.006)
0.002
(0.005)
-0.035***
(0.006)
-0.002
(0.003)
0.289***
(0.003)
-0.001
(0.002)
-0.012***
(0.002)
-0.001
(0.002)
0.001
(0.002)
N 1,632 1,632 1,412 1,412 1,056 1,056 797 797 253 253
R
2
0.1802 0.4038 0.1990 0.3722 0.1552 0.3785 0.0787 0.3113 0.0629 0.4692
Group C (Age 45-54) -0.022***
(0.006)
-0.091***
(0.005)
-0.002
(0.003)
-0.024***
(0.002)
0.004**
(0.002)
-0.003
(0.002)
3.232×10
-4

(18.277×10
-4
)
-0.003*
(0.002)
0.004***
(0.001)
0.006***
(0.002)
N 1,610 1,610 1,284 1,284 887 887 615 615 172 172
R
2
0.1373 0.3941 0.0968 0.3342 0.0782 0.3716 0.0729 0.3516 0.2071 0.4557
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

65

Table 5.3. Parameter Estimates for 2005 MSA Population Density (with amenity variables)
2006 ALL 2006 BAPLUS 2006 MAPLUS 2006 SCC 2006 BOH
Group A (Age 25-34) -0.014
(0.022)
-0.594***
(0.091)
0.008
(0.013)
-0.726
(0.611)
0.003
(0.007)
-0.024***
(0.009)
0.002
(0.004)
0.464***
(0.171)
0.004
(0.003)
-0.463***
(0.078)
N 1,342 1,342 1,273 1,273 1,019 1,019 947 947 406 406
R
2
0.2084 0.4020 0.2353 0.3900 0.1974 0.3497 0.1253 0.2922 0.0779 0.2362
Group B (Age 35-44) -0.015
(0.013)
2.659***
(0.350)
0.006
(0.006)
-2.888***
(0.257)
0.002
(0.004)
-0.260***
(0.089)
0.002
(0.002)
-0.015
(0.014)
-0.002
(0.003)
-0.031*
(0.018)
N 1,330 1,330 1,165 1,165 886 886 669 669 220 220
R
2
0.2039 0.4110 0.2094 0.3754 0.1837 0.3913 0.1120 0.3048 0.1052 0.4258
Group C (Age 45-54) -0.010
(0.007)
0.088
(0.242)
-0.002
(0.003)
0.413***
(0.023)
0.005**
(0.002)
0.184**
(0.071)
-0.001
(0.002)
0.118
(0.184)
0.004**
(0.002)
0.633**
(0.284)
N 1,313 1,313 1,057 1,057 737 737 512 512 140 140
R
2
0.1705 0.4039 0.1104 0.3008 0.1158 0.3341 0.0938 0.3236 0.3149 0.4862
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

66

Table 5.4. Parameter Estimates for 2008 MSA Population Density (without amenity variables)
2009 ALL 2009 BAPLUS 2009 MAPLUS 2009 SCC 2009 BOH
Group A (Age 25-34) -0.019*
(0.011)
0.678***
(0.063)
0.013*
(0.008)
0.624***
(0.059)
0.006
(0.004)
0.233***
(0.033)
0.006*
(0.003)
0.256***
(0.021)
0.008***
(0.002)
0.036***
(0.004)
N 1,708 1,708 1,614 1,614 1,251 1,251 1,129 1,129 424 424
R
2
0.1409 0.3320 0.1490 0.3017 0.1343 0.2856 0.1417 0.3090 0.1150 0.3457
Group B (Age 35-44) -0.020**
(0.008)
0.449***
(0.030)
-0.002
(0.004)
0.216***
(0.021)
-0.001
(0.002)
0.039***
(0.003)
0.001
(0.002)
0.063***
(0.004)
0.002
(0.002)
0.003*
(0.002)
N 1,679 1,679 1,395 1,395 1,014 1,014 805 805 231 231
R
2
0.1248 0.3538 0.1666 0.3525 0.1378 0.3624 0.1052 0.3506 0.0976 0.4934
Group C (Age 45-54) -0.016***
(0.004)
0.168***
(0.016)
-0.003*
(0.002)
0.074***
(0.008)
-0.004**
(0.001)
0.048***
(0.004)
-0.001
(0.001)
0.015***
(0.005)
-4.060×10
-4

(17.191×10
-4
)
-0.003
(0.007)
N 1,648 1,648 1,236 1,236 799 799 546 546 139 139
R
2
0.1102 0.3346 0.1174 0.3208 0.0881 0.3507 0.0463 0.3234 0.1648 0.7373
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

67

Table 5.5. Parameter Estimates for 2008 MSA Population Density (with amenity variables)
2009 ALL 2009 BAPLUS 2009 MAPLUS 2009 SCC 2009 BOH
Group A (Age 25-34) 0.002
(0.016)
-3.587***
(0.247)
0.024**
(0.012)
-7.123***
(2.274)
0.010
(0.006)
0.069
(0.046)
0.007
(0.005)
0.090***
(0.022)
0.012***
(0.003)
0.057
(0.035)
N 1,308 1,308 1,244 1,244 973 973 892 892 335 335
R
2
0.1944 0.3867 0.2094 0.3497 0.1804 0.3117 0.1864 0.3367 0.1756 0.3267
Group B (Age 35-44) -0.015
(0.011)
-0.639***
(0.036)
0.001
(0.006)
-0.073
(0.059)
0.002
(0.004)
-0.694***
(0.095)
0.002
(0.003)
0.078**
(0.033)
0.005
(0.003)
0.013
(0.031)
N 1,291 1,291 1,093 1,093 811 811 644 644 195 195
R
2
0.1648 0.3965 0.1937 0.3640 0.1700 0.3719 0.1195 0.3604 0.1417 0.4834
Group C (Age 45-54) -0.017***
(0.007)
-4.236***
(0.559)
-0.002
(0.002)
0.007
(0.010)
-0.003*
(0.002)
-0.150***
(0.017)
0.001
(0.002)
0.096***
(0.033)
0.003
(0.003)
0.320***
(0.060)
N 1,264 1,264 967 967 643 643 442 442 109 109
R
2
0.1493 0.3451 0.1471 0.3227 0.1096 0.3372 0.0564 0.2938 0.2045 0.7309
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

68

5.2.2. Results for PUMA Population Density
Table 5.6 and Table 5.7 (with amenity variables) report the parameter estimates for 2005
PUMA population density from the migration regression models for 2006 ALL and
talented in-migrants for three age groups. In both the Census Division and MSA fixed
effects models, most parameter estimates are statistically significant and positively
associated with the number of in-migrants. The magnitude for parameter estimates of
PUMA population density for BAPLUS and MAPLUS in-migrants are greater than that
of SCC and BOH groups. The density variable is however not statistically significant for
the SCC and BOH (MSA fixed effects model without amenity variables) in-migrants
belonging to the 45-54 age group. In addition, the magnitude of the density variable
decreases as we move up the age ladder. Table 5.8 and Table 5.9 (with amenity variables)
present the parameter estimates for 2008 PUMA population density from migration
regression models for 2009 ALL and talented in-migrants for three age groups. The
findings are consistent with the ones shown in Table 5.6 and 5.7. In sum, these findings
suggest that when PUMA population density is used most talented migrants do prefer
denser places. However, the effect of the density variable becomes weaker when these
talented individuals (SCC and BOH in particular) get older.
2

2
For complete regression results, see Appendices D and E.
69

Table 5.6. Parameter Estimates for 2005 PUMA Population Density (without amenity variables)
2006 ALL 2006 BAPLUS 2006 MAPLUS 2006 SCC 2006 BOH
Group A (Age 25-34) 0.004**
(0.001)
0.005**
(0.003)
0.005***
(0.002)
0.005**
(0.002)
0.002***
(0.001)
0.002**
(0.001)
18.497×10
-4
***
(4.682×10
-4
)
20.045×10
-4
***
(6.436×10
-4
)
5.714×10
-4
***
(1.398×10
-4
)
5.246×10
-4
***
(1.744×10
-4
)
N 1,647 1,647 1,557 1,557 1,226 1,226 1,120 1,120 465 465
R
2
0.1698 0.4104 0.2362 0.4188 0.1990 0.3821 0.1430 0.3246 0.1020 0.2760
Group B (Age 35-44) 3.482×10
-4

(6.342×10
-4
)
0.001
(0.001)
11.835×10
-4
**
(4.617×10
-4
)
14.740×10
-4
**
(6.167×10
-4
)
8.116×10
-4
***
(2.041×10
-4
)
9.657×10
-4
***
(2.836×10
-4
)
4.515×10
-4
**
(1.910×10
-4
)
4.738×10
-4
**
(2.343×10
-4
)
3.269×10
-4
***
(0.624×10
-4
)
3.787×10
-4
***
(0.694×10
-4
)
N 1,632 1,632 1,412 1,412 1,056 1,056 797 797 253 253
R
2
0.1722 0.4070 0.2102 0.3870 0.1710 0.3985 0.0879 0.3199 0.0895 0.4993
Group C (Age 45-54) -0.501×10
-4

(3.704×10
-4
)
5.631×10
-4

(5.420×10
-4
)
6.594×10
-4
***
(1.654×10
-4
)
7.535×10
-4
***
(2.084×10
-4
)
6.137×10
-4
***
(1.873×10
-4
)
5.349×10
-4
***
(1.971×10
-4
)
1.713×10
-4

(1.436×10
-4
)
0.113×10
-4

(1.127×10
-4
)
1.935×10
-4
*
(1.136×10
-4
)
0.344
(2.546)
N 1,610 1,610 1,284 1,284 887 887 615 615 172 172
R
2
0.1230 0.3956 0.1068 0.3458 0.0967 0.3862 0.0752 0.3516 0.1992 0.4560
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

70

Table 5.7. Parameter Estimates for 2005 PUMA Population Density (with amenity variables)
2006 ALL 2006 BAPLUS 2006 MAPLUS 2006 SCC 2006 BOH
Group A (Age 25-34) 0.003***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.002***
(0.001)
0.002***
(0.001)
0.002***
(2.807×10
-4
)
0.002***
(3.883×10
-4
)
4.872×10
-4
***
(7.630×10
-5
)
4.658×10
-4
***
(1.075×10
-4
)
N 1,342 1,342 1,273 1,273 1,019 1,019 947 947 406 406
R
2
0.2199 0.4188 0.2684 0.4281 0.2316 0.3859 0.1618 0.3305 0.1040 0.2590
Group B (Age 35-44) 2.289×10
-4

(2.590×10
-4
)
6.455×10
-4

(2.483×10
-4
)
9.454×10
-4
***
(1.302×10
-4
)
0.001***
(2.126×10
-4
)
0.001***
(9.430×10
-5
)
0.001***
(1.391×10
-4
)
3.833×10
-4
***
(9.010×10
-5
)
3.828×10
-4
***
(9.340×10
-5
)
3.506×10
-4
***
(6.190×10
-5
)
4.328×10
-4
***
(5.400×10
-5
)
N 1,330 1,330 1,165 1,165 886 886 669 669 220 220
R
2
0.2021 0.4121 0.2171 0.3876 0.1978 0.4098 0.1186 0.3110 0.1352 0.4689
Group C (Age 45-54) -3.830×10
-5

(2.683×10
-4
)
2.359×10
-4

(2.050×10
-4
)
5.209×10
-4
***
(9.530×10
-5
)
0.001***
(6.700×10
-5
)
4.733×10
-4
***
(7.250×10
-5
)
4.432×10
-4
***
(7.810×10
-5
)
8.430×10
-5

(9.460×10
-5
)
1.350×10
-5

(1.199×10
-4
)
2.316×10
-4
***
(5.120×10
-5
)
1.853×10
-4
***
(5.530×10
-5
)
N 1,313 1,313 1,057 1,057 737 737 512 512 140 140
R
2
0.1687 0.4042 0.1175 0.3107 0.1265 0.3459 0.0940 0.3237 0.3198 0.4958
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

71

Table 5.8. Parameter Estimates for 2008 PUMA Population Density (without amenity variables)
2009 ALL 2009 BAPLUS 2009 MAPLUS 2009 SCC 2009 BOH
Group A (Age 25-34) 0.006
(0.002)
0.008**
(0.004)
0.005**
(0.002)
0.006*
(0.003)
0.003**
(0.001)
0.003*
(0.001)
0.002*
(0.001)
0.002
(0.001)
5.897×10
-4
***
(2.191×10
-4
)
4.578×10
-4
**
(2.169×10
-4
)
N 1,708 1,708 1,614 1,614 1,251 1,251 1,129 1,129 424 424
R
2
0.1737 0.3746 0.1938 0.3467 0.1791 0.3277 0.1727 0.3368 0.1380 0.3654
Group B (Age 35-44) -0.225
(2.961)
6.104×10
-4

(5.193×10
-4
)
6.085×10
-4
*
(3.249×10
-4
)
8.012×10
-4
*
(4.388×10
-4
)
5.571×10
-4
***
(1.912×10
-4
)
5.354×10
-4
**
(2.098×10
-4
)
3.691×10
-4
**
(1.514×10
-4
)
3.511×10
-4
**
(14.510×10
-4
)
1.996×10
-4
***
(0.620×10
-4
)
1.953×10
-4
***
(0.565×10
-4
)
N 1,679 1,679 1,395 1,395 1,014 1,014 805 805 231 231
R
2
0.1175 0.3552 0.1710 0.3592 0.1475 0.3702 0.1144 0.3576 0.1062 0.5009
Group C (Age 45-54) 1.074×10
-4

(1.739×10
-4
)
4.969×10
-4
**
(2.132×10
-4
)
3.323×10
-4
***
(0.900×10
-4
)
4.331×10
-4
***
(1.221×10
-4
)
3.884×10
-4
***
(0.654×10
-4
)
4.296×10
-4
***
(0.582×10
-4
)
0.785×10
-4

(0.477×10
-4
)
0.639×10
-4

(0.438×10
-4
)
0.873×10
-4

(0.781×10
-4
)
0.477×10
-4

(1.206×10
-4
)
N 1,648 1,648 1,236 1,236 799 799 546 546 139 139
R
2
0.1010 0.3364 0.1196 0.3265 0.0959 0.3646 0.0466 0.3240 0.1676 0.7379
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

72

Table 5.9. Parameter Estimates for 2008 PUMA Population Density (with amenity variables)
2009 ALL 2009 BAPLUS 2009 MAPLUS 2009 SCC 2009 BOH
Group A (Age 25-34) 0.005***
(0.001)
0.006***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.002***
(4.442×10
-4
)
0.002***
(0.001)
0.001***
(3.350×10
-4
)
0.001***
(4.293×10
-4
)
4.682×10
-4
***
(1.202×10
-4
)
3.685×10
-4
***
(1.232×10
-4
)
N 1,308 1,308 1,244 1,244 973 973 892 892 335 335
R
2
0.2257 0.4232 0.2469 0.3903 0.2145 0.3482 0.2077 0.3590 0.1828 0.3431
Group B (Age 35-44) -4.820×10
-5

(2.537×10
-4
)
4.031×10
-4
**
(1.786×10
-4
)
3.565×10
-4
***
(1.168×10
-4
)
0.001***
(1.520×10
-4
)
4.028×10
-4
***
(1.137×10
-4
)
4.622×10
-4
***
(1.122×10
-4
)
2.437×10
-4
***
(5.510×10
-5
)
2.559×10
-4
***
(6.360×10
-5
)
2.377×10
-4
***
(6.850×10
-5
)
2.329×10
-4
***
(4.350×10
-5
)
N 1,291 1,291 1,093 1,093 811 811 644 644 195 195
R
2
0.1619 0.3973 0.1955 0.3683 0.1757 0.3790 0.1241 0.3652 0.1520 0.4980
Group C (Age 45-54) 1.449×10
-4

(2.293×10
-4
)
4.132×10
-4
**
*
(1.231×10
-4
)
3.008×10
-4
***
(5.950×10
-5
)
3.907×10
-4
***
(4.910×10
-5
)
3.776×10
-4
***
(5.960×10
-5
)
4.171×10
-4
***
(5.230×10
-5
)
9.540×10
-5
**
(4.230×10
-5
)
8.370×10
-5
**
(4.150×10
-5
)
1.117×10
-4
**
(5.330×10
-5
)
6.780×10
-5

(1.059×10
-4
)
N 1,264 1,264 967 967 643 643 442 442 109 109
R
2
0.1419 0.3466 0.1504 0.3286 0.1215 0.3535 0.0584 0.2953 0.2010 0.7332
Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
73

5.2.3. Results for PUMA “Talented” People Population Density
Table 5.10 and Table 5.11 (with amenity variables) report the parameter estimates for
2005 PUMA “talented” people population density from migration regression models for
2006 talented in-migrants for three age groups. In both the Census Division and MSA
fixed effects models, most parameter estimates are statistically significant and positively
associated with the number of in-migrants. However, the magnitude of the density
variable decreases as we move up the age ladder. Density does not matter for the SCC
migrants belonging to the 45-54 age group. For the 25-34 age group, the density has the
largest effect on the SCC group, followed by BAPLUS and MAPLUS, and BOH
in-migrants. For the 35-44 and 45-54 age groups, the density effects are similar across
different talented groups.
Table 5.12 and Table 5.13 (with amenity variables) report the parameter estimates
for 2008 PUMA “talented” people population density from migration regression models
for 2009 talented in-migrants for three age groups. In both the Census Division and MSA
fixed effects models, most parameter estimates are statistically significant and positively
associated with the number of in-migrants. The magnitude of the density variable
decreases as we move up the age ladder. The density variable is in general not
statistically significant for the SCC and BOH groups belonging to the 45-54 age group.
3


3
For complete regression results, see Appendices F and G.
74

Table 5.10. Parameter Estimates for 2005 PUMA “Talented” People Population Density (without amenity variables)
2006 BAPLUS 2006 MAPLUS 2006 SCC 2006 BOH
Group A (Age 25-34) 0.016**
(0.007)
0.015**
(0.007)
0.016**
(0.007)
0.014**
(0.007)
0.020***
(0.007)
0.018**
(0.008)
0.009***
(0.001)
0.008***
(0.001)
N 1,557 1,557 1,226 1,226 1,120 1,120 465 465
R
2
0.2795 0.4392 0.2413 0.4033 0.1729 0.3372 0.1010 0.2788
Group B (Age 35-44) 0.004**
(0.002)
0.004**
(0.002)
0.005***
(0.002)
0.005**
(0.002)
0.005**
(0.002)
0.004*
(0.002)
0.005***
(0.001)
0.005***
(0.001)
N 1,412 1,412 1,056 1,056 797 797 253 253
R
2
0.2284 0.3940 0.1890 0.4045 0.0946 0.3217 0.0866 0.4932
Group C (Age 45-54) 0.002***
(0.001)
0.002***
(0.001)
0.003***
(0.001)
0.003***
(0.001)
0.002*
(0.001)
0.001
(0.001)
0.004***
(0.001)
0.003*
(0.002)
N 1,284 1,284 887 887 615 615 172 172
R
2
0.1220 0.3555 0.1101 0.3951 0.0779 0.3518 0.2122 0.4685
Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

75

Table 5.11. Parameter Estimates for 2005 PUMA “Talented” People Population Density (with amenity variables)
2006 BAPLUS 2006 MAPLUS 2006 SCC 2006 BOH
Group A (Age 25-34) 0.013***
(0.004)
0.012***
(0.004)
0.014***
(0.005)
0.013***
(0.005)
0.018***
(0.006)
0.016***
(0.006)
0.008***
(0.001)
0.008***
(0.001)
N 1,273 1,273 1,019 1,019 947 947 406 406
R
2
0.3032 0.4446 0.2662 0.4027 0.1890 0.3418 0.1102 0.2668
Group B (Age 35-44) 0.004***
(0.001)
0.003***
(0.001)
0.005***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.003**
(0.001)
0.004***
(0.001)
0.005***
(0.001)
N 1,165 1,165 886 886 669 669 220 220
R
2
0.2314 0.3923 0.2125 0.4133 0.1220 0.3108 0.1273 0.4544
Group C (Age 45-54) 0.002***
(1.612×10
-4
)
0.002***
(1.888×10
-4
)
0.003***
(3.205×10
-4
)
0.002***
(3.083×10
-4
)
0.001
(0.001)
3.529×10
-4

(9.137×10
-4
)
0.005***
(0.001)
0.004***
(0.001)
N 1,057 1,057 737 737 512 512 140 140
R
2
0.1298 0.3175 0.1382 0.3545 0.0962 0.3238 0.3433 0.5150
Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

76

Table 5.12. Parameter Estimates for 2008 PUMA “Talented” People Population Density (without amenity variables)
2009 BAPLUS 2009 MAPLUS 2009 SCC 2009 BOH
Group A (Age 25-34) 0.019*
(0.010)
0.017
(0.011)
0.016*
(0.009)
0.015
(0.009)
0.020*
(0.011)
0.017
(0.012)
0.014***
(0.003)
0.012***
(0.003)
N 1,614 1,614 1,251 1,251 1,129 1,129 424 424
R
2
0.2349 0.3677 0.2124 0.3435 0.2053 0.3530 0.1622 0.3860
Group B (Age 35-44) 0.003**
(0.001)
0.003*
(0.002)
0.004**
(0.002)
0.003**
(0.002)
0.004***
(0.001)
0.004***
(0.001)
0.002***
(0.001)
0.002*
(0.001)
N 1,395 1,395 1,014 1,014 805 805 231 231
R
2
0.1856 0.3655 0.1632 0.3778 0.1249 0.3624 0.1001 0.4959
Group C (Age 45-54) 12.059×10
-4
***
(3.458×10
-4
)
12.213×10
-4
***
(4.042×10
-4
)
0.003***
(0.001)
0.002***
(0.001)
0.276×10
-4

(2.498×10
-4
)
-1.819×10
-4

(3.070×10
-4
)
4.891×10
-4

(6.429×10
-4
)
3.845×10
-4

(11.235×10
-4
)
N 1,236 1,236 799 799 546 546 139 139
R
2
0.1263 0.3307 0.1164 0.3791 0.0454 0.3235 0.1654 0.7377
Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

77

Table 5.13. Parameter Estimates for 2008 PUMA “Talented” People Population Density (with amenity variables)
2009 BAPLUS 2009 MAPLUS 2009 SCC 2009 BOH
Group A (Age 25-34) 0.013***
(0.005)
0.012**
(0.005)
0.012***
(0.005)
0.011**
(0.005)
0.015**
(0.006)
0.013*
(0.007)
0.012***
(0.001)
0.010***
(0.001)
N 1,244 1,244 973 973 892 892 335 335
R
2
0.2755 0.4036 0.2430 0.3626 0.2376 0.3754 0.2102 0.3637
Group B (Age 35-44) 0.002***
(0.001)
0.002***
(0.001)
0.003***
(0.001)
0.003***
(0.001)
0.004***
(0.001)
0.003***
(4.284×10
-4
)
0.002***
(0.001)
0.002**
(0.001)
N 1,093 1,093 811 811 644 644 195 195
R
2
0.2058 0.3731 0.1897 0.3867 0.1356 0.3713 0.1415 0.4879
Group C (Age 45-54) 0.001***
(1.081×10
-4
)
0.001***
(1.503×10
-4
)
0.002***
(2.624×10
-4
)
0.002***
(2.646×10
-4
)
1.456×10
-4

(2.265×10
-4
)
3.870×10
-6
(3.128×10
-4
)
0.001*
(4.061×10
-4
)
0.001
(0.001)
N 967 967 643 643 442 442 109 109
R
2
0.1557 0.3324 0.1384 0.3664 0.0561 0.2938 0.1973 0.7352
Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
78

5.3. Discussion
This chapter specifically examines the density-migration nexus for talented individuals
with respect to their life cycle while addressing the two research questions posed in the
previous chapter. For the first research question, whether talented movers prefer to live in
denser places, the estimation results using MSA population density are not definite. The
results for a given group are sensitive to model specification. This may be due to the fact
that MSA population density is a crude measure based on large geographic units. It may
have reflected some aspects embedded in amenity variables and geographic dummy
variables incorporated in the regression models. On the contrary, when PUMA
population and PUMA “talented” people densities are utilized, estimation results do show
that various talented groups prefer denser places.
For the second research question, whether the importance of population density for
talented movers change over the recent business cycle, models with MSA population
density do not show a clear pattern. When PUMA population and PUMA “talented”
people population densities are used, the estimation results of this study suggest that the
recent economic downturn did not change the taste for density for talented movers. They
still prefer to live in denser places.
As for the third research question, whether talented migrants at different phase of
their life cycle have different preferences for density, model results with MSA population
density remain confounding. Models with PUMA population and PUMA “talented”
people population densities show that overall the three age groups investigated in this
study all prefer denser places. The finding for migrants aged 35 to 54 is not in line with
what existing literature on migration and life cycle suggests. Talented movers as opposed
79

to the general public apparently make different calculations when making their moving
decisions with respect to destination population density. Nonetheless the regression
results do show that the effect of population densities wanes as we move up the age
stratum. Also the estimated correlation coefficients between PUMA population density
and number of in-migrants for various groups are higher for the younger groups. All
these suggest that density plays a more important role in drawing younger people. Neal
(1999) argues that young people move in order to obtain “more information about
potential matches with various careers” (p. 256). If so, the greater preferences over denser
places for talented movers aged 25-34 may have to do with the fact that denser places
provide them with the greater economic diversity that supports their career goals. Another
noticeable point is that by and large PUMA population densities do not matter for the
SCC and BOH migrants belonging to the 45-54 age group. This suggests that the more
mature members of the creative class may have idiosyncratic tastes when making their
migration decisions.

80

CHAPTER 6
CONCLUSION

6.1. Summary of Findings
This dissertation sought to examine whether various measures of urban density are
statistically significant in explaining the in-migration of talented individuals in the
continental United States. The dissertation addresses three research questions. First, are
denser places especially attractive to talented people? This study tests three measures of
destination density – MSA population density, PUMA population density, and PUMA
“talented” people density. Estimation results from multivariate tests show that, although
the effects of various density measures are small when PUMA densities are utilized,
talented individuals do prefer denser places. This finding corroborates what the theories
of the agglomeration economies suggest in terms of density preferences for talented
individuals.
Compared with MSA population density, population densities at the PUMA level
yield more consistent results for the density-talent migration nexus. Both PUMA
population and “Talented” population densities exert positive effects on talented
in-migration across different educational and occupational groups. The mixed effects of
81

MSA population density may be due to the fact that MSA population density varies
dramatically within any large MSA. In addition, MSA population density or other density
measures based on large geographic areas may reflect the dimensions of both production
and consumption amenities that may be highly correlated with other control variables and
thus produced mixed findings for the density-talent migration linkage.
The second question this study addresses is: Does density become more or less
important for the talented migrants in the face of the recent economic downturn?
Correlation coefficients do show that PUMA population density is relatively stronger in
the “bust” year (2009). Multivariate tests using MSA population density show that
although the density variable exerts mixed effects on various talented groups in 2006, it
has positive effects on the talented in 2009. This suggests that density has become more
important for talented migrants in the “bust” year (2009). This may not be surprising,
since denser places can offer people more networking opportunities that in turn may offer
better employment opportunities. When PUMA population density and PUMA “Talented”
people population density are tested however, the difference between the “boom” and
“bust” years of density effects is not obvious. Talented movers do not systematically
change their density preferences over the recent business cycle.

82

The third research question posed is: Does the density-talent migration nexus vary
with age, educational or occupational groups? Although theories suggest that younger
people may prefer denser places whereas older people may prefer places with lower
density, the estimation results of this study do not find the support for the prediction for
the older talented groups. Various talented groups all prefer denser places. Nonetheless,
both the correlation coefficients and multivariate tests do show that the effect of
population densities wanes as we move up the age stratum. Density is more important for
the youngest (25-34) migrant group. Another point worth mentioning is that by and large
PUMA population densities do not matter for the SCC and BOH migrants belonging to
the 45-54 age group. This may suggest that the more mature members of the creative
class may have idiosyncratic tastes when making their migration decisions.

6.2. Contributions and Policy Implications
Given that talented individuals and density are thought to be important for urban and
regional development, this dissertation examines the various possible linkages between
population density and talent migration and contributes to the existing literature in four
ways. First, most researchers use density measures based on geographic units such as
MSA or county that are too large. Due to the polycentric and otherwise varied features of
83

contemporary urban spatial structure, the conventional density measures may mask
considerable variation across vast urban landscape. The results of multivariate tests using
the MSA population density have demonstrated that this type of crude density measure
has offered mixed and confounding results for the density-talent migration nexus. This
suggests that the simple idea of “density” used by most researchers is inadequate. By
developing more refined population density measures at PUMA level, this study provides
alternative approaches to examine population density and various variables in the fields
of urban planning and economics.
Second, although there has been some attention given to the density-talent migration
nexus, a more detailed analysis can be useful. Theories of agglomeration economies posit
that high density serves as an attractor to talented individuals whereas the theories of
residential location choice also suggest that some talented people prefer to live at low
density. There are a few studies that econometrically test the effect of population density
on talent migration. However the density-migration link for talented movers was not their
central interest. By focusing specifically on the density-talent migration nexus, this study
adds to the accumulating empirical research on urban population density and its influence
on talent migration.

84

Third, while some studies have examined how business cycles and life cycles affect
migration choices, most of them did not focus on talented migrants. This study
specifically focuses on various talented groups, which are thought by some to be the key
driving force of economic development. This is the pertinent to the modern era of
knowledge-driven economies. This study therefore offers further insights into the
density-talent migration nexus by taking the recent business cycle and life cycle into
account.
Fourth, existing theoretical and empirical studies suggest that both talented
professionals and urban density matter for urban and regional economic development.
Based on the findings of this study on the density-talent migration nexus, one might ask:
should planners and policymakers promote a high-density planning paradigm? Or put
differently, are there dense area settings that are most attractive to talented individuals
planners and policymakers can craft to bolster economic development? Can planners and
policymakers find optimal densities for various talented individuals? The focus on just
“higher” densities is much too simple.
In Table 4.2, I have shown that both high- and low-density PUMAs successfully
attract BAPLUS migrants. Although theories of agglomeration economies on production
and consumption postulate that talented people prefer high-density places, as mentioned
85

previously, some scholars note that various talented individuals also prefer living in
low-density suburbs. Gordon and Ikeda (2011) also argue that innovative and creative
interactions also occur in “neighborhood lifestyle centers” in suburbs. In addition, both
high-density Manhattan and lower-density Silicon Valley have successfully attracted
talented professionals. In other words, talented people live and work at various densities.
Therefore, there is no one “good” or “bad” density. These suggest that there is no optimal
density, which planners and policymakers can pursue in order to attract talented
individuals and bolster economic development.

6.3. Directions for Future Research
Although this study has shed light on the density-talent migration nexus and elaborated
the concept of density, there are various elaborations that might usefully extend the
results of this study. Existing literature suggests that inter-metropolitan migrants have
different motivations of moving than intra-metropolitan movers. For example,
inter-metropolitan movers are more likely to be involved in job-related migration.
Because further understanding on migration behaviors of these two groups may help
policymakers to craft plans to bolster their local economic development, this study
suggests this as a priority for future research.
86

Regression results of this study suggest that the members of the creative class – SCC
and BOH – seem to be more idiosyncratic when making their moving decisions. One line
of criticism directed at the concept of the creative class as proposed by Richard Florida is
that the creative class in fact conflates heterogeneous types of professionals to a single
group. For example, computer engineers may prefer places that may not be attractive to
artists. Breaking down the creative class into more detailed groups may address this issue
and help researchers to better understand the migration behaviors of the creative class.
Future research into this subject can also examine various finer demographic and
socio-economic attributes of talented migrants. For example, people with young children
would take into account school quality. In this case, the empirical estimations should
include neighborhood school quality measures. In addition, people from different racial
groups, ethnicities, origins of birth, immigration status, marital status (single, married,
divorced, separated, widowed), gender, income, and employment status may weigh
different considerations when making their moving decisions. Future studies can and
should explore these directions.

87

REFERENCES

Abel, Jaison R., Ishita Dey, and Todd M. Gabe. 2012. “Productivity and the Density of
Human Capital,” Journal of Regional Science, 52(4), 562–586.
Abel, Jaison R., and Todd M. Gabe. 2011. “Human Capital and Economic Activity in
Urban America,” Regional Studies, 45(8), 1079–1090.
Acs, Zoltan J. 2006. Entrepreneurship, Geography, and American Economic Growth.
New York: Cambridge University Press.
Aharonovitz, Gilad David. 2011. “Knowledge-based Spatial Differences in Economic
Activity, Job Related Migration and Housing Related Migration,” The Annals of
Regional Science, 46(1), 159–188.
Anas, Alex, Richard Arnott, and Kenneth A. Small. 1998. “Urban Spatial Structure,”
Journal of Economic Literature, 36(3), 1426–1464.
Andersson, Fredrik, Simon Burgess, and Julia I. Lane. 2007. “Cities, Matching and the
Productivity Gains of Agglomeration,” Journal of Urban Economics, 61(1), 112–
128.
Audretsch, David B., and Max Keilbach. 2005. “Entrepreneurship Capital and Regional
Growth,” The Annuals of Regional Science, 39(3), 457–469.
Baptista, Rui. 2003. “Productivity and the Density of Local Clusters,” in Johannes
Bröcker, Dirk Dohse, and Rüdiger Soltwedel (eds.), Innovation Clusters and
Interregional Competition. Berlin: Springer.
Barrett, G. 1996. “The Transport Dimension,” in M. Jenks, E. Burton, and K. Williams
(eds.), The Compact City: A Sustainable Urban Form? London: E & FN Spon.
88

Baum-Snow, Nathaniel. 2007. “Did Highways Cause Suburbanization?” Quarterly
Journal of Economics, 122(2), 775–805.
Bröcker, Johannes, Dirk Dohse, and Rüdiger Soltwedel. 2003. “Clusters and Competition
as Engines of Innovation—An Introduction,” in Johannes Bröcker, Dirk Dohse, and
Rüdiger Soltwedel (eds.), Innovation Clusters and Interregional Competition.
Berlin: Springer.
Bures, Regina M. 1997. “Migration and the Life Course: Is there a Retirement
Transition?” International Journal of Population Geography, 3(2), 109–119.
Chapple, Karen, Ann Markusen, Greg Schrock, Daisaku Yamamoto, and Pingkang Yu.
2004. “Gauging Metropolitan “High-Tech” and “I-Tech” Activity,” Economic
Development Quarterly, 18(1), 10–29.
Chen, Yong, and Stuart S. Rosenthal. 2008. “Local Amenities and Life-cycle Migration:
Do People Move for Jobs or Fun?” Journal of Urban Economics, 64(3), 519–537.
Christensen, Clayton M., Scott D. Anthony, and Erik A. Roth. 2004. Seeing What’s Next:
Using the Theories of Innovation to Predict Industry Change. Boston: Harvard
Business School Press.
Ciccone, Antonio. 2002. “Agglomeration Effects in Europe,” European Economic
Review, 46(2), 213–227.
Ciccone, Antonio, and Robert E. Hall. 1996. “Productivity and the Density of Economic
Activity,” The American Economic Review, 86(1), 54–70.
Clark, William A.V., Youqin Huang, and Suzanne Withers. 2003. “Does Commuting
Distance Matter? Commuting Tolerance and Residential Change,” Regional Science
and Urban Economics, 33(2), 199–221.
89

Cumbers, Andy, and Danny MacKinnon. 2004. “Introduction: Clusters in Urban and
Regional Development,” Urban Studies, 41(5/6), 959–969.
Dawkins, Casey J., and Arthur C. Nelson. 2003. “State Growth Management Programs
and Central-City Revitalization,” Journal of the American Planning Association,
69(4), 381–396.
Di Addario, Sabrina, and Eleonora Patacchini. 2008. “Wages and the City: Evidence
from Italy,” Labour Economics, 15(5), 1040–1061.
DiPasquale, Denise, and Matthew E. Kahn. 1999. “Measuring Neighborhood
Investments: An Examination of Community Choice,” Real Estate Economics,
27(3), 389–424.
Downs, Anthony. 1994. New Visions for Metropolitan America. Washington, DC: The
Brookings Institution; Cambridge: Lincoln Institute of Land Policy.
Florida, Richard. 2002. The Rise of the Creative Class. New York: Basic Books.
Florida, Richard. 2003. “Cities and the Creative Class,” City & Community, 2(1), 3–19.
Florida, Richard. 2005. Cities and the Creative Class. New York: Routledge.
Florida, Richard. 2009. “How the Crash Will Reshape America,” Atlantic, Last accessed
June 15, 2013, http://www.theatlantic.com/magazine/archive/2009/03/how-the-
crash-will-reshape-america/307293/?single_page=true.
Frey, William H. 2009. The Great American Migration Slowdown: Regional and
Metropolitan Dimensions. Washington, DC: The Brookings Institution.
Frey, William H. 2011. “Migration Declines Further: Stalling Brain Gains and
Ambitions,” Last accessed June 16, 2013,
http://www.brookings.edu/research/opinions/2011/01/12-migration-frey.
90

Gabe, Todd M., Kristen Colby, and Kathleen P. Bell. 2007. “The Effects of Workforce
Creativity on Earnings in U.S. Counties,” Agricultural and Resource Economics
Review, 36(1), 71–83.
Glaeser, Edward L. 1994. “Cities, Information and Economic Growth,” Cityscape, 1(1),
9–47.
Glaeser, Edward L. 1998. “Are Cities Dying?” Journal of Economic Perspectives,12(2),
139–160.
Glaeser, Edward L. 1999. “Learning in Cities,” Journal of Urban Economics, 46(2), 254–
277.
Glaeser, Edward L. 2005. “Review of Richard Florida’s The Rise of the Creative Class,”
Regional Science and Urban Economics, 35(5), 593–596.
Glaeser, Edward L., and Joshua D. Gottlieb. 2006. “Urban Resurgence and the Consumer
City,” Urban Studies, 43(8), 1275–1299.
Glaeser, Edward L., and Matthew E. Kahn. 2001. “Decentralized Employment and the
Transformation of the American City,” Brookings-Wharton Papers on Urban
Affairs, 1–63.
Glaeser, Edward L., Hedi D. Kallal, José A. Scheinkman, and Andrei Shleifer. 1992.
“Growth in Cities,” Journal of Political Economy, 100(6), 1126–1152.
Glaeser, Edward L., Jed Kolko, and Albert Saiz. 2001. “Consumer city,” Journal of
Economic Geography, 1(1), 27–50.
Glaeser, Edward L., and Matthew G. Resseger. 2010. “The Complementarity between
Cities and Skills,” Journal of Regional Science, 50(1), 221–244.
91

Glaeser, Edward L., José A. Scheinkman, and Andrei Shleifer. 1995. “Economic Growth
in a Cross-section of Cities,” Journal of Monetary Economics, 36(1), 117-143.
Glaeser, Edward L., and Jesse M. Shapiro. 2003. “Urban Growth in the 1990s: Is City
Living Back?” Journal of Regional Science, 43(1), 139–165.
Gordon, I. 1985. “The Cyclical Interaction between Regional Migration, Employment
and Unemployment: A Time Series Analysis for Scotland,” Scottish Journal of
Political Economy, 32(2), 135–158.
Gordon, Peter. 2013. “Thinking about Economic Growth: Cities, Networks, Creativity
and Supply Chains for Ideas.” The Annals of Regional Science, 50(3), 667–684.
Gordon, Peter, and Sanford Ikeda. 2011. “Does Density Matter?” in David Emanuel
Andersson, Åke E. Andersson, and Charlotta Mellander (eds.), Handbook of
Creative Cities. Northampton: Edward Elgar.
Gordon, Peter, and Harry W. Richardson. 1997. “Are Compact Cities a Desirable
Planning Goal?” Journal of the American Planning Association, 63(1), 95–106.
Greenwood, Michael J. 1997. “Internal Migration in Developed Countries,” in M. R.
Rosenzweig, and O. Stark (eds.), Handbook of Population and Family Economics,
Vol. 1B. North Holland: Elsevier Science B.V.
Greenwood, Michael J., and Gary L. Hunt. 1989. “Jobs versus Amenities in the Analysis
of Metropolitan Migration,” Journal of Urban Economics, 25(1), 1–16.
Greenwood, Michael J., Gary L. Hunt, and John M. McDowell. 1986. “Migration and
Employment Change: Empirical Evidence on the Spatial and Temporal Dimensions
of the Linkage,” Journal of Regional Science, 26(2), 223–234.
92

Hansen, Høgni Kalsø, and Thomas Niedomysl. 2009. “Migration of the Creative Class:
Evidence from Sweden,” Journal of Economic Geography, 9(2), 191–206.
Haurin, Donald, and Jean Haurin. 1991. “Youth Migration in the United States: Analysis
of a Deindustrializing Region,” in John Stillwell, and Peter Congdon (eds.),
Migration Models: Macro and Micro Approaches. London: Belhaven Press.
Jaffe, Adam B., Manuel Trajtenberg, and Rebecca Henderson. 1993. “Geographic
Localization of Knowledge Spillovers as Evidenced by Patent Citations,” The
Quarterly Journal of Economics, 108(3), 577–598.
Jargowsky, Paul A., and Yoonhwan Park. 2009. “Cause or Consequence?
Suburbanization and Crime in U.S. Metropolitan Areas,” Crime & Delinquency,
55(1), 28–50.
Karlsson, Charlie. 2011. “Clusters, Networks and Creativity,” in David Emanuel
Andersson, Åke E. Andersson, and Charlotta Mellander (eds.), Handbook of
Creative Cities. Northampton: Edward Elgar.
King, Karen M. 2011. “Technology, Talent and Tolerance and Inter-regional Migration
in Canada,” in David Emanuel Andersson, Åke E. Andersson, and Charlotta
Mellander (eds.), Handbook of Creative Cities. Northampton: Edward Elgar.
Kiriakos, Carol Marie. 2011. “Why Being There Matters: Finnish Professionals in Silicon
Valley,” in David Emanuel Andersson , Åke E. Andersson, and Charlotta Mellander
(eds.), Handbook of Creative Cities. Northampton: Edward Elgar.
Knudsen, Brian, Richard Florida, Kevin Stolarick, and Gary Gates. 2008. “Density and
Creativity in U.S. Regions,” Annals of the Association of American Geographers,
98(2), 461–478.
93

Kolko, Jed. 2012. “Even After the Housing Bust, Americans Still Love the Suburbs,”
Last accessed June 16, 2013, http://trends.truliablog.com/category/city-vs-suburban-
growth/.
Kopecky, Karen A., and Richard M. H. Suen. 2010. “A Quantitative Analysis of
Suburbanization and the Diffusion of the Automobile,” International Economic
Review, 51(4), 1003–1037.
Krugman, Paul. 1991. “Increasing Returns and Economic Geography,” Journal of
Political Economy, 99(3), 483–499.
Lang, Rob, and Karen Danielsen. 2005. “Review Roundtable: Cities and the Creative
Class,” Journal of the American Planning Association, 71(2), 203–220.
Leamer, Edward E., and Michael Storper. 2001. “The Economic Geography of the
Internet Age,” Journal of International Business Studies, 32(4), 641–665.
Lee, Everett S. 1966. “A Theory of Migration,” Demography, 3(1), 47–57.
Levine, Jonathan. 1998. “Rethinking Accessibility and Jobs Housing Balance,” Journal
of the American Planning Association, 64(2), 133–149.
Linneman, P., and P.E. Graves. 1983. “Migration and Job Change: A Multinomial Logit
Approach,” Journal of Urban Economics, 14(3), 263–279.
Lohr, S. 2007. “Silicon Valley Shaped by Technology and Traffic,” The New York Times,
Last accessed June 15, 2013,
http://www.nytimes.com/2007/12/20/technology/20cluster.html?pagewanted=all.
Long, Larry. 1988. Migration and Residential Mobility in the United States. New York:
Russell Sage Foundation.
94

Lucas, Robert E., Jr. 1988. “On the Mechanics of Economic Development,” Journal of
Monetary Economics, 22(1), 3–42.
Lucas, Robert E., Jr. 1993. “Making a Miracle,” Econometrica, 61(2), 251–272.
Makower, H.J., J. Marschak, and H.W. Robinson. 1939. “Studies in Mobility of Labor:
Analysis for Great Britain, Part I,” Oxford Economic Papers, 2, 70–97.
Markusen, Ann. 2006. “Urban Development and the Politics of a Creative Class:
Evidence from a Study of Artists,” Environment and Planning A, 38(10), 1921–
1940.
Martin-Brelot, Helene, Michel Grossetti, Denis Eckert, Olga Gritsai, and Zoltán Kovács.
2010. “The Spatial Mobility of the ‘Creative Class’: A European Perspective,”
International Journal of Urban and Regional Research, 34(4), 854–870.
McGranahan, David, and Timothy Wojan. 2007. “Recasting the Creative Class to
Examine Growth Processes in Rural and Urban Counties,” Regional Studies, 41(2),
197–216.
Mieszkowski, Peter, and Edwin S. Mills. 1993. “The Causes of Metropolitan
Suburbanization,” Journal of Economic Perspectives, 7(3), 135–147.
Miller, Ann R. 1977. “Interstate Migrants in the United States: Some Social-economic
Differences by Type of Move,” Demography, 14(1), 1–17.
Millington, Jim. 2000. “Migration and Age: The Effect of Age on Sensitivity to
Migration Stimuli,” Regional Studies, 34(6), 521–533.
Molho, I. 1984. “A Dynamic Model of Interregional Migration Flows in Great Britain,”
Journal of Regional Science, 24(3), 317–337.
95

Molloy, Raven, Christopher L. Smith, and Abigail Wozniak. 2011. “Internal Migration in
the United States,” Journal of Economic Perspectives, 25(3), 173–196.
Mommaas, Hans. 2004. “Cultural Clusters and the Post-industrial City: Towards the
Remapping of Urban Cultural Policy,” Urban Studies, 41(3), 507–532.
Moretti, Enrico. 2012. The New Geography of Jobs. New York: Houghton Mifflin
Harcourt.
Neal, Derek. 1999. “The Complexity of Job Mobility among Young Men,” Journal of
Labor Economics, 17(2), 237–261.
Niedomysl, Thomas, and Høgni Kalsø Hansen. 2010. “What Matters More for the
Decision to Move: Jobs versus Amenities,” Environment and Planning A, 42(7),
1636–1649.
Plane, D.A. 1993. “Demographic Influences on Migration,” Regional Studies, 27(4),
375–383.
Porter, M. 1990. The Competitive Advantage of Nations. Basingstoke: Macmillan.
Puga, Diego. 2010. “The Magnitude and Causes of Agglomeration Economies,” Journal
of Regional Science, 50(1), 203–219.
Rappaport, Jordan. 2007. “Moving to Nice Weather,” Regional Science and Urban
Economics, 37(3), 375–398.
Rappaport, Jordan. 2008. “Consumption Amenities and City Population Density,”
Regional Science and Urban Economics, 38(6), 533–552.
Rappaport, Jordan. 2009. “The Increasing Importance of Quality of Life,” Journal of
Economic Geography, 9(6), 779–804.
96

Ravenstein, E.G. 1885. “The Laws of Migration,” Journal of the Statistical Society of
London, 48(2), 167–235.
Ravenstein, E.G. 1889. “The Laws of Migration,” Journal of the Royal Statistical
Society, 52(2), 241–301.
Rickman, D. S., and Rickman, S. D. 2011. “Population Growth in High-amenity
Nonmetropolitan Areas: What's the Prognosis?” Journal of Regional Science, 51(5),
863–879.
Ritsilä, Jari, and Marko Ovaskainen. 2001. “Migration and Regional Centralization of
Human Capital,” Applied Economics, 33(3), 317–325.
Romer, Paul M. 1986. “Increasing Returns and Long-run Growth,” Journal of Political
Economy, 94(5), 1002–1037.
Romer, Paul M. 1990. “Endogenous Technological Change,” Journal of Political
Economy, 98(5), Part 2, S71–S102.
Rossi, Peter H. 1955. Why Families Move: A Study in the Social Psychology of Urban
Residential Mobility. Glencoe: The Free Press.
Saks, Raven E., and Abigail Wozniak. 2011. “Labor Reallocation over the Business Cycle:
New Evidence from Internal Migration,” Journal of Labor Economics, 29(4), 697–
739.
Santos, A., N. McGuckin, H.Y. Nakamoto, D. Gray, and S. Liss. 2011. Summary of
Travel Trends: 2009 National Household Travel Survey. Washington, DC: U.S.
Department of Transportation.
Saxenian, AnnaLee. 1990. “Regional Networks and the Resurgence of Silicon Valley,”
California Management Review, 33(1), 89–112.
97

Saxenian, AnnaLee. 1994. Regional Advantage: Culture and Competition in Silicon
Valley and Route 128. Cambridge: Harvard University Press.
Schultz, Theodore W. 1961. “Investment in Human Capital,” American Economic
Review, 51(1), 1–17.
Scott, Allen J. 2010. “Jobs or Amenities? Destination Choices of Migrant Engineers in
the USA,” Papers in Regional Science, 89(1), 43–64.
Shaw, R.P. 1985. Intermetropolitan Migration in Canada: Changing Determinants over
Three Decades. Toronto: NC Press Ltd.
Silverstein, Merril. 1995. “Stability and Change in Temporal Distance between the
Elderly and Their Children,” Demography, 32(1), 29–45.
Simon, Curtis J., and Clark Nardinelli. 1996. “The Talk of the Town: Human Capital,
Information, and the Growth of English Cities, 1861 to 1961,” Explorations in
Economic History, 33(3), 384–413.
Simon, Curtis J., and Clark Nardinelli. 2002. “Human Capital and the Rise of American
Cities, 1900–1990,” Regional Science and Urban Economics, 32(1), 59–96.
Simonton, Dean Keith. 2011. “Big-C Creativity in the Big City,” in David Emanuel
Andersson, Åke E. Andersson, and Charlotta Mellander (eds.), Handbook of
Creative Cities. Northampton: Edward Elgar.
Sjaastad, Larry A. 1962. “The Costs and Returns of Human Migration,” Journal of
Political Economy, 70(5), Part 2: Investment in Human Beings, 80–93.
Stillwell, John, and Peter Congdon, eds., 1991. Migration Models: Macro and Micro
Approaches. London: Belhaven Press.
98

Stoll, Michael A. 2013. Great Recession Spurs a Shift to Local Moves. New York:
Russell Sage Foundation.
Storper, Michael, and Anthony J. Venables. 2004. “Buzz: Face-to-face Contact and the
Urban Economy,” Journal of Economic Geography, 4(4), 351–370.
U.S. Census Bureau. 2009. Design and Methodology: American Community Survey.
Washington, DC: U.S. Government Printing Office.
van der Linde, Claas. 2003. “The Demography of Clusters—Findings from the Cluster
Meta-Study,” in Bröcker, Johannes, Dirk Dohse, and Rüdiger Soltwedel (eds.),
Innovation Clusters and Interregional Competition. Berlin: Springer.
Walters, William H. 2000. “Types and Patterns of Later-life Migration,” Geografiska
Annaler: Series B, Human Geography, 82(3), 129–147.
Wheeler, Christopher H. 2001. “Search, Sorting, and Urban Agglomeration,” Journal of
Labor Economics, 19(4), 879–899.
Wheeler, Christopher H. 2004. “Wage Inequality and Urban Density,” Journal of
Economic Geography, 4(4), 421–437.
Whisler, Ronald L., Brigitte S. Waldorf, Gordon F. Mulligan, and David A. Plane. 2008.
“Quality of Life and the Migration of the College-Educated: A Life-Course
Approach,” Growth and Change, 39(1), 58–94.
Wojan, Timothy R., Dayton M. Lambert, and David A. McGranahan. 2007. “Emoting
with Their Feet: Bohemian Attraction to Creative Milieu,” Journal of Economic
Geography, 7(6), 711–736.
Wolfe, David A., and Meric S. Gertler. 2004. “Clusters from the Inside and Out: Local
Dynamics and Global Linkages,” Urban Studies, 41(5/6), 1071–1093.
99

Appendices
Appendix A. Occupational Types of the Creative Class
Appendix A.1. Super-creative Core
Type PUMS Data Dictionary
Code Occupation
Architecture and
engineering
1300 ENG-Architects, except naval
1310 ENG-Surveyors, cartographers, and photogrammetrists
1320 ENG-Aerospace engineers
1340 ENG-Biomedical and agricultural engineers
1350 ENG-Chemical engineers
1360 ENG-Civil engineers
1400 ENG-Computer hardware engineers
1410 ENG-Electrical and electronics engineers
1420 ENG-Environmental engineers
1430 ENG-Industrial engineers, including health and safety
1440 ENG-Marine engineers and naval architects
1450 ENG-Materials engineers
1460 ENG-Mechanical engineers
1520 ENG-Petroleum, mining and geological engineers, including mining safety engineers
1530 ENG-Miscellaneous engineers, including nuclear engineers
1540 ENG-Drafters
1550 ENG-Engineering technicians, except drafters
1560 ENG-Surveying and mapping technicians
Computer and
math
1000 CMM-Computer scientists and systems analysts
1010 CMM-Computer programmers
1020 CMM-Computer software engineers
1040 CMM-Computer support specialists
1060 CMM-Database administrators
1100 CMM-Network and computer systems administrators
1110 CMM-Network systems and data communications analysts
1200 CMM-Actuaries
1220 CMM-Operations research analysts
1240 CMM-Miscellaneous mathematical science occupations, including mathematicians and statisticians
Life, physical, and
social science
1600 SCI-Agricultural and food scientists
1610 SCI-Biological scientists
1640 SCI-Conservation scientists and foresters
1650 SCI-Medical scientists
1700 SCI-Astronomers and physicists
1710 SCI-Atmospheric and space scientists
1720 SCI-Chemists and materials scientists
1740 SCI-Environmental scientists and geoscientists
1760 SCI-Physical scientists, all other
1800 SCI-Economists
1810 SCI-Market and survey researchers
1820 SCI-Psychologists
1840 SCI-Urban and regional planners
1860 SCI-Miscellaneous social scientists, including sociologists
1900 SCI-Agricultural and food science technicians
1910 SCI-Biological technicians
1920 SCI-Chemical technicians
100

1930 SCI-Geological and petroleum technicians
1960 SCI-Miscellaneous life, physical, and social science technicians, including social science research assistants and nuclear technicians
Education,
training, and
library
2200 EDU-Postsecondary teachers
2300 EDU-Preschool and kindergarten teachers
2310 EDU-Elementary and middle school teachers
2320 EDU-Secondary school teachers
2330 EDU-Special education teachers
2340 EDU-Other teachers and instructors
2400 EDU-Archivists, curators, and museum technicians
2430 EDU-Librarians
2440 EDU-Library technicians
2540 EDU-Teacher assistants
2550 EDU-Other education, training, and library workers
Arts, design,
entertainment,
sports, and media
2600 ENT-Artists and related workers
2630 ENT-Designers
2700 ENT-Actors
2710 ENT-Producers and directors
2720 ENT-Athletes, coaches, umpires, and related workers
2740 ENT-Dancers and choreographers
2750 ENT-Musicians, singers, and related workers
2760 ENT-Entertainers and performers, sports and related workers, all other
2800 ENT-Announcers
2810 ENT-News analysts, reporters and correspondents
2820 ENT-Public relations specialists
2830 ENT-Editors
2840 ENT-Technical writers
2850 ENT-Writers and authors
2860 ENT-Miscellaneous media and communication workers
2900 ENT-Broadcast and sound engineering technicians and radio operators, and media and communication equipment workers, all other
2910 ENT-Photographers
2920 ENT-Television, video, and motion picture camera operators and editors
Analysts 0710 BUS-Management analysts
0820 FIN-Budget analysts
0830 FIN-Credit analysts
0840 FIN-Financial analysts
Notes: Engineers including “4930 SAL-Sales engineers,” “8610 PRD-Stationary engineers and boiler operators,” “9030 TRN-Aircraft pilots and flight engineers,” ” 9200 TRN-Locomotive engineers
and operators,” and “9300 TRN-Sailors and marine oilers, and ship engineers” are excluded.
101

Appendix A.2. Bohemians
T y p e PUMS Data Dictionary
C o d e Occ u p a t i o n
Authors 2850 ENT-WRITERS AND AUTHORS
Designers 2630 ENT-DESIGNERS
Musicians 2750 ENT-MUSICIANS, SINGERS, AND RELATED WORKERS
Actors 2700 ENT-ACTORS
Directors 2710 ENT-PRODUCERS AND DIRECTORS
Photographers 2910 ENT-PHOTOGRAPHERS
Dancers 2740 ENT-DANCERS AND CHOREOGRAPHERS
Artists 2600 ENT-ARTISTS AND RELATED WORKERS
Performers 2760 ENT-ENTERTAINERS AND PERFORMERS, SPORTS AND RELATED WORKERS, ALL OTHER
Others 2810 ENT-NEWS ANALYSTS, REPORTERS AND CORRESPONDENTS
2830 ENT-EDITORS
2840 ENT-TECHNICAL WRITERS
2900 ENT-BROADCAST AND SOUND ENGINEERING TECHNICIANS AND RADIO OPERATORS, AND MEDIA AND COMMUNICATION EQUIPMENT
WORKERS, ALL OTHER
2920 ENT-TELEVISION, VIDEO, AND MOTION PICTURE CAMERA OPERATORS AND EDITORS

102

Appendix B. Regression Results for the Migration Models – MSA Population Density (without amenity variables)
Appendix B.1. Regression Results for the Migration Models – Group A (Age 25-34) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_A Six_BAPLUS_Inmig_Pc_A Six_MAPLUS_Inmig_Pc_A Six_SCC_Inmig_Pc_A Six_BOH_Inmig_Pc_A
Constant -592.474***
(195.237)
-1782.267 ***
(221.373)
-869.104***
(169.945)
-1489.500***
(174.933)
-432.843***
(128.068)
507.011***
(100.084)
-277.421***
(81.888)
-581.384***
(118.258)
51.957
(36.781)
-78.960**
(35.685)
Five_MSA_Pop_Den_BEA -0.041**
(0.017)
-0.165***
(0.011)
-0.001
(0.010)
-0.048***
(0.011)
0.005
(0.005)
0.273***
(0.046)
0.001
(0.003)
-0.007
(0.005)
0.003
(0.002)
0.007***
(0.002)
Five_MSA_Inc_BEA_Ln 53.217***
(19.710)
184.406***
(24.709)
77.681***
(17.837)
145.383***
(21.137)
42.405***
(12.921)
-49.832***
(9.252)
26.464***
(8.331)
57.316***
(13.323)
-4.598
(3.594)
8.600**
(4.313)
Five_Pct_M_Vo -4.000***
(0.624)
-4.752***
(1.439)
-3.083***
(0.602)
-5.043***
(1.293)
-1.374***
(0.348)
-2.344***
(0.667)
-0.808***
(0.279)
-1.382**
(0.549)
-0.329**
(0.128)
-0.660**
(0.283)
Five_Emp_Pop_Ratio 3.539***
(0.499)
3.251***
(0.562)
2.716***
(0.401)
2.763***
(0.506)
0.802***
(0.225)
0.875***
(0.277)
0.782***
(0.188)
0.830***
(0.244)
0.145**
(0.067)
0.106
(0.123)
Five_PUMA_Self_Emp_Pop_Ratio -2.525***
(0.811)
-0.821
(0.646)
0.352
(0.526)
1.924***
(0.518)
0.023
(0.279)
0.577
(0.402)
-0.297
(0.238)
0.107
(0.332)
-0.027
(0.140)
-0.123
(0.271)
Five_PUMA_Mttw 0.752
(1.144)
-2.225*
(1.282)
-1.606***
(0.589)
-3.680***
(1.010)
-1.129***
(0.299)
-2.188***
(0.739)
-0.672***
(0.216)
-1.354***
(0.513)
0.035
(0.077)
-0.085
(0.140)
New_England 5.632
(16.759)
5.880
(16.083)
7.918
(10.774)
-1.767
(5.839)
1.056
(1.930)

Middle_Atlantic -39.601***
(11.386)
-6.320
(6.563)
-1.376
(3.212)
-2.626
(2.782)
1.464
(1.686)

East_North_Central 1.989
(22.947)
12.643
(11.787)
5.902
(5.130)
-1.052
(3.650)
-1.146
(1.881)

West_North_Central -22.062*
(12.071)
-14.200*
(8.249)
-6.555
(4.250)
-8.701**
(4.240)
-2.311
(2.020)

South_Atlantic 6.644
(13.102)
0.178
(8.619)
-0.098
(4.676)
-4.594
(3.512)
-2.235
(1.749)

East_South_Central 10.255
(17.131)
5.443
(7.052)
0.337
(3.575)
-7.825**
(3.594)
-2.322
(1.850)

West_South_Central 8.846**
(12.435)
-0.897
(6.669)
-1.214
(3.062)
-5.612*
(3.093)
-1.869
(1.667)

Mountain -7.169
(14.135)
-16.732**
(7.720)
-8.318**
(3.369)
-9.828**
(4.130)
-1.926
(2.273)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 402 402 392 392 325 325 284 284 164 164
N 1,647 1,647 1,557 1,557 1,226 1,226 1,120 1,120 465 465
R
2
0.1657 0.3885 0.1991 0.3761 0.1591 0.3437 0.1031 0.2839 0.0655 0.2483
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

103

Appendix B.2. Regression Results for the Migration Models – Group B (Age 35-44) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_B Six_BAPLUS_Inmig_Pc_B Six_MAPLUS_Inmig_Pc_B Six_SCC_Inmig_Pc_B Six_BOH_Inmig_Pc_B
Constant -436.462***
(132.048)
-1753.226***
(146.366)
-478.698***
(99.669)
-436.578***
(72.046)
-310.773***
(62.756)
211.274***
(43.199)
-185.667***
(37.831)
-297.812***
(25.466)
-81.556**
(33.014)
44.659
(28.244)
Five_MSA_Pop_Den_BEA -0.025**
(0.011)
-0.145***
(0.006)
0.002
(0.005)
-0.035***
(0.006)
-0.002
(0.003)
0.289***
(0.003)
-0.001
(0.002)
-0.012***
(0.002)
-0.001
(0.002)
0.001
(0.002)
Five_MSA_Inc_BEA_Ln 36.075***
(13.533)
174.145***
(16.410)
39.811***
(10.215)
40.182***
(8.952)
28.392***
(5.994)
-26.294***
(5.207)
18.245***
(3.839)
30.567***
(3.289)
9.207***
(3.147)
-3.138***
(1.108)
Five_Pct_M_Vo -1.288***
(0.333)
-1.059
(0.756)
-0.864***
(0.288)
-1.146*
(0.652)
-0.511***
(0.173)
-0.733*
(0.407)
-0.059
(0.129)
-0.431
(0.261)
-0.049
(0.089)
-0.188
(0.220)
Five_Emp_Pop_Ratio 1.703***
(0.324)
1.706***
(0.403)
1.410***
(0.214)
1.464***
(0.290)
0.593***
(0.134)
0.637***
(0.154)
0.229**
(0.091)
0.252**
(0.117)
-0.074
(0.108)
-0.020
(0.218)
Five_PUMA_Self_Emp_Pop_Ratio -0.434
(0.351)
0.081
(0.430)
0.552**
(0.249)
1.248***
(0.348)
0.086
(0.192)
0.470**
(0.200)
0.112
(0.136)
0.138
(0.219)
0.088
(0.116)
0.201
(0.203)
Five_PUMA_Mttw 1.694**
(0.704)
0.139
(0.771)
-0.043
(0.296)
-0.956**
(0.433)
-0.215
(0.130)
-0.670***
(0.249)
-0.277**
(0.123)
-0.462**
(0.185)
-0.065
(0.086)
-0.093
(0.149)
New_England 5.274
(4.872)
-1.552
(4.768)
-1.216
(3.418)
-1.163
(1.693)
-1.839
(1.410)

Middle_Atlantic -16.031**
(6.632)
-6.674**
(3.353)
-0.959
(1.955)
-1.800
(1.505)
1.564
(1.722)

East_North_Central 5.277
(11.881)
4.305
(5.509)
2.025
(2.728)
0.550
(2.153)
0.868
(1.670)

West_North_Central -4.632
(6.541)
-8.167**
(3.339)
-3.610*
(2.110)
-2.559
(2.619)
1.490
(2.063)

South_Atlantic 25.988***
(5.960)
4.841
(3.487)
-0.022
(2.438)
-0.882
(1.764)
0.062
(1.354)

East_South_Central 22.508**
(9.239)
7.373*
(3.938)
3.834
(2.524)
-1.279
(2.088)
-1.525
(1.534)

West_South_Central 23.709***
(6.568)
5.434*
(3.271)
0.327
(1.522)
-0.150
(1.790)
0.576
(1.658)

Mountain 3.912
(9.295)
-4.243
(4.699)
-6.356***
(2.404)
-2.956
(2.186)
2.018
(2.378)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 400 400 362 362 295 295 235 235 116 116
N 1,632 1,632 1,412 1,412 1,056 1,056 797 797 253 253
R
2
0.1802 0.4038 0.1990 0.3722 0.1552 0.3785 0.0787 0.3113 0.0629 0.4692
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

104

Appendix B.3. Regression Results for the Migration Models – Group C (Age 45-54) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_C Six_BAPLUS_Inmig_Pc_C Six_MAPLUS_Inmig_Pc_C Six_SCC_Inmig_Pc_C Six_BOH_Inmig_Pc_C
Constant -104.736
(78.422)
-376.891***
(129.571)
-144.469***
(38.708)
-468.724***
(50.915)
-67.310**
(31.068)
-231.894***
(42.069)
-88.099***
(27.293)
-111.760***
(41.862)
12.925
(21.680)
213.281***
(71.563)
Five_MSA_Pop_Den_BEA -0.022***
(0.006)
-0.091***
(0.005)
-0.002
(0.003)
-0.024***
(0.002)
0.004**
(0.002)
-0.003
(0.002)
3.232×10
-4

(18.277×10
-4
)
-0.003*
(0.002)
0.004***
(0.001)
0.006***
(0.002)
Five_MSA_Inc_BEA_Ln 8.138
(7.807)
41.364***
(14.344)
12.291***
(3.945)
46.597***
(6.015)
6.792**
(3.064)
23.947***
(4.984)
8.492***
(2.808)
11.644**
(4.881)
-0.912
(2.205)
-19.335***
(6.949)
Five_Pct_M_Vo -0.690***
(0.258)
-0.480
(0.628)
-0.397***
(0.137)
-0.617*
(0.327)
-0.260***
(0.095)
-0.545***
(0.207)
-0.100
(0.077)
-0.373**
(0.156)
-0.160*
(0.092)
-0.329
(0.280)
Five_Emp_Pop_Ratio 0.777***
(0.232)
0.729**
(0.325)
0.499***
(0.120)
0.361*
(0.187)
0.163*
(0.086)
0.088
(0.143)
0.180***
(0.069)
0.128
(0.121)
0.090
(0.079)
0.032
(0.197)
Five_PUMA_Self_Emp_Pop_Ratio 0.285
(0.307)
0.499
(0.444)
0.547***
(0.164)
0.963***
(0.268)
0.319**
(0.134)
0.593***
(0.174)
0.143
(0.104)
0.285
(0.180)
0.102
(0.119)
0.072
(0.292)
Five_PUMA_Mttw 0.975**
(0.466)
0.343
(0.590)
-0.054
(0.129)
-0.499***
(0.178)
-0.239**
(0.100)
-0.463***
(0.175)
-0.168**
(0.079)
-0.306**
(0.150)
-0.173**
(0.068)
-0.280*
(0.146)
New_England 5.165
(4.119)
0.813
(2.223)
3.516*
(1.793)
-0.334
(1.590)
-0.163
(1.005)

Middle_Atlantic -8.393***
(3.217)
-0.858
(1.390)
0.234
(0.945)
-1.011
(1.063)
0.794
(0.871)

East_North_Central 2.665
(7.047)
1.764
(2.288)
3.181*
(1.667)
0.964
(1.619)
4.731***
(1.764)

West_North_Central 4.224
(5.267)
0.094
(3.229)
1.044
(2.163)
-0.528
(1.550)
5.203***
(1.131)

South_Atlantic 14.898***
(4.329)
3.834**
(1.840)
1.842
(1.594)
-0.460
(0.946)
0.459
(0.787)

East_South_Central 22.702***
(7.038)
5.943***
(2.082)
3.378**
(1.602)
4.180***
(1.477)
10.273***
(3.185)

West_South_Central 14.648***
(3.602)
4.508**
(1.900)
2.739**
(1.104)
1.572
(1.119)
1.698
(1.209)

Mountain 4.884
(5.730)
1.837
(2.271)
1.758
(1.517)
-0.004
(1.476)
0.478
(1.117)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 401 401 349 349 284 284 222 222 87 87
N 1,610 1,610 1,284 1,284 887 887 615 615 172 172
R
2
0.1373 0.3941 0.0968 0.3342 0.0782 0.3716 0.0729 0.3516 0.2071 0.4557
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

105

Appendix B.4. Regression Results for the Migration Models – Group A (Age 25-34) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_A Nine_BAPLUS_Inmig_Pc_A Nine_MAPLUS_Inmig_Pc_A Nine_SCC_Inmig_Pc_A Nine_BOH_Inmig_Pc_A
Constant -765.646***
(175.679)
783.913***
(140.490)
-821.095***
(179.323)
630.841***
(129.229)
-435.376***
(111.916)
153.232**
(63.005)
-346.030***
(79.385)
76.000***
(27.926)
-43.706
(43.397)
321.736**
(155.493)
Eight_MSA_Pop_Den_BEA -0.019*
(0.011)
0.678***
(0.063)
0.013*
(0.008)
0.624***
(0.059)
0.006
(0.004)
0.233***
(0.033)
0.006*
(0.003)
0.256***
(0.021)
0.008***
(0.002)
0.036***
(0.004)
Eight_MSA_Inc_BEA_Ln 83.032***
(18.762)
-64.951***
(13.183)
80.978***
(19.103)
-58.184***
(14.870)
44.798***
(11.434)
-11.117
(6.763)
36.123***
(8.030)
-5.512***
(1.929)
6.110
(4.649)
-26.393**
(13.136)
Eight_Pct_M_Vo -5.564***
(1.173)
-7.866***
(2.783)
-3.958***
(0.994)
-6.812***
(2.202)
-1.834***
(0.547)
-2.979***
(1.123)
-1.256***
(0.416)
-2.414***
(0.838)
-0.426**
(0.196)
-1.362***
(0.497)
Eight_Emp_Pop_Ratio 1.927***
(0.500)
1.932***
(0.661)
1.544***
(0.500)
1.550*
(0.809)
0.394*
(0.225)
0.402
(0.370)
0.290*
(0.155)
0.251
(0.273)
-0.070
(0.149)
-0.001
(0.336)
Eight_PUMA_Self_Emp_Pop_Ratio -3.730***
(1.102)
-1.985
(1.239)
-0.339
(0.737)
0.939
(0.790)
-0.368
(0.320)
0.172
(0.490)
-0.397
(0.302)
-0.070
(0.471)
0.118
(0.248)
-0.078
(0.463)
Eight_PUMA_Mttw -0.979
(1.259)
-3.705**
(1.468)
-1.923***
(0.735)
-4.585***
(1.025)
-0.969***
(0.325)
-2.290***
(0.570)
-0.822***
(0.283)
-1.718***
(0.384)
-0.212
(0.188)
-0.519
(0.324)
New_England 32.858
(22.563)
20.196
(19.158)
7.359
(7.969)
6.534
(7.173)
3.284
(2.262)

Middle_Atlantic -32.856***
(8.160)
-10.320*
(6.205)
-1.197
(3.329)
-6.330**
(2.737)
-0.377
(1.352)

East_North_Central 25.262
(23.284)
28.270*
(16.010)
17.020**
(7.403)
9.367
(5.733)
5.093*
(2.579)

West_North_Central 0.731
(13.841)
-3.506
(9.148)
-2.465
(5.004)
-4.297
(3.652)
3.349
(2.581)

South_Atlantic 9.774
(11.982)
2.265
(9.256)
1.437
(4.527)
-2.670
(3.335)
-0.398
(2.306)

East_South_Central 33.247**
(13.848)
14.533**
(6.716)
6.922**
(3.412)
-1.911
(2.853)
2.852
(2.419)

West_South_Central 9.002
(9.510)
-3.696
(6.629)
-0.005
(3.565)
-5.671**
(2.516)
-0.580
(1.637)

Mountain -12.659
(12.845)
-17.907**
(7.743)
-6.453*
(3.637)
-6.833**
(3.128)
-0.319
(2.375)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 403 403 384 384 317 317 285 285 151 151
N 1,708 1,708 1,614 1,614 1,251 1,251 1,129 1,129 424 424
R
2
0.1409 0.3320 0.1490 0.3017 0.1343 0.2856 0.1417 0.3090 0.1150 0.3457
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

106

Appendix B.5. Regression Results for the Migration Models – Group B (Age 35-44) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_B Nine_BAPLUS_Inmig_Pc_B Nine_MAPLUS_Inmig_Pc_B Nine_SCC_Inmig_Pc_B Nine_BOH_Inmig_Pc_B
Constant -364.006***
(107.469)
654.002***
(167.698)
-394.330***
(72.808)
318.967***
(92.345)
-241.499***
(53.883)
-295.774***
(55.153)
-137.513***
(38.589)
-99.149***
(23.502)
-32.052
(33.137)
137.766***
(23.790)
Eight_MSA_Pop_Den_BEA -0.020**
(0.008)
0.449***
(0.030)
-0.002
(0.004)
0.216***
(0.021)
-0.001
(0.002)
0.039***
(0.003)
0.001
(0.002)
0.063***
(0.004)
0.002
(0.002)
0.003*
(0.002)
Eight_MSA_Inc_BEA_Ln 37.443***
(11.302)
-60.987***
(18.446)
37.315***
(7.237)
-30.661***
(9.356)
24.841***
(5.351)
32.492***
(4.209)
14.353***
(3.944)
11.754***
(2.979)
4.579
(3.293)
-11.780***
(3.684)
Eight_Pct_M_Vo -1.393***
(0.382)
-1.153
(0.923)
-1.070***
(0.237)
-1.319**
(0.535)
-0.714***
(0.172)
-1.199***
(0.362)
-0.511***
(0.125)
-0.690***
(0.248)
-0.039
(0.117)
-0.136
(0.281)
Eight_Emp_Pop_Ratio 0.516
(0.426)
0.396
(0.584)
0.592***
(0.157)
0.569***
(0.198)
0.125
(0.111)
0.018
(0.158)
0.118
(0.095)
0.079
(0.123)
-0.126
(0.113)
-0.063
(0.205)
Eight_PUMA_Self_Emp_Pop_Ratio -0.209
(0.426)
0.758
(0.476)
0.770**
(0.321)
1.611***
(0.430)
0.310
(0.224)
0.670**
(0.259)
0.148
(0.158)
0.179
(0.230)
0.228
(0.154)
0.242
(0.275)
Eight_PUMA_Mttw 0.689
(0.530)
-0.436
(0.579)
-0.484**
(0.228)
-1.403***
(0.352)
-0.441***
(0.148)
-1.056***
(0.277)
-0.293***
(0.087)
-0.771***
(0.178)
-0.105
(0.107)
-0.128
(0.298)
New_England 11.042
(6.965)
4.567
(6.057)
4.540
(4.871)
1.173
(3.223)
3.464
(2.893)

Middle_Atlantic -17.897***
(5.197)
-3.491
(2.725)
1.006
(1.907)
0.138
(1.572)
1.098
(1.507)

East_North_Central 4.926
(9.878)
4.985
(4.317)
5.855**
(2.494)
6.890***
(2.217)
8.569***
(2.489)

West_North_Central -1.787
(6.656)
-5.879
(3.729)
-2.828
(2.469)
1.368
(2.453)
7.146
(2.728)

South_Atlantic 15.108**
(6.519)
4.777
(4.001)
2.401
(2.718)
0.981
(1.544)
0.745
(1.488)

East_South_Central 13.363
(10.187)
10.369***
(3.432)
4.346*
(2.496)
2.924*
(1.590)
3.866
(2.794)

West_South_Central 9.663*
(5.495)
2.461
(2.809)
0.611
(2.097)
-0.386
(1.431)
2.044
(1.701)

Mountain -4.489
(5.744)
-5.974**
(2.849)
-5.556***
(2.039)
-1.422
(1.766)
2.313
(3.952)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 398 398 345 345 278 278 234 234 105 105
N 1,679 1,679 1,395 1,395 1,014 1,014 805 805 231 231
R
2
0.1248 0.3538 0.1666 0.3525 0.1378 0.3624 0.1052 0.3506 0.0976 0.4934
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

107

Appendix B.6. Regression Results for the Migration Models – Group C (Age 45-54) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_C Nine_BAPLUS_Inmig_Pc_C Nine_MAPLUS_Inmig_Pc_C Nine_SCC_Inmig_Pc_C Nine_BOH_Inmig_Pc_C
Constant -158.299**
(66.172)
-189.836**
(82.496)
-190.533***
(38.259)
-167.932***
(18.449)
-130.100***
(36.414)
-87.816***
(18.366)
-60.903**
(25.362)
-267.699***
(29.051)
12.112
(33.216)
86.138***
(28.920)
Eight_MSA_Pop_Den_BEA -0.016***
(0.004)
0.168***
(0.016)
-0.003*
(0.002)
0.074***
(0.008)
-0.004**
(0.001)
0.048***
(0.004)
-0.001
(0.001)
0.015***
(0.005)
-4.060×10
-4

(17.191×10
-4
)
-0.003
(0.007)
Eight_MSA_Inc_BEA_Ln 18.097***
(6.682)
20.874**
(8.561)
19.534***
(3.903)
16.899***
(1.816)
14.554***
(3.582)
10.385***
(2.283)
7.619***
(2.650)
28.308***
(3.861)
1.002
(3.424)
-6.080*
(3.347)
Eight_Pct_M_Vo -0.858***
(0.250)
-0.859
(0.589)
-0.449***
(0.121)
-0.558*
(0.291)
-0.267**
(0.132)
-0.604**
(0.276)
-0.131
(0.116)
-0.323
(0.295)
0.089
(0.089)
-0.117
(0.230)
Eight_Emp_Pop_Ratio 0.095
(0.183)
0.040
(0.247)
0.072
(0.086)
0.034
(0.104)
-0.042
(0.071)
-0.037
(0.096)
-0.104
(0.097)
-0.190
(0.155)
-0.245***
(0.084)
-0.083
(0.307)
Eight_PUMA_Self_Emp_Pop_Ratio 0.357
(0.297)
0.531
(0.364)
0.685***
(0.138)
1.002***
(0.242)
0.142
(0.096)
0.170
(0.177)
0.099
(0.106)
0.102
(0.218)
0.110
(0.121)
-0.088
(0.324)
Eight_PUMA_Mttw 0.399
(0.328)
-0.001
(0.435)
-0.301***
(0.112)
-0.531***
(0.202)
-0.349***
(0.065)
-0.559***
(0.116)
-0.131
(0.086)
-0.322*
(0.182)
-0.046
(0.137)
-0.154
(0.359)
New_England 8.498**
(3.790)
2.557
(2.702)
-0.230
(1.907)
-0.260
(0.952)
2.356
(2.141)

Middle_Atlantic -6.782**
(2.618)
0.497
(1.494)
2.196**
(1.102)
1.050
(1.224)
3.315**
(1.648)

East_North_Central 9.709**
(4.430)
5.522**
(2.210)
3.249**
(1.500)
3.123**
(1.270)
3.244**
(1.551)

West_North_Central 7.926**
(3.887)
4.395**
(1.841)
2.894
(1.861)
2.979
(2.065)
5.503*
(3.150)

South_Atlantic 15.850***
(3.598)
3.950**
(1.536)
1.464
(1.248)
0.304
(0.831)
0.190
(1.271)

East_South_Central 14.032
(9.159)
1.942
(2.412)
-0.699
(1.713)
-1.300
(0.963)
-2.165
(1.610)

West_South_Central 9.930***
(3.373)
1.467
(1.762)
-0.485
(1.277)
0.457
(1.149)
2.254
(1.980)

Mountain 5.415
(4.197)
1.096
(1.836)
-1.524
(1.470)
0.291
(1.403)
3.470*
(2.023)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 399 399 322 322 243 243 193 193 87 87
N 1,648 1,648 1,236 1,236 799 799 546 546 139 139
R
2
0.1102 0.3346 0.1174 0.3208 0.0881 0.3507 0.0463 0.3234 0.1648 0.7373
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

108

Appendix C. Regression Results for the Migration Models – MSA Population Density (with amenity variables)
Appendix C.1. Regression Results for the Migration Models – Group A (Age 25-34) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_A Six_BAPLUS_Inmig_Pc_A Six_MAPLUS_Inmig_Pc_A Six_SCC_Inmig_Pc_A Six_BOH_Inmig_Pc_A
Constant -279.635
(245.555)
6895.429***
(1749.211)
-574.692***
(180.003)
-3620.804***
(840.798)
-321.944***
(118.154)
-12173.080***
(1710.541)
-212.220**
(84.588)
3888.712***
(925.581)
93.171**
(45.201)
-1514.736***
(306.768)
Five_MSA_Pop_Den_BEA -0.014
(0.022)
-0.594***
(0.091)
0.008
(0.013)
-0.726
(0.611)
0.003
(0.007)
-0.024***
(0.009)
0.002
(0.004)
0.464***
(0.171)
0.004
(0.003)
-0.463***
(0.078)
Five_MSA_Pop_BEA_M 19.243**
(8.796)
372.737***
(66.000)
17.260***
(5.604)
72.704
(106.325)
6.222*
(3.229)
-285.984***
(49.390)
4.037
(2.524)
-48.479***
(16.057)
2.849**
(1.289)
36.741***
(7.022)
Five_MSA_Pop_BEA_L 0.170
(14.156)
509.605***
(103.658)
10.932
(10.726)
80.320
(148.220)
9.818
(7.271)
-465.292***
(81.437)
6.021
(4.184)
-8.407
(20.220)
2.599
(1.974)
27.281**
(11.304)
Five_MSA_Inc_BEA_Ln 27.870
(24.236)
-624.315***
(156.493)
57.442***
(18.436)
364.296***
(86.224)
35.229***
(12.335)
1135.986***
(156.838)
21.897**
(8.474)
-397.189***
(93.815)
-7.212*
(4.171)
151.264***
(30.879)
Five_Pct_M_Vo -3.805***
(0.478)
-3.625***
(1.008)
-3.173***
(0.532)
-3.982***
(0.954)
-1.503***
(0.370)
-1.839***
(0.624)
-0.788**
(0.307)
-1.092*
(0.564)
-0.331**
(0.152)
-0.648**
(0.314)
Five_Emp_Pop_Ratio 3.712***
(0.505)
3.537***
(0.566)
2.457***
(0.335)
2.454***
(0.391)
0.846***
(0.224)
0.907***
(0.307)
0.813***
(0.184)
0.847***
(0.228)
0.043
(0.077)
0.067
(0.131)
Five_PUMA_Self_Emp_Pop_Ratio -2.079***
(0.733)
-0.855
(0.714)
0.756
(0.476)
2.197***
(0.516)
0.101
(0.304)
0.769*
(0.408)
-0.264
(0.266)
0.199
(0.345)
-0.006
(0.167)
-0.121
(0.285)
Five_PUMA_Mttw -0.734
(0.765)
-3.384***
(0.995)
-2.533***
(0.522)
-4.295***
(1.046)
-1.546***
(0.451)
-2.435***
(0.918)
-1.001***
(0.314)
-1.614**
(0.637)
-0.036
(0.097)
-0.071
(0.148)
January_Average_Temperature -0.907*
(0.467)
-11.711***
(2.491)
-0.807**
(0.316)
-0.804
(0.297)
-0.350**
(0.164)
13.254***
(2.185)
-0.265**
(0.124)
14.518***
(1.594)
-0.137**
(0.053)
2.506***
(0.526)
Five_Violent_Crime -0.007
(0.022)
0.405***
(0.096)
-0.014
(0.011)
0.039
(0.094)
-0.011*
(0.006)
-0.324***
(0.086)
-0.005
(0.006)
0.501***
(0.165)
-0.003
(0.003)
0.111***
(0.023)
Five_Property_Crime 0.001
(0.004)
-0.043***
(0.011)
-0.003
(0.003)
-0.010
(0.019)
-0.002
(0.002)
0.049***
(0.008)
8.870×10
-5

(0.002)
-0.145***
(0.025)
1.618×10
-4

(0.001)
-0.036***
(0.007)
New_England -9.205
(22.553)
-5.810
(19.032)
1.511
(12.844)
-6.021
(7.044)
-0.772
(2.342)

Middle_Atlantic -51.642***
(17.229)
-23.697**
(11.824)
-12.505**
(5.678)
-8.383**
(4.038)
-0.610
(2.378)

East_North_Central -51.097***
(14.610)
-24.403**
(9.828)
-11.916**
(5.262)
-13.141***
(4.713)
-5.076***
(1.898)

West_North_Central -54.585***
(15.428)
-39.972***
(10.745)
-19.275***
(6.495)
-16.670***
(6.048)
-4.681**
(2.299)

South_Atlantic 10.059
(11.919)
1.557
(6.821)
-0.496
(3.727)
-4.749
(3.142)
-2.314
(1.854)

East_South_Central 2.614
(17.377)
-5.502
(7.740)
-3.352
(3.606)
-12.571***
(3.998)
-4.588**
(2.126)

West_South_Central 12.517
(14.763)
0.080
(8.838)
-2.555
(4.263)
-7.489*
(4.013)
-1.890
(2.001)

Mountain -23.836
(15.001)
-26.898***
(9.983)
-12.598***
(4.465)
-13.190**
(5.362)
-2.578
(2.223)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 277 277 276 276 241 241 218 218 136 136
N 1,342 1,342 1,273 1,273 1,019 1,019 947 947 406 406
R
2
0.2084 0.4020 0.2353 0.3900 0.1974 0.3497 0.1253 0.2922 0.0779 0.2362
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
109

Appendix C.2. Regression Results for the Migration Models – Group B (Age 35-44) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_B Six_BAPLUS_Inmig_Pc_B Six_MAPLUS_Inmig_Pc_B Six_SCC_Inmig_Pc_B Six_BOH_Inmig_Pc_B
Constant -461.689***
(156.665)
-13905.200***
(581.027)
-440.770***
(109.154)
-948.409*
(484.294)
-269.577***
(65.759)
1379.582**
(535.800)
-165.599***
(44.871)
-1956.821**
(837.433)
-80.617**
(34.376)
312.607
(660.342)
Five_MSA_Pop_Den_BEA -0.015
(0.013)
2.659***
(0.350)
0.006
(0.006)
-2.888***
(0.257)
0.002
(0.004)
-0.260***
(0.089)
0.002
(0.002)
-0.015
(0.014)
-0.002
(0.003)
-0.031*
(0.018)
Five_MSA_Pop_BEA_M 1.387
(5.476)
-733.485***
(65.656)
5.495**
(2.744)
484.464***
(48.942)
2.679
(1.917)
68.010***
(14.126)
3.235**
(1.406)
-20.544**
(8.691)
1.602
(1.239)
11.205
(14.017)
Five_MSA_Pop_BEA_L -8.804
(7.011)
-1168.450***
(91.649)
2.087
(5.592)
662.230***
(70.820)
1.249
(4.340)
146.534***
(26.404)
0.771
(1.956)
-79.619***
(27.582)
2.136
(1.703)
28.239
(41.158)
Five_MSA_Inc_BEA_Ln 35.531**
(15.711)
1182.992***
(52.722)
37.985***
(10.989)
187.828***
(41.922)
25.611***
(6.384)
-121.192**
(50.851)
16.007***
(4.305)
176.320**
(78.945)
9.140***
(3.281)
-30.113
(66.115)
Five_Pct_M_Vo -0.818***
(0.273)
-0.316
(0.459)
-0.663**
(0.261)
-0.564
(0.431)
-0.489**
(0.191)
-0.500
(0.358)
-0.026
(0.118)
-0.329
(0.201)
-0.015
(0.103)
-0.075
(0.189)
Five_Emp_Pop_Ratio 1.973***
(0.338)
2.018***
(0.418)
1.279***
(0.208)
1.352***
(0.297)
0.581***
(0.141)
0.721***
(0.170)
0.317***
(0.093)
0.369***
(0.133)
0.020
(0.120)
0.076
(0.270)
Five_PUMA_Self_Emp_Pop_Ratio -0.208
(0.434)
0.381
(0.438)
0.842***
(0.290)
1.432***
(0.322)
0.244
(0.188)
0.502**
(0.208)
0.208
(0.158)
0.199
(0.222)
0.151
(0.123)
0.215
(0.197)
Five_PUMA_Mttw 1.153**
(0.558)
-0.299
(0.491)
-0.433*
(0.252)
-1.273***
(0.392)
-0.367**
(0.167)
-0.835***
(0.282)
-0.416***
(0.117)
-0.583***
(0.176)
-0.192**
(0.096)
-0.182
(0.151)
January_Average_Temperature 0.121
(0.248)
24.333***
(1.483)
-0.215
(0.142)
-8.791***
(1.224)
-0.269***
(0.092)
-3.894***
(0.859)
-0.055
(0.061)
8.604***
(0.576)
-0.051
(0.065)
-0.063
(0.154)
Five_Violent_Crime 0.008
(0.014)
0.147***
(0.054)
-0.001
(0.006)
-0.272***
(0.036)
0.002
(0.005)
-0.040
(0.039)
-0.006*
(0.003)
0.077*
(0.045)
0.002
(0.003)
0.039*
(0.023)
Five_Property_Crime 0.002
(0.003)
0.137***
(0.012)
3.826×10
-4
(0.001)
-0.088***
(0.009)
8.180×10
-5

(0.001)
0.015***
(0.002)
0.001
(0.001)
-0.060***
(0.004)
-0.001
(0.001)
-0.003
(0.005)
New_England 5.768
(7.700)
-5.276
(6.936)
-6.539
(4.865)
-2.531
(1.851)
-4.406**
(1.826)

Middle_Atlantic -8.034
(10.231)
-8.992
(6.005)
-6.617
(4.002)
-1.890
(2.180)
-0.264
(1.896)

East_North_Central -3.718
(7.542)
-6.789
(4.627)
-6.294*
(3.356)
-2.189
(2.303)
-0.924
(2.228)

West_North_Central -4.244
(8.751)
-12.249**
(5.302)
-10.691***
(3.625)
-2.908
(3.013)
-0.605
(3.845)

South_Atlantic 29.614***
(7.318)
6.455*
(3.548)
-0.226
(2.090)
0.631
(1.689)
0.212
(1.260)

East_South_Central 19.192**
(9.602)
4.791
(4.494)
0.168
(2.990)
-0.682
(2.364)
-1.907
(1.812)

West_South_Central 25.464***
(8.479)
5.173
(4.508)
-0.625
(2.541)
0.448
(2.055)
-0.424
(1.720)

Mountain 1.624
(9.978)
-7.252
(5.378)
-9.326***
(2.885)
-4.257*
(2.330)
1.838
(2.357)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 276 276 260 260 226 226 183 183 94 94
N 1,330 1,330 1,165 1,165 886 886 669 669 220 220
R
2
0.2039 0.4110 0.2094 0.3754 0.1837 0.3913 0.1120 0.3048 0.1052 0.4258
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

110

Appendix C.3. Regression Results for the Migration Models – Group C (Age 45-54) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_C Six_BAPLUS_Inmig_Pc_C Six_MAPLUS_Inmig_Pc_C Six_SCC_Inmig_Pc_C Six_BOH_Inmig_Pc_C
Constant -162.406*
(82.550)
-7501.785***
(356.397)
-130.921***
(42.270)
-1240.326***
(199.527)
-20.580
(27.620)
804.881***
(222.593)
-74.982**
(28.990)
132.214
(432.582)
15.656
(21.678)
-3351.153*
(1853.628)
Five_MSA_Pop_Den_BEA -0.010
(0.007)
0.088
(0.242)
-0.002
(0.003)
0.413***
(0.023)
0.005**
(0.002)
0.184**
(0.071)
-0.001
(0.002)
0.118
(0.184)
0.004**
(0.002)
0.633**
(0.284)
Five_MSA_Pop_BEA_M 0.100
(3.267)
-159.369***
(42.694)
3.514**
(1.514)
-56.157***
(6.457)
3.415***
(1.212)
-7.034
(6.392)
2.276***
(0.836)
-11.073
(16.395)
-0.003
(0.888)
-263.330**
(119.922)
Five_MSA_Pop_BEA_L -11.609***
(3.947)
-326.963***
(58.242)
2.537
(2.601)
-97.877***
(12.715)
3.405
(2.067)
6.507
(7.224)
1.758
(1.064)
-15.771
(16.014)
-0.073
(1.058)
-488.398**
(232.200)
Five_MSA_Inc_BEA_Ln 10.823
(8.072)
684.592***
(35.009)
12.111***
(4.169)
108.556***
(19.659)
3.886
(2.805)
-79.012***
(21.990)
7.885***
(2.804)
-15.390
(46.881)
-0.859
(2.105)
354.327*
(192.851)
Five_Pct_M_Vo -0.294
(0.191)
0.080
(0.293)
-0.326***
(0.106)
-0.389*
(0.210)
-0.285***
(0.086)
-0.452***
(0.159)
-0.065
(0.072)
-0.284*
(0.154)
-0.110
(0.113)
-0.162
(0.269)
Five_Emp_Pop_Ratio 0.929***
(0.213)
0.859***
(0.287)
0.448***
(0.141)
0.407*
(0.214)
0.079
(0.102)
0.010
(0.159)
0.139*
(0.082)
0.113
(0.142)
0.089
(0.132)
0.139
(0.190)
Five_PUMA_Self_Emp_Pop_Ratio 0.451
(0.331)
0.629
(0.454)
0.693***
(0.197)
1.019***
(0.286)
0.496***
(0.138)
0.693***
(0.169)
0.242**
(0.115)
0.423**
(0.182)
0.006
(0.127)
0.032
(0.236)
Five_PUMA_Mttw 0.636
(0.399)
-0.110
(0.400)
-0.218
(0.135)
-0.644***
(0.150)
-0.359***
(0.125)
-0.542***
(0.183)
-0.264***
(0.098)
-0.323*
(0.193)
-0.236**
(0.098)
-0.308
(0.272)
January_Average_Temperature 0.100
(0.170)
8.740***
(0.894)
-0.090
(0.088)
3.772***
(0.306)
-0.169***
(0.061)
-0.199
(0.133)
-0.064
(0.045)
0.478*
(0.253)
0.032
(0.040)
-1.169
(0.731)
Five_Violent_Crime 0.006
(0.010)
-0.039
(0.039)
-0.005
(0.003)
-0.284***
(0.025)
-0.007***
(0.003)
0.023***
(0.005)
0.001
(0.002)
0.004
(0.025)
-0.005*
(0.003)
-0.277**
(0.108)
Five_Property_Crime 0.003*
(0.002)
0.037***
(0.008)
-4.005×10
-4

(0.001)
0.015***
(0.001)
2.810×10
-5
(0.001)
0.003
(0.002)
-3.442×10
-4

(0.001)
2.746×10
-4

(0.011)
8.780×10
-5

(0.001)
-0.015**
(0.007)
New_England 4.930
(5.351)
-4.463
(3.071)
0.007
(2.767)
-3.004
(1.949)
-0.960
(1.359)

Middle_Atlantic 1.336
(4.988)
-2.378
(2.781)
-3.194
(1.935)
-1.591
(1.547)
1.759
(1.320)

East_North_Central -1.960
(4.764)
-1.559
(2.578)
-0.200
(1.966)
-2.106
(1.497)
1.571
(1.465)

West_North_Central 7.357
(5.892)
-1.262
(3.676)
-4.028
(2.499)
-3.213*
(1.845)
4.440**
(1.848)

South_Atlantic 18.312***
(4.867)
5.541***
(1.695)
3.152**
(1.477)
-0.600
(1.013)
0.608
(0.886)

East_South_Central 23.339***
(6.972)
5.486**
(2.461)
2.051
(1.727)
2.206
(1.630)
16.773***
(1.517)

West_South_Central 16.489***
(4.954)
5.115**
(2.266)
2.940*
(1.652)
1.092
(1.166)
1.590
(1.526)

Mountain 3.235
(6.157)
2.082
(2.850)
0.886
(1.598)
-0.381
(1.710)
0.631
(1.257)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 276 276 247 247 208 208 170 170 65 65
N 1,313 1,313 1,057 1,057 737 737 512 512 140 140
R
2
0.1705 0.4039 0.1104 0.3008 0.1158 0.3341 0.0938 0.3236 0.3149 0.4862
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

111

Appendix C.4. Regression Results for the Migration Models – Group A (Age 25-34) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_A Nine_BAPLUS_Inmig_Pc_A Nine_MAPLUS_Inmig_Pc_A Nine_SCC_Inmig_Pc_A Nine_BOH_Inmig_Pc_A
Constant -224.975
(236.236)
-41214.080***
(2732.499)
-445.328**
(215.449)
-147853.700***
(48264.160)
-291.558**
(114.332)
-721.570
(604.171)
-256.186***
(93.821)
132.707
(233.452)
14.840
(44.502)
17.839
(179.327)
Eight_MSA_Pop_Den_BEA 0.002
(0.016)
-3.587***
(0.247)
0.024**
(0.012)
-7.123***
(2.274)
0.010
(0.006)
0.069
(0.046)
0.007
(0.005)
0.090***
(0.022)
0.012***
(0.003)
0.057
(0.035)
Five_MSA_Pop_BEA_M 22.610**
(9.147)
-118.615***
(18.732)
23.373***
(6.130)
-1432.550***
(480.550)
10.091***
(3.029)
-25.733
(18.312)
6.900***
(2.540)
-17.821***
(2.744)
4.796***
(1.661)
-14.685**
(6.120)
Five_MSA_Pop_BEA_L 20.107
(15.448)
-56.841***
(7.228)
23.737**
(11.846)
-1204.045***
(415.947)
9.589*
(5.235)
6.966
(14.081)
7.225*
(3.848)
12.267***
(4.140)
4.166
(3.388)
15.668***
(4.597)
Eight_MSA_Inc_BEA_Ln 46.331**
(22.101)
3978.654***
(262.564)
58.049***
(20.659)
13963.260***
(4553.774)
36.229***
(11.568)
78.358
(58.470)
30.760***
(8.788)
-4.240
(20.455)
3.058
(4.244)
4.649
(16.147)
Eight_Pct_M_Vo -5.222***
(0.685)
-5.862***
(1.250)
-3.722***
(0.681)
-5.274***
(1.118)
-1.727***
(0.430)
-2.183***
(0.554)
-1.113***
(0.279)
-1.693***
(0.368)
-0.261*
(0.145)
-0.934***
(0.280)
Eight_Emp_Pop_Ratio 1.256**
(0.520)
1.510***
(0.538)
0.732**
(0.327)
0.800**
(0.380)
0.151
(0.209)
0.172
(0.258)
0.152
(0.131)
0.074
(0.154)
-0.290*
(0.168)
-0.292
(0.296)
Eight_PUMA_Self_Emp_Pop_Ratio -2.117**
(1.064)
-1.255
(1.102)
0.871
(0.672)
1.505**
(0.642)
0.118
(0.227)
0.483*
(0.264)
-0.017
(0.250)
0.273
(0.269)
0.414*
(0.227)
0.236
(0.340)
Eight_PUMA_Mttw -2.688***
(0.708)
-4.599***
(1.049)
-3.243***
(0.669)
-4.971***
(1.093)
-1.567***
(0.371)
-2.448***
(0.671)
-1.300***
(0.252)
-1.900***
(0.354)
-0.566***
(0.146)
-0.862***
(0.201)
January_Average_Temperature -1.729***
(0.458)
28.101***
(2.159)
-1.261***
(0.323)
150.078***
(50.013)
-0.569***
(0.166)
0.275
(0.942)
-0.285**
(0.129)
-0.536*
(0.308)
-0.241**
(0.097)
-0.089
(0.389)
Eight_Violent_Crime -0.026
(0.023)
-1.172***
(0.103)
-0.031**
(0.012)
-8.198***
(2.748)
-0.011*
(0.006)
-0.042
(0.059)
-0.014**
(0.005)
0.022*
(0.011)
-0.010*
(0.005)
-0.027***
(0.009)
Eight_Property_Crime -0.001
(0.005)
-0.081***
(0.006)
1.484×10
-4

(0.004)
0.225***
(0.082)
-3.961×10
-4

(0.002)
-0.005*
(0.003)
-3.873×10
-4

(0.001)
-0.006***
(0.002)
0.002*
(0.001)
-0.001
(0.001)
New_England 8.237
(31.142)
5.293
(25.777)
-1.444
(10.985)
3.585
(9.478)
1.921
(2.218)

Middle_Atlantic -65.343***
(13.232)
-32.189***
(10.783)
-11.181*
(6.003)
-10.817***
(3.998)
-3.246
(2.320)

East_North_Central -37.840***
(14.394)
-15.342
(10.624)
-5.055
(5.763)
-5.622
(4.541)
-5.880**
(2.581)

West_North_Central -37.168*
(20.711)
-31.442**
(15.359)
-14.230**
(6.021)
-12.638***
(4.522)
-4.057
(2.523)

South_Atlantic 14.321
(9.666)
5.322
(7.061)
2.234
(3.734)
-1.088
(3.377)
-0.277
(1.856)

East_South_Central 22.719
(19.024)
3.802
(9.546)
-0.553
(4.288)
-4.748
(3.470)
-2.540
(2.510)

West_South_Central 10.961
(13.875)
-4.301
(8.677)
-0.871
(4.240)
-5.898*
(3.422)
-2.806
(2.888)

Mountain -28.132*
(15.307)
-30.620***
(10.630)
-12.906**
(5.483)
-9.822**
(3.866)
-2.526
(2.316)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 241 241 209 209 196 196 110 110
N 1,308 1,308 1,244 1,244 973 973 892 892 335 335
R
2
0.1944 0.3867 0.2094 0.3497 0.1804 0.3117 0.1864 0.3367 0.1756 0.3267
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

112

Appendix C.5. Regression Results for the Migration Models – Group B (Age 35-44) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_B Nine_BAPLUS_Inmig_Pc_B Nine_MAPLUS_Inmig_Pc_B Nine_SCC_Inmig_Pc_B Nine_BOH_Inmig_Pc_B
Constant -209.712
(165.818)
1453.548***
(442.400)
-305.406***
(87.738)
1498.518***
(490.933)
-183.245***
(59.573)
-5640.498***
(564.559)
-120.650**
(53.833)
707.872***
(137.836)
-29.388
(34.139)
181.463
(257.706)
Eight_MSA_Pop_Den_BEA -0.015
(0.011)
-0.639***
(0.036)
0.001
(0.006)
-0.073
(0.059)
0.002
(0.004)
-0.694***
(0.095)
0.002
(0.003)
0.078**
(0.033)
0.005
(0.003)
0.013
(0.031)
Eight_MSA_Pop_BEA_M 4.316
(4.706)
168.654***
(12.015)
5.739**
(2.648)
50.857***
(4.839)
2.509
(1.996)
15.898
(9.774)
2.142
(1.553)
2.162
(3.756)
-0.145
(1.688)
6.698
(12.924)
Eight_MSA_Pop_BEA_L 4.149
(11.029)
181.257***
(11.970)
5.363
(5.644)
70.819***
(7.932)
1.898
(4.014)
29.604***
(10.723)
2.388
(2.845)
28.741***
(1.923)
-2.139
(2.184)
6.391
(14.377)
Eight_MSA_Inc_BEA_Ln 21.915
(15.543)
-114.084***
(42.426)
31.060***
(8.459)
-133.180***
(43.723)
21.192***
(5.822)
545.158***
(55.667)
13.419***
(5.071)
-57.108***
(12.327)
3.447
(3.279)
-16.373
(23.687)
Eight_Pct_M_Vo -1.209***
(0.309)
-0.339
(4.475)
-1.041***
(0.208)
-0.876***
(0.309)
-0.823***
(0.167)
-0.913***
(0.236)
-0.504***
(0.105)
-0.540***
(0.150)
-0.076
(0.148)
-0.129
(0.337)
Eight_Emp_Pop_Ratio 0.976***
(0.263)
1.023***
(3.303)
0.539***
(0.189)
0.600**
(0.235)
0.048
(0.135)
0.006
(0.174)
0.069
(0.117)
0.006
(0.139)
-0.097
(0.139)
0.059
(0.233)
Eight_PUMA_Self_Emp_Pop_Ratio 0.278
(0.476)
0.878*
(0.496)
1.111***
(0.347)
1.714***
(0.446)
0.477**
(0.201)
0.741***
(0.259)
0.268
(0.172)
0.331
(0.206)
0.288**
(0.126)
0.341
(0.220)
Eight_PUMA_Mttw 0.205
(0.342)
-0.747*
(0.401)
-0.702**
(0.281)
-1.428***
(0.422)
-0.583**
(0.235)
-1.106***
(0.323)
-0.384***
(0.128)
-0.769***
(0.190)
0.018
(0.103)
0.018
(0.259)
January_Average_Temperature -0.492*
(0.266)
-5.618***
(0.671)
-0.334**
(0.161)
-3.066***
(0.733)
-0.166
(0.101)
3.407***
(0.378)
-0.104*
(0.062)
-1.119***
(0.139)
-0.042
(0.081)
-0.338**
(0.139)
Eight_Violent_Crime -0.002
(0.016)
0.382***
(0.045)
-0.014**
(0.006)
0.222***
(0.042)
-0.015***
(0.004)
-0.066**
(0.028)
-0.008**
(0.003)
-0.156***
(0.013)
-0.001
(0.005)
0.030
(0.047)
Eight_Property_Crime 0.002
(0.003)
-0.041***
(0.004)
0.001
(0.001)
-0.022***
(0.004)
2.781×10
-4

(0.001)
-0.013***
(0.003)
0.001
(0.001)
0.004***
(0.002)
0.002*
(0.001)
-0.005
(0.014)
New_England 1.696
(10.368)
0.334
(8.191)
3.086
(6.191)
-0.089
(4.460)
1.440
(2.657)

Middle_Atlantic -25.981***
(7.985)
-10.282**
(5.080)
-2.221
(3.681)
-1.163
(2.088)
1.445
(2.007)

East_North_Central -21.577**
(8.263)
-7.227
(4.792)
1.392
(3.712)
2.328
(3.208)
8.170*
(4.636)

West_North_Central -14.332*
(8.437)
-13.299**
(6.510)
-4.631
(4.297)
0.981
(4.243)
8.044**
(3.250)

South_Atlantic 11.346*
(6.290)
5.262*
(2.984)
4.512*
(2.352)
0.725
(1.699)
0.441
(1.775)

East_South_Central 9.499
(10.819)
10.242**
(4.653)
7.365**
(3.196)
2.267
(2.467)
3.690
(3.491)

West_South_Central 3.981
(8.231)
2.185
(3.867)
1.803
(2.680)
-1.243
(2.029)
1.887
(1.989)

Mountain -14.331*
(7.315)
-10.590***
(3.922)
-6.729**
(2.659)
-4.100**
(1.762)
-1.664
(2.073)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 234 234 198 198 165 165 84 84
N 1,291 1,291 1,093 1,093 811 811 644 644 195 195
R
2
0.1648 0.3965 0.1937 0.3640 0.1700 0.3719 0.1195 0.3604 0.1417 0.4834
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

113

Appendix C.6. Regression Results for the Migration Models – Group C (Age 45-54) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_C Nine_BAPLUS_Inmig_Pc_C Nine_MAPLUS_Inmig_Pc_C Nine_SCC_Inmig_Pc_C Nine_BOH_Inmig_Pc_C
Constant -53.588
(93.034)
-86843.600***
(11603.830)
-112.081***
(37.029)
-660.499
(678.596)
-96.265**
(37.515)
828.054***
(110.552)
-16.117
(26.503)
265.660*
(159.727)
63.886*
(33.530)
1625.942***
(194.629)
Eight_MSA_Pop_Den_BEA -0.017***
(0.007)
-4.236***
(0.559)
-0.002
(0.002)
0.007
(0.010)
-0.003*
(0.002)
-0.150***
(0.017)
0.001
(0.002)
0.096***
(0.033)
0.003
(0.003)
0.320***
(0.060)
Eight_MSA_Pop_BEA_M 5.736*
(3.044)
-800.140***
(114.003)
2.289*
(1.159)
-0.498
(12.637)
0.378
(0.927)
35.991***
(1.800)
1.934**
(0.939)
-4.921
(3.109)
-0.892
(1.112)
14.308***
(0.780)
Eight_MSA_Pop_BEA_L 6.646
(4.352)
-691.753***
(98.684)
4.855***
(1.635)
0.482
(13.580)
1.627
(1.493)
37.083***
(2.893)
1.609
(1.180)
0.778
(3.018)
-0.538
(1.553)
11.749
(9.459)
Eight_MSA_Inc_BEA_Ln 7.231
(8.807)
8196.632***
(1094.578)
12.605***
(3.770)
65.856
(62.467)
11.847***
(3.664)
-67.715***
(10.451)
3.416
(2.652)
-24.304
(15.662)
-4.750
(3.368)
-152.040***
(18.123)
Eight_Pct_M_Vo -0.438**
(0.191)
-0.098
(0.348)
-0.355***
(0.115)
-0.260
(0.231)
-0.203
(0.138)
-0.409*
(0.225)
0.006
(0.136)
-0.191
(0.300)
0.116
(0.088)
-0.010
(0.272)
Eight_Emp_Pop_Ratio 0.215
(0.193)
0.231
(0.244)
0.106
(0.095)
0.094
(0.110)
0.008
(0.072)
0.004
(0.098)
-0.095
(0.099)
-0.142
(0.167)
-0.131*
(0.071)
-0.232
(0.375)
Eight_PUMA_Self_Emp_Pop_Ratio 0.590*
(0.312)
0.720**
(0.357)
0.814***
(0.166)
1.052***
(0.236)
0.210*
(0.112)
0.226
(0.175)
0.247*
(0.133)
0.119
(0.226)
0.185
(0.116)
0.085
(0.353)
Eight_PUMA_Mttw 0.105
(0.205)
-0.223
(0.291)
-0.470***
(0.130)
-0.555**
(0.236)
-0.395***
(0.087)
-0.562***
(0.146)
-0.141
(0.112)
-0.225
(0.196)
0.028
(0.145)
0.036
(0.328)
January_Average_Temperature -0.139
(0.151)
89.222***
(12.001)
-0.159***
(0.057)
0.208
(0.921)
-0.183***
(0.046)
-2.328***
(0.141)
-0.118***
(0.039)
-0.125
(0.081)
-0.089
(0.053)
-4.156***
(0.806)
Eight_Violent_Crime 0.021*
(0.011)
-4.963***
(0.658)
-0.001
(0.003)
-0.033
(0.056)
0.001
(0.003)
-0.011**
(0.005)
-0.001
(0.002)
-0.037***
(0.007)
0.001
(0.002)
0.058***
(0.018)
Eight_Property_Crime -0.001
(0.002)
0.135***
(0.019)
8.270×10
-5

(0.001)
-0.002
(0.004)
-6.890×10
-5

(0.001)
0.005***
(0.001)
1.401×10
-4
(4.902×10
-4
)
0.005***
(4.762×10
-4
)
4.956×10
-4

(0.001)
0.015***
(0.003)
New_England 12.055***
(4.206)
4.139
(2.577)
-3.398
(2.461)
-1.942
(1.257)
1.373
(3.472)

Middle_Atlantic -7.123*
(4.225)
-2.432
(1.807)
-2.082
(1.596)
-1.213
(1.233)
0.516
(1.431)

East_North_Central 2.409
(4.918)
-0.986
(2.316)
-3.297*
(1.866)
0.636
(1.552)
-0.448
(2.487)

West_North_Central 5.839
(5.367)
1.882
(2.221)
-4.166*
(2.243)
3.175
(3.462)
-2.047
(2.211)

South_Atlantic 14.830***
(4.502)
4.527***
(1.529)
0.409
(1.162)
1.075
(0.974)
-0.120
(1.462)

East_South_Central 15.205*
(8.357)
1.381
(2.614)
-3.382*
(2.027)
-2.340*
(1.389)
-2.721
(1.933)

West_South_Central 8.145
(5.049)
0.686
(1.843)
-2.424*
(1.410)
0.312
(1.180)
0.603
(1.723)

Mountain 2.917
(4.938)
0.593
(1.848)
-3.764**
(1.656)
0.099
(1.463)
1.665
(2.021)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 219 219 172 172 143 143 67 67
N 1,264 1,264 967 967 643 643 442 442 109 109
R
2
0.1493 0.3451 0.1471 0.3227 0.1096 0.3372 0.0564 0.2938 0.2045 0.7309
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

114

Appendix D. Regression Results for the Migration Models – PUMA Population Density (without amenity variables)
Appendix D.1. Regression Results for the Migration Models – Group A (Age 25-34) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_A Six_BAPLUS_Inmig_Pc_A Six_MAPLUS_Inmig_Pc_A Six_SCC_Inmig_Pc_A Six_BOH_Inmig_Pc_A
Constant -187.345
(193.531)
-879.961***
(183.004)
-729.224***
(158.652)
-998.533***
(150.157)
-404.350***
(110.938)
-416.153***
(115.752)
-235.998***
(71.613)
-505.480***
(96.365)
44.475
(29.869)
-70.779**
(34.115)
Five_PUMA_Pop_Den 0.004**
(0.001)
0.005**
(0.003)
0.005***
(0.002)
0.005**
(0.002)
0.002***
(0.001)
0.002**
(0.001)
18.497×10
-4
***
(4.682×10
-4
)
20.045×10
-4
***
(6.436×10
-4
)
5.714×10
-4
***
(1.398×10
-4
)
5.246×10
-4
***
(1.744×10
-4
)
Five_MSA_Inc_BEA_Ln 11.039
(20.174)
91.228***
(20.823)
61.458***
(16.997)
93.857***
(18.803)
38.320***
(10.949)
42.244***
(13.425)
21.425***
(7.295)
48.621***
(10.901)
-4.255
(2.953)
7.331*
(4.123)
Five_Pct_M_Vo -3.884***
(0.650)
-3.906***
(1.228)
-2.773***
(0.547)
-4.081***
(1.130)
-1.177***
(0.314)
-1.835***
(0.611)
-0.641**
(0.251)
-0.972**
(0.484)
-0.230**
(0.103)
-0.475*
(0.251)
Five_Emp_Pop_Ratio 4.297***
(0.506)
3.818***
(0.547)
3.275***
(0.486)
3.327***
(0.597)
1.044***
(0.248)
1.131***
(0.308)
0.973***
(0.205)
1.012***
(0.253)
0.203***
(0.065)
0.192
(0.127)
Five_PUMA_Self_Emp_Pop_Ratio -3.114***
(0.743)
-0.576
(0.669)
0.222
(0.481)
2.181***
(0.571)
-0.003
(0.245)
0.666*
(0.398)
-0.379
(0.237)
0.155
(0.418)
-0.027
(0.159)
-0.132
(0.311)
Five_PUMA_Mttw -0.386
(1.105)
-2.741**
(1.274)
-2.228***
(0.585)
-4.213***
(0.947)
-1.338***
(0.298)
-2.399***
(0.696)
-0.858***
(0.219)
-1.499***
(0.465)
0.006
(0.066)
-0.127
(0.131)
New_England 9.237
(20.277)
6.378
(16.629)
8.023
(10.815)
-1.307
(5.803)
1.004
(1.710)

Middle_Atlantic -54.064***
(16.730)
-15.559*
(8.273)
-5.022*
(2.998)
-5.880*
(3.392)
0.220
(1.571)

East_North_Central 0.169
(26.058)
9.327
(12.299)
4.230
(4.898)
-2.325
(3.783)
-2.101
(1.675)

West_North_Central -20.271
(16.829)
-18.326**
(9.291)
-9.261**
(4.234)
-10.093**
(4.307)
-3.482*
(1.852)

South_Atlantic 13.492
(17.473)
2.676
(9.473)
0.690
(4.871)
-3.342
(3.590)
-2.084
(1.607)

East_South_Central 15.998
(21.047)
5.200
(8.041)
-0.644
(3.447)
-7.974**
(3.616)
-2.949*
(1.685)

West_South_Central 14.356
(16.636)
-0.064
(7.504)
-1.221
(3.016)
-5.185*
(3.080)
-2.159
(1.511)

Mountain -0.375
(17.229)
-17.783**
(8.363)
-9.556***
(3.306)
-9.890**
(4.087)
-2.436
(2.199)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 402 402 392 392 325 325 284 284 164 164
N 1,647 1,647 1,557 1,557 1,226 1,226 1,120 1,120 465 465
R
2
0.1698 0.4104 0.2362 0.4188 0.1990 0.3821 0.1430 0.3246 0.1020 0.2760
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

115

Appendix D.2. Regression Results for the Migration Models – Group B (Age 35-44) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_B Six_BAPLUS_Inmig_Pc_B Six_MAPLUS_Inmig_Pc_B Six_SCC_Inmig_Pc_B Six_BOH_Inmig_Pc_B
Constant -249.913**
(121.276)
-1152.252***
(131.978)
-456.594***
(89.454)
-405.034***
(68.538)
-272.418***
(55.660)
13.436
(38.666)
-167.121***
(31.517)
-246.303***
(24.432)
-59.039**
(28.920)
36.060
(29.562)
Five_PUMA_Pop_Den 3.482×10
-4

(6.342×10
-4
)
0.001
(0.001)
11.835×10
-4
**
(4.617×10
-4
)
14.740×10
-4
**
(6.167×10
-4
)
8.116×10
-4
***
(2.041×10
-4
)
9.657×10
-4
***
(2.836×10
-4
)
4.515×10
-4
**
(1.910×10
-4
)
4.738×10
-4
**
(2.343×10
-4
)
3.269×10
-4
***
(0.624×10
-4
)
3.787×10
-4
***
(0.694×10
-4
)
Five_MSA_Inc_BEA_Ln 17.316
(12.911)
112.931***
(15.077)
37.008***
(9.371)
35.350***
(8.656)
24.249***
(5.303)
-1.820
(4.643)
16.127***
(3.254)
24.996***
(3.266)
6.766**
(2.602)
-2.717*
(1.505)
Five_Pct_M_Vo -1.327***
(0.354)
-0.863
(0.656)
-0.772***
(0.267)
-0.857
(0.596)
-0.440***
(0.162)
-0.525
(0.370)
-0.015
(0.122)
-0.318
(0.244)
-0.013
(0.089)
-0.076
(0.256)
Five_Emp_Pop_Ratio 1.931***
(0.312)
1.828***
(0.381)
1.536***
(0.231)
1.613***
(0.307)
0.680***
(0.140)
0.720***
(0.155)
0.290***
(0.088)
0.302**
(0.117)
-0.025
(0.107)
0.015
(0.210)
Five_PUMA_Self_Emp_Pop_Ratio -0.737*
(0.387)
0.140
(0.419)
0.552**
(0.248)
1.331***
(0.357)
0.041
(0.194)
0.505**
(0.211)
0.099
(0.135)
0.172
(0.227)
0.063
(0.106)
0.220
(0.202)
Five_PUMA_Mttw 1.261*
(0.649)
0.025
(0.764)
-0.170
(0.297)
-1.086**
(0.432)
-0.329**
(0.145)
-0.731***
(0.250)
-0.342***
(0.119)
-0.503**
(0.196)
-0.104
(0.094)
-0.087
(0.156)
New_England 7.138
(7.694)
-1.514
(4.900)
-1.030
(3.545)
-0.987
(1.659)
-1.557
(1.273)

Middle_Atlantic -21.126**
(8.612)
-8.751***
(3.307)
-3.269
(2.408)
-2.932*
(1.763)
0.469
(1.872)

East_North_Central 5.286
(13.343)
3.441
(5.416)
1.335
(2.849)
0.167
(2.143)
0.457
(1.603)

West_North_Central -1.985
(8.760)
-9.353**
(3.596)
-3.995
(2.437)
-2.946
(2.629)
1.260
(2.083)

South_Atlantic 29.091***
(8.529)
5.279
(3.687)
0.636
(2.750)
-0.508
(1.720)
0.244
(1.320)

East_South_Central 25.872**
(11.283)
7.134*
(4.124)
4.018
(2.751)
-1.270
(2.072)
-1.687
(1.487)

West_South_Central 26.626***
(8.869)
5.521
(3.425)
0.677
(1.861)
0.048
(1.720)
0.671
(1.579)

Mountain 8.384
(10.618)
-4.757
(4.921)
-6.065**
(2.616)
-2.866
(2.098)
1.965
(2.411)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 400 400 362 362 295 295 235 235 116 116
N 1,632 1,632 1,412 1,412 1,056 1,056 797 797 253 253
R
2
0.1722 0.4070 0.2102 0.3870 0.1710 0.3985 0.0879 0.3199 0.0895 0.4993
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

116

Appendix D.3. Regression Results for the Migration Models – Group C (Age 45-54) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_C Six_BAPLUS_Inmig_Pc_C Six_MAPLUS_Inmig_Pc_C Six_SCC_Inmig_Pc_C Six_BOH_Inmig_Pc_C
Constant 54.541
(67.926)
-10.930
(113.849)
-112.274***
(33.355)
-441.285***
(48.966)
-74.976***
(27.062)
-217.978***
(40.201)
-85.163***
(20.931)
-109.219***
(41.426)
-12.480
(19.081)
190.525***
(69.677)
Five_PUMA_Pop_Den -0.501×10
-4

(3.704×10
-4
)
5.631×10
-4

(5.420×10
-4
)
6.594×10
-4
***
(1.654×10
-4
)
7.535×10
-4
***
(2.084×10
-4
)
6.137×10
-4
***
(1.873×10
-4
)
5.349×10
-4
***
(1.971×10
-4
)
1.713×10
-4

(1.436×10
-4
)
0.113×10
-4

(1.127×10
-4
)
1.935×10
-4
*
(1.136×10
-4
)
0.344
(2.546)
Five_MSA_Inc_BEA_Ln -7.715
(6.941)
4.141
(12.702)
8.866***
(3.398)
42.983***
(5.773)
7.338***
(2.682)
22.258***
(4.749)
8.090***
(2.174)
11.334**
(4.831)
1.468
(1.889)
-17.081**
(6.528)
Five_Pct_M_Vo -0.736***
(0.264)
-0.388
(0.569)
-0.350***
(0.131)
-0.475
(0.309)
-0.205**
(0.086)
-0.435**
(0.193)
-0.088
(0.076)
-0.371**
(0.160)
-0.133
(0.094)
-0.321
(0.324)
Five_Emp_Pop_Ratio 0.941***
(0.216)
0.787**
(0.304)
0.569***
(0.123)
0.420**
(0.191)
0.179**
(0.085)
0.108
(0.140)
0.198***
(0.066)
0.129
(0.123)
0.093
(0.086)
0.038
(0.233)
Five_PUMA_Self_Emp_Pop_Ratio 0.018
(0.343)
0.526
(0.436)
0.512***
(0.164)
1.000***
(0.264)
0.333***
(0.123)
0.599***
(0.157)
0.141
(0.101)
0.285
(0.179)
0.146
(0.121)
0.070
(0.292)
Five_PUMA_Mttw 0.638
(0.422)
0.291
(0.592)
-0.150
(0.136)
-0.545***
(0.200)
-0.215**
(0.092)
-0.452**
(0.181)
-0.175**
(0.078)
-0.306**
(0.150)
-0.153**
(0.067)
-0.278*
(0.144)
New_England 6.593
(5.651)
1.028
(2.333)
3.609**
(1.762)
-0.321
(1.571)
-0.143
(1.069)

Middle_Atlantic -12.623**
(5.390)
-2.447
(1.586)
0.061
(0.908)
-1.289
(0.994)
1.504
(1.016)

East_North_Central 2.633
(8.068)
1.287
(2.363)
2.878*
(1.520)
0.805
(1.573)
4.460**
(1.922)

West_North_Central 6.802
(6.572)
-0.223
(3.378)
0.566
(2.111)
-0.705
(1.506)
4.491***
(1.225)

South_Atlantic 17.240***
(5.794)
4.301**
(1.997)
1.907
(1.537)
-0.426
(0.939)
0.290
(0.839)

East_South_Central 25.506***
(8.145)
6.103***
(2.180)
3.029*
(1.561)
4.134***
(1.443)
9.752***
(3.198)

West_South_Central 17.005***
(5.400)
4.685**
(2.107)
2.528**
(1.029)
1.548
(1.103)
1.277
(1.224)

Mountain 8.776
(6.639)
2.102
(2.392)
1.359
(1.440)
-0.071
(1.460)
-0.196
(1.125)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 401 401 349 349 284 284 222 222 87 87
N 1,610 1,610 1,284 1,284 887 887 615 615 172 172
R
2
0.1230 0.3956 0.1068 0.3458 0.0967 0.3862 0.0752 0.3516 0.1992 0.4560
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

117

Appendix D.4. Regression Results for the Migration Models – Group A (Age 25-34) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_A Nine_BAPLUS_Inmig_Pc_A Nine_MAPLUS_Inmig_Pc_A Nine_SCC_Inmig_Pc_A Nine_BOH_Inmig_Pc_A
Constant -410.722**
(202.717)
-4131.886***
(245.005)
-723.978***
(184.465)
-2495.233***
(238.170)
-384.208***
(101.141)
-461.153***
(115.830)
-333.614***
(73.022)
-208.199***
(41.382)
-68.047*
(40.561)
-583.759***
(41.465)
Eight_PUMA_Pop_Den 0.006
(0.002)
0.008**
(0.004)
0.005**
(0.002)
0.006*
(0.003)
0.003**
(0.001)
0.003*
(0.001)
0.002*
(0.001)
0.002
(0.001)
5.897×10
-4
***
(2.191×10
-4
)
4.578×10
-4
**
(2.169×10
-4
)
Eight_MSA_Inc_BEA_Ln 46.540**
(22.344)
409.450***
(26.324)
69.288***
(20.286)
246.854***
(23.757)
38.645***
(10.489)
50.178***
(13.800)
34.089***
(7.593)
25.824***
(5.406)
7.948*
(4.543)
59.266***
(5.065)
Eight_Pct_M_Vo -5.006***
(1.085)
-6.378***
(2.408)
-3.401***
(0.850)
-5.510***
(1.854)
-1.532***
(0.473)
-2.352**
(0.969)
-1.057***
(0.361)
-2.000***
(0.732)
-0.335**
(0.168)
-1.143**
(0.450)
Eight_Emp_Pop_Ratio 2.627***
(0.566)
2.665***
(0.730)
2.053***
(0.600)
2.145**
(0.928)
0.630**
(0.264)
0.663
(0.417)
0.442**
(0.178)
0.422
(0.298)
-0.009
(0.135)
0.085
(0.320)
Eight_PUMA_Self_Emp_Pop_Ratio -4.168***
(0.995)
-1.901
(1.162)
-0.413
(0.702)
1.016
(0.871)
-0.441
(0.313)
0.152
(0.523)
-0.436
(0.285)
-0.091
(0.443)
0.213
(0.230)
-0.087
(0.411)
Eight_PUMA_Mttw -2.251**
(1.100)
-4.471***
(1.340)
-2.524***
(0.647)
-5.185***
(0.927)
-1.190***
(0.294)
-2.480***
(0.515)
-0.934***
(0.265)
-1.855***
(0.369)
-0.174
(0.174)
-0.602*
(0.309)
New_England 37.396
(23.798)
20.362
(19.188)
7.694
(7.774)
6.393
(7.064)
2.560
(2.093)

Middle_Atlantic -47.402***
(14.312)
-19.307**
(8.062)
-5.396
(4.122)
-8.983***
(3.208)
-1.445
(1.893)

East_North_Central 21.469
(25.128)
23.385
(15.644)
14.694**
(7.102)
7.716
(5.495)
3.924
(2.561)

West_North_Central -1.153
(16.743)
-10.405
(9.634)
-5.417
(4.962)
-6.764**
(3.352)
1.131
(2.491)

South_Atlantic 16.539
(13.901)
4.223
(9.205)
2.548
(4.507)
-2.366
(3.172)
-1.096
(2.310)

East_South_Central 35.758**
(16.622)
12.237*
(6.998)
5.795*
(3.346)
-2.951
(2.626)
1.330
(2.410)

West_South_Central 13.688
(11.753)
-4.421
(6.364)
-0.351
(3.286)
-6.346***
(2.217)
-1.877
(1.638)

Mountain -9.262
(13.919)
-21.039***
(7.574)
-7.445**
(3.374)
-8.028***
(2.879)
-2.305
(2.438)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 403 403 384 384 317 317 285 285 151 151
N 1,708 1,708 1,614 1,614 1,251 1,251 1,129 1,129 424 424
R
2
0.1737 0.3746 0.1938 0.3467 0.1791 0.3277 0.1727 0.3368 0.1380 0.3654
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

118

Appendix D.5. Regression Results for the Migration Models – Group B (Age 35-44) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_B Nine_BAPLUS_Inmig_Pc_B Nine_MAPLUS_Inmig_Pc_B Nine_SCC_Inmig_Pc_B Nine_BOH_Inmig_Pc_B
Constant -226.026**
(100.766)
-1576.647***
(80.067)
-363.954***
(67.619)
-848.933***
(73.221)
-215.002***
(50.518)
-531.291***
(54.224)
-127.522***
(33.783)
-137.391***
(22.172)
-33.540
(29.292)
99.104***
(27.569)
Eight_PUMA_Pop_Den -0.225
(2.961)
6.104×10
-4

(5.193×10
-4
)
6.085×10
-4
*
(3.249×10
-4
)
8.012×10
-4
*
(4.388×10
-4
)
5.571×10
-4
***
(1.912×10
-4
)
5.354×10
-4
**
(2.098×10
-4
)
3.691×10
-4
**
(1.514×10
-4
)
3.511×10
-4
**
(14.510×10
-4
)
1.996×10
-4
***
(0.620×10
-4
)
1.953×10
-4
***
(0.565×10
-4
)
Eight_MSA_Inc_BEA_Ln 24.221**
(10.518)
159.174***
(9.860)
34.200***
(6.735)
83.896***
(8.031)
22.106***
(5.038)
55.189***
(6.569)
13.222***
(3.537)
16.517***
(2.800)
4.634
(3.018)
-8.308***
(3.104)
Eight_Pct_M_Vo -1.407***
(0.401)
-1.026
(0.875)
-1.007***
(0.221)
-1.134**
(0.493)
-0.645***
(0.164)
-1.051***
(0.352)
-0.455***
(0.113)
-0.582**
(0.225)
0.007
(0.117)
-0.062
(0.247)
Eight_Emp_Pop_Ratio 0.602
(0.408)
0.457
(0.568)
0.651***
(0.163)
0.637***
(0.211)
0.172
(0.117)
0.057
(0.163)
0.146
(0.100)
0.105
(0.130)
-0.119
(0.114)
-0.055
(0.207)
Eight_PUMA_Self_Emp_Pop_Ratio -0.410
(0.433)
0.764
(0.481)
0.726**
(0.325)
1.622***
(0.455)
0.255
(0.217)
0.663**
(0.261)
0.129
(0.150)
0.170
(0.230)
0.221
(0.142)
0.233
(0.262)
Eight_PUMA_Mttw 0.356
(0.484)
-0.501
(0.563)
-0.599***
(0.222)
-1.470***
(0.335)
-0.514***
(0.151)
-1.063***
(0.271)
-0.315***
(0.087)
-0.778***
(0.171)
-0.087
(0.113)
-0.083
(0.331)
New_England 13.185
(8.255)
4.857
(6.095)
4.873
(4.863)
1.249
(3.204)
3.311
(2.879)

Middle_Atlantic -20.632***
(7.463)
-5.081
(3.110)
-0.374
(2.353)
-0.693
(1.598)
0.280
(1.509)

East_North_Central 5.305
(11.073)
4.446
(4.366)
5.279**
(2.507)
6.420***
(2.153)
8.005***
(2.552)

West_North_Central 1.350
(8.181)
-6.121
(3.888)
-2.898
(2.492)
1.083
(2.501)
6.640**
(2.651)

South_Atlantic 17.508**
(8.017)
5.340
(4.155)
2.909
(2.779)
1.220
(1.567)
0.612
(1.401)

East_South_Central 16.146
(11.271)
10.511***
(3.510)
4.365*
(2.527)
2.754*
(1.618)
3.443
(2.758)

West_South_Central 12.610*
(7.003)
2.779
(2.844)
0.853
(2.027)
-0.357
(1.425)
1.709
(1.648)

Mountain -0.451
(6.976)
-5.713**
(2.857)
-5.233***
(1.925)
-1.456
(1.766)
1.987
(3.942)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 398 398 345 345 278 278 234 234 105 105
N 1,679 1,679 1,395 1,395 1,014 1,014 805 805 231 231
R
2
0.1175 0.3552 0.1710 0.3592 0.1475 0.3702 0.1144 0.3576 0.1062 0.5009
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

119

Appendix D.6. Regression Results for the Migration Models – Group C (Age 45-54) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_C Nine_BAPLUS_Inmig_Pc_C Nine_MAPLUS_Inmig_Pc_C Nine_SCC_Inmig_Pc_C Nine_BOH_Inmig_Pc_C
Constant -44.047
(73.841)
-1103.992***
(54.971)
-158.952***
(37.871)
-312.765***
(24.573)
-93.970***
(34.819)
-119.958***
(15.998)
-51.021**
(24.361)
-135.057**
(61.250)
18.863
(33.345)
94.781
(66.334)
Eight_PUMA_Pop_Den 1.074×10
-4

(1.739×10
-4
)
4.969×10
-4
**
(2.132×10
-4
)
3.323×10
-4
***
(0.900×10
-4
)
4.331×10
-4
***
(1.221×10
-4
)
3.884×10
-4
***
(0.654×10
-4
)
4.296×10
-4
***
(0.582×10
-4
)
0.785×10
-4

(0.477×10
-4
)
0.639×10
-4

(0.438×10
-4
)
0.873×10
-4

(0.781×10
-4
)
0.477×10
-4

(1.206×10
-4
)
Eight_MSA_Inc_BEA_Ln 7.108
(7.302)
110.618***
(6.172)
16.408***
(3.832)
32.012***
(2.849)
10.997***
(3.432)
14.331***
(1.972)
6.664**
(2.563)
16.090**
(6.804)
0.392
(3.463)
-6.950
(6.835)
Eight_Pct_M_Vo -0.856***
(0.255)
-0.758
(0.573)
-0.414***
(0.117)
-0.457*
(0.272)
-0.210
(0.131)
-0.482*
(0.256)
-0.119
(0.117)
-0.304
(0.298)
0.096
(0.087)
-0.101
(0.261)
Eight_Emp_Pop_Ratio 0.172
(0.176)
0.082
(0.241)
0.103
(0.087)
0.056
(0.106)
-0.013
(0.075)
-0.028
(0.099)
-0.099
(0.097)
-0.189
(0.156)
-0.251***
(0.083)
-0.094
(0.333)
Eight_PUMA_Self_Emp_Pop_Ratio 0.203
(0.300)
0.544
(0.352)
0.657***
(0.142)
1.024***
(0.246)
0.098
(0.094)
0.196
(0.176)
0.084
(0.106)
0.102
(0.220)
0.105
(0.121)
-0.086
(0.332)
Eight_PUMA_Mttw 0.115
(0.310)
-0.051
(0.429)
-0.383***
(0.106)
-0.556***
(0.181)
-0.443***
(0.071)
-0.572***
(0.103)
-0.151*
(0.080)
-0.315*
(0.185)
-0.051
(0.131)
-0.144
(0.382)
New_England 10.340**
(4.285)
3.054
(2.684)
0.566
(1.810)
-0.112
(0.960)
2.546
(2.099)

Middle_Atlantic -9.073**
(4.230)
-0.445
(2.027)
1.055
(1.669)
0.598
(1.406)
2.914
(1.816)

East_North_Central 9.928**
(5.003)
5.413**
(2.229)
3.190*
(1.624)
3.101**
(1.257)
3.171**
(1.528)

West_North_Central 10.321**
(4.359)
4.800***
(1.796)
3.514*
(1.847)
3.159
(2.065)
5.686*
(3.028)

South_Atlantic 17.839***
(4.140)
4.562***
(1.582)
2.355*
(1.404)
0.503
(0.851)
0.296
(1.166)

East_South_Central 16.249*
(9.376)
2.446
(2.412)
-0.064
(1.745)
-1.145
(0.969)
-2.127
(1.545)

West_South_Central 12.392***
(3.863)
2.062
(1.686)
0.242
(1.300)
0.669
(1.137)
2.293
(1.881)

Mountain 8.591*
(4.373)
1.787
(1.741)
-0.464
(1.455)
0.556
(1.392)
3.550*
(1.937)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 399 399 322 322 243 243 193 193 87 87
N 1,648 1,648 1,236 1,236 799 799 546 546 139 139
R
2
0.1010 0.3364 0.1196 0.3265 0.0959 0.3646 0.0466 0.3240 0.1676 0.7379
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

120

Appendix E. Regression Results for the Migration Models – PUMA Population Density (with amenity variables)
Appendix E.1. Regression Results for the Migration Models – Group A (Age 25-34) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_A Six_BAPLUS_Inmig_Pc_A Six_MAPLUS_Inmig_Pc_A Six_SCC_Inmig_Pc_A Six_BOH_Inmig_Pc_A
Constant -192.643
(264.853)
-3625.412***
(1052.930)
-552.658***
(163.276)
-3399.141***
(911.285)
-305.579***
(102.468)
-8120.691***
(848.246)
-203.660***
(76.363)
-4280.037**
(2151.313)
78.490*
(43.297)
1166.560***
(179.644)
Five_PUMA_Pop_Den 0.003***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.002***
(0.001)
0.002***
(0.001)
0.002***
(2.807×10
-4
)
0.002***
(3.883×10
-4
)
4.872×10
-4
***
(7.630×10
-5
)
4.658×10
-4
***
(1.075×10
-4
)
Five_MSA_Pop_BEA_M 18.551**
(8.745)
-221.428***
(21.914)
17.230***
(5.483)
-78.720***
(18.976)
6.110**
(3.079)
-115.610***
(14.334)
3.706
(2.437)
-57.580***
(15.353)
2.808**
(1.212)
47.585***
(8.187)
Five_MSA_Pop_BEA_L -8.986
(12.650)
-396.925***
(44.898)
8.502
(9.241)
-123.426***
(39.275)
8.612
(6.213)
-191.768***
(28.823)
4.689
(3.787)
-176.840***
(59.609)
2.976*
(1.659)
96.908***
(16.521)
Five_MSA_Inc_BEA_Ln 16.403
(26.432)
313.355***
(96.711)
52.299***
(16.657)
309.321***
(84.593)
32.049***
(10.541)
758.895***
(78.408)
19.798**
(7.577)
391.321*
(207.974)
-6.243
(4.009)
-104.708***
(16.135)
Five_Pct_M_Vo -3.560***
(0.479)
-2.900***
(0.865)
-2.822***
(0.494)
-3.083***
(0.790)
-1.299***
(0.343)
-1.341**
(0.553)
-0.615**
(0.278)
-0.708
(0.482)
-0.232*
(0.120)
-0.483*
(0.272)
Five_Emp_Pop_Ratio 4.223***
(0.567)
4.049***
(0.620)
2.950***
(0.432)
2.987***
(0.497)
1.107***
(0.271)
1.180***
(0.367)
1.006***
(0.202)
1.040***
(0.242)
0.098
(0.075)
0.148
(0.133)
Five_PUMA_Self_Emp_Pop_Ratio -2.187***
(0.692)
-0.677
(0.714)
0.751
(0.455)
2.396***
(0.556)
0.057
(0.251)
0.830**
(0.410)
-0.314
(0.288)
0.220
(0.435)
-0.002
(0.186)
-0.132
(0.319)
Five_PUMA_Mttw -1.231
(0.850)
-3.850***
(0.869)
-2.984***
(0.562)
-4.830***
(0.865)
-1.756***
(0.441)
-2.671***
(0.837)
-1.151***
(0.294)
-1.785***
(0.538)
-0.067
(0.087)
-0.111
(0.139)
January_Average_Temperature -0.980**
(0.421)
-11.628***
(2.184)
-0.770**
(0.319)
-0.978
(1.815)
-0.345*
(0.182)
13.174***
(1.932)
-0.259**
(0.124)
14.327***
(1.347)
-0.111**
(0.053)
-8.801***
(1.486)
Five_Violent_Crime -0.006
(0.023)
2.239***
(0.207)
-0.010
(0.012)
0.516***
(0.180)
-0.008
(0.007)
-0.811***
(0.188)
-0.003
(0.006)
0.192
(0.190)
-0.001
(0.003)
-0.133***
(0.021)
Five_Property_Crime 0.002
(0.005)
0.005
(0.006)
-0.003
(0.003)
0.009*
(0.005)
-0.002
(0.002)
0.038***
(0.004)
2.121×10
-4

(0.002)
-0.108***
(0.027)
-3.160×10
-5

(0.001)
0.078***
(0.013)
New_England -9.861
(21.902)
-5.440
(19.302)
1.731
(13.024)
-5.522
(7.019)
-0.374
(2.251)

Middle_Atlantic -57.873***
(17.463)
-28.121**
(12.450)
-15.287**
(6.474)
-10.256**
(4.448)
-1.137
(2.261)

East_North_Central -55.214***
(14.237)
-27.893***
(9.677)
-13.746**
(5.346)
-14.368***
(4.720)
-5.478***
(1.860)

West_North_Central -57.493***
(15.796)
-43.667***
(11.151)
-21.326***
(6.766)
-17.913***
(6.102)
-5.251**
(2.247)

South_Atlantic 13.779
(12.502)
2.694
(7.468)
0.229
(4.276)
-3.797
(3.215)
-2.368
(1.827)

East_South_Central 2.727
(17.415)
-7.378
(7.747)
-4.501
(3.445)
-13.175***
(3.958)
-5.140**
(2.068)

West_South_Central 17.069
(14.520)
0.406
(8.127)
-2.192
(3.910)
-7.001*
(3.820)
-2.420
(1.962)

Mountain -24.303
(15.043)
-29.446***
(9.898)
-13.606***
(4.392)
-13.640**
(5.352)
-2.876
(2.187)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 277 277 276 276 241 241 218 218 136 136
N 1,342 1,342 1,273 1,273 1,019 1,019 947 947 406 406
R
2
0.2199 0.4188 0.2684 0.4281 0.2316 0.3859 0.1618 0.3305 0.1040 0.2590
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

121

Appendix E.2. Regression Results for the Migration Models – Group B (Age 35-44) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_B Six_BAPLUS_Inmig_Pc_B Six_MAPLUS_Inmig_Pc_B Six_SCC_Inmig_Pc_B Six_BOH_Inmig_Pc_B
Constant -414.485**
(161.868)
-15775.900***
(668.201)
-446.374***
(103.216)
951.544*
(562.124)
-266.187***
(59.264)
1860.489***
(465.326)
-170.546***
(40.749)
2293.814***
(292.818)
-65.156**
(29.369)
200.351
(614.945)
Five_PUMA_Pop_Den 2.289×10
-4

(2.590×10
-4
)
6.455×10
-4

(2.483×10
-4
)
9.454×10
-4
***
(1.302×10
-4
)
0.001***
(2.126×10
-4
)
0.001***
(9.430×10
-5
)
0.001***
(1.391×10
-4
)
3.833×10
-4
***
(9.010×10
-5
)
3.828×10
-4
***
(9.340×10
-5
)
3.506×10
-4
***
(6.190×10
-5
)
4.328×10
-4
***
(5.400×10
-5
)
Five_MSA_Pop_BEA_M 0.887
(5.500)
-250.652***
(13.082)
5.591**
(2.812)
-47.902***
(10.374)
2.541
(1.943)
28.366***
(8.743)
3.190**
(1.426)
14.847***
(1.568)
1.162
(1.282)
3.494
(9.885)
Five_MSA_Pop_BEA_L -14.275*
(7.816)
-502.725***
(24.514)
2.802
(4.933)
-70.056***
(19.079)
0.864
(3.768)
104.609***
(19.904)
0.992
(1.896)
48.048***
(4.922)
0.612
(1.845)
12.883
(32.527)
Five_MSA_Inc_BEA_Ln 30.387*
(16.190)
1454.340***
(60.505)
37.902***
(10.303)
-97.788*
(51.259)
24.777***
(5.712)
-177.212***
(43.190)
16.189***
(3.850)
-233.519***
(29.297)
7.080***
(2.663)
-20.161
(61.410)
Five_Pct_M_Vo -0.815***
(0.297)
-0.200
(0.464)
-0.571**
(0.253)
-0.305
(0.380)
-0.417**
(0.184)
-0.304
(0.308)
0.024
(0.111)
-0.233
(0.195)
0.029
(0.104)
0.066
(0.201)
Five_Emp_Pop_Ratio 2.073***
(0.348)
2.095***
(0.422)
1.375***
(0.213)
1.489***
(0.298)
0.654***
(0.151)
0.803***
(0.172)
0.360***
(0.090)
0.420***
(0.128)
0.105
(0.104)
0.149
(0.233)
Five_PUMA_Self_Emp_Pop_Ratio -0.290
(0.433)
0.410
(0.438)
0.868***
(0.308)
1.491***
(0.339)
0.236
(0.195)
0.523**
(0.219)
0.223
(0.165)
0.222
(0.230)
0.112
(0.116)
0.232
(0.200)
Five_PUMA_Mttw 1.038*
(0.571)
-0.371
(0.525)
-0.511*
(0.285)
-1.406***
(0.355)
-0.421**
(0.193)
-0.907***
(0.255)
-0.448***
(0.125)
-0.630***
(0.177)
-0.243**
(0.106)
-0.194
(0.155)
January_Average_Temperature 0.049
(0.231)
24.351***
(1.514)
-0.190
(0.148)
-8.832***
(1.168)
-0.261***
(0.099)
-4.001***
(0.747)
-0.045
(0.057)
8.616***
(0.614)
-0.054
(0.065)
0.026
(0.179)
Five_Violent_Crime 0.006
(0.015)
-1.076***
(0.144)
0.001
(0.006)
1.111***
(0.105)
0.003
(0.005)
0.028
(0.035)
-0.005
(0.003)
0.427***
(0.041)
0.002
(0.003)
0.046**
(0.021)
Five_Property_Crime 0.003
(0.003)
0.075***
(0.004)
1.952×10
-4

(0.001)
-0.020***
(0.003)
9.110×10
-5

(0.001)
0.028***
(0.002)
4.579×10
-4

(0.001)
-0.100***
(0.008)
-0.001
(0.001)
-0.004
(0.005)
New_England 4.991
(7.608)
-4.906
(7.122)
-6.406
(4.901)
-2.369
(1.831)
-4.250**
(1.889)

Middle_Atlantic -10.127
(10.348)
-9.585
(6.056)
-7.619*
(4.296)
-2.139
(2.123)
-1.219
(2.124)

East_North_Central -4.885
(7.617)
-7.359
(4.667)
-6.744**
(3.321)
-2.423
(2.275)
-1.310
(2.252)

West_North_Central -4.283
(8.964)
-13.153**
(5.394)
-11.171***
(3.662)
-3.349
(3.022)
-0.969
(3.954)

South_Atlantic 31.752***
(7.452)
6.247*
(3.735)
-0.042
(2.322)
0.608
(1.574)
0.717
(1.168)

East_South_Central 20.284**
(9.753)
4.034
(4.527)
-0.191
(3.048)
-1.063
(2.355)
-2.162
(1.805)

West_South_Central 28.625***
(8.939)
4.521
(3.851)
-0.625
(2.262)
0.121
(1.853)
0.251
(1.534)

Mountain 2.744
(9.984)
-8.081
(5.392)
-9.625***
(2.853)
-4.547**
(2.292)
1.755
(2.389)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 276 276 260 260 226 226 183 183 94 94
N 1,330 1,330 1,165 1,165 886 886 669 669 220 220
R
2
0.2021 0.4121 0.2171 0.3876 0.1978 0.4098 0.1186 0.3110 0.1352 0.4689
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

122

Appendix E.3. Regression Results for the Migration Models – Group C (Age 45-54) (2006)
Dependent variable
Independent variable Six_ALL_Inmig_Pc_C Six_BAPLUS_Inmig_Pc_C Six_MAPLUS_Inmig_Pc_C Six_SCC_Inmig_Pc_C Six_BOH_Inmig_Pc_C
Constant -132.533
(83.500)
-7578.939***
(373.252)
-117.913***
(37.816)
-3413.866***
(351.936)
-31.994
(24.052)
2085.018**
(937.926)
-70.785***
(26.482)
24.365
(267.212)
4.428
(21.195)
697.203**
(318.442)
Five_PUMA_Pop_Den -3.830×10
-5

(2.683×10
-4
)
2.359×10
-4

(2.050×10
-4
)
5.209×10
-4
***
(9.530×10
-5
)
0.001***
(6.700×10
-5
)
4.733×10
-4
***
(7.250×10
-5
)
4.432×10
-4
***
(7.810×10
-5
)
8.430×10
-5

(9.460×10
-5
)
1.350×10
-5

(1.199×10
-4
)
2.316×10
-4
***
(5.120×10
-5
)
1.853×10
-4
***
(5.530×10
-5
)
Five_MSA_Pop_BEA_M -0.245
(0.333)
-144.263***
(9.090)
3.316**
(1.542)
-33.539***
(5.880)
3.396***
(1.261)
34.774***
(12.769)
2.151**
(0.859)
0.093
(2.173)
0.237
(0.902)
9.275**
(4.337)
Five_MSA_Pop_BEA_L -14.976***
(4.027)
-306.019***
(16.762)
1.141
(2.347)
-89.845***
(13.737)
4.046**
(1.914)
96.209***
(35.969)
1.252
(1.250)
-0.790
(9.288)
0.852
(0.860)
36.519**
(15.652)
Five_MSA_Inc_BEA_Ln 7.669
(8.110)
694.657***
(33.568)
10.436***
(3.761)
326.057***
(34.708)
4.816**
(2.393)
-192.645**
(86.210)
7.348***
(2.510)
-2.095
(26.499)
0.175
(1.866)
-68.137**
(31.501)
Five_Pct_M_Vo -0.304
(0.198)
0.122
(0.314)
-0.284**
(0.111)
-0.261
(0.204)
-0.234***
(0.086)
-0.360**
(0.171)
-0.060
(0.073)
-0.282*
(0.157)
-0.076
(0.111)
-0.117
(0.278)
Five_Emp_Pop_Ratio 0.975***
(0.219)
0.887***
(0.287)
0.513***
(0.155)
0.465**
(0.224)
0.090
(0.103)
0.030
(0.159)
0.159*
(0.082)
0.115
(0.146)
0.102
(0.142)
0.176
(0.229)
Five_PUMA_Self_Emp_Pop_Ratio 0.398
(0.344)
0.639
(0.452)
0.679***
(0.198)
1.048***
(0.281)
0.503***
(0.129)
0.692***
(0.157)
0.235**
(0.113)
0.423**
(0.183)
0.006
(0.117)
0.007
(0.223)
Five_PUMA_Mttw 0.586
(0.411)
-0.136
(0.433)
-0.273*
(0.160)
-0.692***
(0.165)
-0.334***
(0.128)
-0.529***
(0.188)
-0.266***
(0.099)
-0.324*
(0.194)
-0.252**
(0.100)
-0.323
(0.294)
January_Average_Temperature 0.054
(0.156)
8.734***
(0.915)
-0.098
(0.082)
4.421***
(0.359)
-0.148**
(0.062)
-4.102**
(2.038)
-0.071*
(0.038)
0.495**
(0.241)
0.049
(0.038)
0.078
(0.197)
Five_Violent_Crime 0.005
(0.011)
-0.073
(0.099)
-0.005
(0.003)
-0.411***
(0.034)
-0.006**
(0.003)
0.048***
(0.013)
0.001
(0.002)
0.030*
(0.017)
-0.004
(0.003)
-0.012
(0.014)
Five_Property_Crime 0.004**
(0.002)
0.035***
(0.002)
-2.523×10
-4

(0.001)
0.013***
(0.002)
-1.870×10
-4

(0.001)
0.017*
(0.010)
-2.803×10
-4

(0.001)
-0.007***
(0.001)
-2.099×10
-4

(0.001)
0.003*
(0.002)
New_England 4.417
(5.319)
-4.532
(3.007)
0.320
(2.769)
-3.100
(1.968)
-0.845
(1.422)

Middle_Atlantic 0.139
(4.960)
-3.310
(2.850)
-2.982
(1.975)
-1.927
(1.507)
1.810
(1.351)

East_North_Central -2.599
(4.704)
-2.037
(2.524)
-0.154
(2.021)
-2.271
(1.470)
1.653
(1.391)

West_North_Central 7.503
(5.989)
-1.587
(3.761)
-4.151*
(2.447)
-3.297*
(1.844)
4.109**
(1.899)

South_Atlantic 19.556***
(4.573)
6.031***
(1.751)
2.922**
(1.466)
-0.452
(0.991)
0.226
(0.861)

East_South_Central 24.082***
(6.962)
5.532**
(2.480)
1.691
(1.775)
2.268
(1.609)
15.937***
(1.705)

West_South_Central 18.417***
(5.228)
5.726**
(2.391)
2.377
(1.484)
1.372
(1.235)
0.943
(1.484)

Mountain 4.034
(6.205)
2.147
(2.866)
0.555
(1.575)
-0.303
(1.694)
0.200
(1.324)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 276 276 247 247 208 208 170 170 65 65
N 1,313 1,313 1,057 1,057 737 737 512 512 140 140
R
2
0.1687 0.4042 0.1175 0.3107 0.1265 0.3459 0.0940 0.3237 0.3198 0.4958
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

123

Appendix E.4. Regression Results for the Migration Models – Group A (Age 25-34) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_A Nine_BAPLUS_Inmig_Pc_A Nine_MAPLUS_Inmig_Pc_A Nine_SCC_Inmig_Pc_A Nine_BOH_Inmig_Pc_A
Constant -125.859
(236.022)
-17049.010***
(784.085)
-455.264**
(209.575)
-172168.900***
(49007.750)
-291.467***
(97.120)
-401.471
(1277.642)
-266.248***
(85.054)
14752.620***
(4378.913)
-20.332
(46.149)
27.970
(145.634)
Eight_PUMA_Pop_Den 0.005***
(0.001)
0.006***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.002***
(4.442×10
-4
)
0.002***
(0.001)
0.001***
(3.350×10
-4
)
0.001***
(4.293×10
-4
)
4.682×10
-4
***
(1.202×10
-4
)
3.685×10
-4
***
(1.232×10
-4
)
Eight_MSA_Pop_BEA_M 21.646**
(9.156)
-337.924***
(24.300)
22.769***
(6.061)
-3180.494***
(906.543)
9.494***
(2.934)
-10.021
(23.647)
6.568***
(2.493)
222.500***
(67.221)
4.539**
(1.777)
-6.583***
(2.104)
Eight_MSA_Pop_BEA_L 15.115
(14.684)
-413.268***
(19.173)
26.057**
(10.522)
-3270.473***
(936.778)
10.094**
(4.378)
22.260
(18.058)
7.696**
(3.388)
241.143***
(64.234)
7.094**
(2.893)
23.594***
(2.837)
Eight_MSA_Inc_BEA_Ln 33.509
(22.198)
1586.041***
(78.070)
55.960***
(19.823)
15886.650***
(4522.720)
34.648***
(9.885)
47.872
(121.897)
30.748***
(7.897)
-1356.748***
(404.107)
5.652
(4.471)
4.566
(13.240)
Eight_Pct_M_Vo -4.659***
(0.717)
-4.622***
(1.245)
-3.217***
(0.619)
-4.184***
(0.983)
-1.461***
(0.389)
-1.653***
(0.454)
-0.946***
(0.250)
-1.347***
(0.321)
-0.166
(0.130)
-0.753***
(0.246)
Eight_Emp_Pop_Ratio 1.866***
(0.554)
2.218***
(0.516)
1.234***
(0.378)
1.349***
(0.433)
0.381
(0.242)
0.414
(0.294)
0.293**
(0.123)
0.228*
(0.138)
-0.207
(0.143)
-0.206
(0.279)
Eight_PUMA_Self_Emp_Pop_Ratio -2.282**
(1.025)
-1.284
(1.100)
0.748
(0.703)
1.490*
(0.782)
0.028
(0.243)
0.422
(0.350)
-0.083
(0.230)
0.229
(0.264)
0.425**
(0.203)
0.213
(0.310)
Eight_PUMA_Mttw -3.326***
(0.836)
-5.234***
(0.959)
-3.662***
(0.731)
-5.475***
(1.007)
-1.691***
(0.380)
-2.612***
(0.621)
-1.380***
(0.275)
-2.016***
(0.341)
-0.557***
(0.159)
-0.918***
(0.198)
January_Average_Temperature -1.770***
(0.444)
13.180***
(0.524)
-1.126***
(0.327)
210.462***
(60.478)
-0.512***
(0.171)
-0.153
(1.625)
-0.241**
(0.122)
-20.828***
(6.133)
-0.159
(0.112)
-0.342
(0.283)
Eight_Violent_Crime -0.020
(0.025)
0.740***
(0.058)
-0.022*
(0.013)
-8.228***
(2.433)
-0.007
(0.007)
-0.006
(0.096)
-0.011**
(0.006)
0.734***
(0.197)
-0.004
(0.007)
-0.025**
(0.011)
Eight_Property_Crime -0.001
(0.005)
-0.004
(0.004)
-0.001
(0.004)
0.675***
(0.196)
-0.001
(0.002)
-0.009**
(0.004)
-0.001
(0.001)
-0.012***
(0..001)
0.001
(0.001)
-0.001
(0.001)
New_England 8.048
(30.844)
7.088
(26.293)
-0.500
(11.189)
4.245
(9.633)
3.146
(2.825)

Middle_Atlantic -73.773***
(13.776)
-35.593***
(11.267)
-12.892**
(6.371)
-11.753***
(4.096)
-2.579
(2.846)

East_North_Central -45.462***
(14.704)
-18.932*
(10.766)
-6.614
(5.863)
-6.529
(4.668)
-5.725*
(3.027)

West_North_Central -43.961**
(22.188)
-37.359**
(16.328)
-16.851**
(6.507)
-14.378***
(4.786)
-5.340*
(2.900)

South_Atlantic 16.845
(10.285)
4.528
(7.800)
2.135
(3.876)
-1.331
(3.401)
-2.366
(2.036)

East_South_Central 18.822
(19.264)
-0.036
(9.877)
-2.161
(4.444)
-5.836
(3.536)
-4.160
(2.687)

West_South_Central 11.243
(13.651)
-8.162
(8.400)
-2.478
(3.942)
-7.203**
(3.097)
-5.616**
(2.789)

Mountain -31.184**
(15.406)
-34.869***
(10.924)
-14.387***
(5.424)
-10.844***
(3.873)
-4.259*
(2.548)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 241 241 209 209 196 196 110 110
N 1,308 1,308 1,244 1,244 973 973 892 892 335 335
R
2
0.2257 0.4232 0.2469 0.3903 0.2145 0.3482 0.2077 0.3590 0.1828 0.3431
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

124

Appendix E.5. Regression Results for the Migration Models – Group B (Age 35-44) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_B Nine_BAPLUS_Inmig_Pc_B Nine_MAPLUS_Inmig_Pc_B Nine_SCC_Inmig_Pc_B Nine_BOH_Inmig_Pc_B
Constant -151.879
(182.994)
-3555.729***
(412.108)
-302.180***
(88.212)
1658.294***
(549.657)
-184.397***
(60.806)
11696.420***
(1915.870)
-122.104**
(55.408)
201.903**
(78.190)
-45.584
(32.145)
253.819
(205.199)
Eight_PUMA_Pop_Den -4.820×10
-5

(2.537×10
-4
)
4.031×10
-4
**
(1.786×10
-4
)
3.565×10
-4
***
(1.168×10
-4
)
0.001***
(1.520×10
-4
)
4.028×10
-4
***
(1.137×10
-4
)
4.622×10
-4
***
(1.122×10
-4
)
2.437×10
-4
***
(5.510×10
-5
)
2.559×10
-4
***
(6.360×10
-5
)
2.377×10
-4
***
(6.850×10
-5
)
2.329×10
-4
***
(4.350×10
-5
)
Eight_MSA_Pop_BEA_M 4.084
(4.712)
-40.045***
(5.279)
5.654**
(2.627)
39.476***
(10.186)
2.341
(1.943)
206.938***
(32.854)
2.056
(1.535)
7.743***
(1.583)
-0.151
(1.671)
11.593
(7.092)
Eight_MSA_Pop_BEA_L -0.298
(10.010)
-58.155***
(6.146)
5.226
(4.837)
55.732***
(14.018)
1.970
(3.404)
198.310***
(31.439)
2.464
(2.569)
34.418***
(1.663)
-1.363
(1.862)
12.416*
(6.378)
Eight_MSA_Inc_BEA_Ln 16.424
(17.285)
319.074***
(38.777)
30.511***
(8.562)
-152.090***
(46.877)
21.021***
(5.952)
-1087.636***
(178.255)
13.357**
(5.267)
-8.986
(8.263)
4.751
(3.017)
-22.839
(19.754)
Eight_Pct_M_Vo -1.213***
(0.343)
-0.245
(0.512)
-0.997***
(0.216)
-0.730**
(0.329)
-0.765***
(0.170)
-0.774***
(0.247)
-0.461***
(0.106)
-0.452***
(0.147)
-0.014
(0.145)
-0.036
(0.281)
Eight_Emp_Pop_Ratio 0.997***
(0.280)
1.074***
(0.319)
0.578***
(0.205)
0.663**
(0.259)
0.082
(0.143)
0.043
(0.185)
0.092
(0.124)
0.028
(0.147)
-0.080
(0.138)
0.084
(0.227)
Eight_PUMA_Self_Emp_Pop_Ratio 0.260
(0.473)
0.875*
(0.502)
1.100***
(0.350)
1.712***
(0.467)
0.457**
(0.194)
0.723***
(0.263)
0.255
(0.173)
0.319
(0.211)
0.274**
(0.116)
0.330
(0.203)
Eight_PUMA_Mttw 0.126
(0.352)
-0.795**
(0.384)
-0.737**
(0.295)
-1.483***
(0.402)
-0.590**
(0.241)
-1.111***
(0.315)
-0.386***
(0.131)
-0.774***
(0.183)
0.048
(0.120)
0.086
(0.293)
January_Average_Temperature -0.603**
(0.270)
3.333***
(0.671)
-0.332**
(0.152)
-2.917***
(0.940)
-0.156
(0.097)
-15.112***
(2.413)
-0.095*
(0.057)
-0.710***
(0.093)
-0.008
(0.068)
-0.245
(0.293)
Eight_Violent_Crime -0.004
(0.016)
0.235***
(0.053)
-0.013**
(0.006)
0.253***
(0.041)
-0.014***
(0.004)
0.987***
(0.159)
-0.007**
(0.003)
-0.145***
(0.009)
-9.810×10
-5

(0.005)
0.045
(0.030)
Eight_Property_Crime 0.003
(0.003)
0.011***
(0.003)
0.001
(0.002)
-0.019***
(0.005)
2.394×10
-4

(0.001)
-0.047***
(0.007)
0.001
(0.001)
3.803×10
-4

(0.001)
0.002*
(0.001)
-0.010*
(0.006)
New_England 0.336
(9.766)
0.406
(8.133)
3.388
(6.185)
0.126
(4.397)
1.841
(2.626)

Middle_Atlantic -28.499***
(8.381)
-10.773**
(4.784)
-2.496
(3.572)
-1.334
(1.977)
1.263
(1.677)

East_North_Central -23.485***
(8.478)
-7.660
(4.790)
1.166
(3.730)
2.173
(3.175)
8.216*
(4.555)

West_North_Central -14.264*
(8.480)
-13.772**
(6.624)
-4.970
(4.324)
0.724
(4.258)
7.394**
(3.195)

South_Atlantic 13.386*
(7.010)
5.341*
(3.180)
4.492*
(2.389)
0.747
(1.766)
-0.196
(1.675)

East_South_Central 10.051
(10.872)
9.911**
(4.707)
6.995**
(3.235)
2.014
(2.501)
3.040
(3.486)

West_South_Central 6.743
(8.048)
2.050
(3.531)
1.500
(2.437)
-1.462
(1.988)
0.967
(1.966)

Mountain -12.896*
(7.292)
-10.853***
(3.926)
-6.995***
(2.583)
-4.231**
(1.772)
-2.233
(2.161)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 234 234 198 198 165 165 84 84
N 1,291 1,291 1,093 1,093 811 811 644 644 195 195
R
2
0.1619 0.3973 0.1955 0.3683 0.1757 0.3790 0.1241 0.3652 0.1520 0.4980
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

125

Appendix E.6. Regression Results for the Migration Models – Group C (Age 45-54) (2009)
Dependent variable
Independent variable Nine_ALL_Inmig_Pc_C Nine_BAPLUS_Inmig_Pc_C Nine_MAPLUS_Inmig_Pc_C Nine_SCC_Inmig_Pc_C Nine_BOH_Inmig_Pc_C
Constant 18.994
(115.746)
-94681.580***
(11832.830)
-97.198**
(40.062)
58.139
(40.795)
-77.553*
(41.540)
-35588.370***
(2854.279)
-15.496
(25.481)
1025.779**
(440.925)
53.041
(37.293)
9.108
(247.563)
Eight_PUMA_Pop_Den 1.449×10
-4

(2.293×10
-4
)
4.132×10
-4
***
(1.231×10
-4
)
3.008×10
-4
***
(5.950×10
-5
)
3.907×10
-4
***
(4.910×10
-5
)
3.776×10
-4
***
(5.960×10
-5
)
4.171×10
-4
***
(5.230×10
-5
)
9.540×10
-5
**
(4.230×10
-5
)
8.370×10
-5
**
(4.150×10
-5
)
1.117×10
-4
**
(5.330×10
-5
)
6.780×10
-5

(1.059×10
-4
)
Eight_MSA_Pop_BEA_M 5.450*
(3.123)
-1714.116***
(218.945)
2.141*
(1.155)
13.467***
(1.415)
0.172
(0.937)
-550.749***
(44.070)
1.904**
(0.950)
11.047***
(2.860)
-0.693
(1.066)
16.367***
(0.789)
Eight_MSA_Pop_BEA_L 1.293
(4.153)
-1786.418***
(226.432)
3.776***
(1.440)
14.765***
(1.287)
0.090
(1.381)
-521.148***
(41.212)
1.704
(1.043)
16.751***
(3.748)
0.490
(1.316)
-6.662
(13.073)
Eight_MSA_Inc_BEA_Ln 0.233
(10.928)
8734.278***
(1091.777)
11.019***
(4.050)
-0.640
(4.153)
9.883**
(4.020)
3298.302***
(264.433)
3.282
(2.550)
-93.891**
(41.223)
-3.806
(3.756)
-2.003
(24.699)
Eight_Pct_M_Vo -0.420**
(0.207)
-0.002
(0.345)
-0.321***
(0.114)
-0.160
(0.195)
-0.144
(0.136)
-0.280
(0.184)
0.020
(0.132)
-0.166
(0.298)
0.150*
(0.089)
0.009
(0.289)
Eight_Emp_Pop_Ratio 0.258
(0.187)
0.277
(0.232)
0.133
(0.097)
0.120
(0.113)
0.035
(0.076)
0.020
(0.103)
-0.088
(0.100)
-0.138
(0.167)
-0.126*
(0.071)
-0.246
(0.395)
Eight_PUMA_Self_Emp_Pop_Ratio 0.577*
(0.315)
0.726**
(0.351)
0.817***
(0.170)
1.067***
(0.241)
0.209*
(0.114)
0.246
(0.177)
0.244*
(0.135)
0.118
(0.229)
0.176
(0.116)
0.087
(0.358)
Eight_PUMA_Mttw -0.018
(0.215)
-0.270
(0.267)
-0.501***
(0.115)
-0.580***
(0.208)
-0.432***
(0.086)
-0.576***
(0.124)
-0.131
(0.112)
-0.213
(0.202)
0.038
(0.154)
0.056
(0.348)
January_Average_Temperature -0.269*
(0.140)
116.694***
(14.602)
-0.180***
(0.056)
-0.687***
(0.104)
-0.213***
(0.050)
48.529***
(3.948)
-0.114***
(0.038)
-0.439***
(0.151)
-0.068
(0.046)
-4.107***
(0.823)
Eight_Violent_Crime 0.018*
(0.011)
-4.690***
(0.586)
-0.001
(0.003)
0.008
(0.021)
3.527×10
-4

(0.003)
-1.744***
(0.133)
-0.001
(0.002)
0.029
(0.023)
0.001
(0.002)
0.348***
(0.073)
Eight_Property_Crime 2.427×10
-4

(0.002)
0.375***
(0.047)
2.936×10
-4

(0.001)
-0.006***
(0.001)
2.762×10
-4

(0.001)
0.014***
(0.001)
9.090×10
-5

(4.341×10
-4
)
-0.004
(0.003)
3.785×10
-4

(6.795×10
-4
)
0.008***
(0.002)
New_England 10.562**
(4.178)
3.911
(2.618)
-3.797
(2.592)
-1.844
(1.247)
1.820
(3.473)

Middle_Atlantic -10.312**
(4.679)
-3.297
(2.145)
-3.364*
(1.961)
-1.363
(1.308)
0.768
(1.343)

East_North_Central 0.014
(5.060)
-1.587
(2.314)
-4.096**
(1.956)
0.593
(1.565)
-0.405
(2.423)

West_North_Central 5.789
(5.557)
1.710
(2.278)
-4.343*
(2.337)
3.082
(3.423)
-2.020
(2.260)

South_Atlantic 17.424***
(4.604)
5.046***
(1.541)
1.273
(1.294)
1.007
(0.959)
-0.578
(1.318)

East_South_Central 15.819*
(8.405)
1.361
(2.627)
-3.281
(2.090)
-2.489*
(1.397)
-3.210*
(1.899)

West_South_Central 11.535**
(4.758)
1.250
(1.686)
-1.608
(1.336)
0.152
(1.090)
-0.164
(1.553)

Mountain 4.568
(5.010)
0.793
(1.825)
-3.333*
(1.743)
-0.033
(1.454)
1.440
(2.037)

Census Division Fixed Effect Yes No Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes No Yes
Number of MSA Clusters 248 248 219 219 172 172 143 143 67 67
N 1,264 1,264 967 967 643 643 442 442 109 109
R
2
0.1419 0.3466 0.1504 0.3286 0.1215 0.3535 0.0584 0.2953 0.2010 0.7332
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

126

Appendix F. Regression Results for the Migration Models – PUMA “Talented” People Population Density (w/o amenity variables)
Appendix F.1. Regression Results for the Migration Models – Group A (Age 25-34) (2006)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc_A Six_MAPLUS_Inmig_Pc_A Six_SCC_Inmig_Pc_A Six_BOH_Inmig_Pc_A
Constant -711.315***
(148.895)
-1210.988***
(139.433)
-400.345***
(102.902)
-389.970***
(112.527)
-225.256***
(65.674)
-492.593***
(96.692)
39.864
(33.497)
-78.904**
(34.232)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den /
Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.016**
(0.007)
0.015**
(0.007)
0.016**
(0.007)
0.014**
(0.007)
0.020***
(0.007)
0.018**
(0.008)
0.009***
(0.001)
0.008***
(0.001)
Five_MSA_Inc_BEA_Ln 63.292***
(15.851)
118.355***
(17.054)
39.937***
(10.258)
42.075***
(13.112)
21.772***
(6.699)
48.914***
(10.901)
-3.223
(3.231)
8.860**
(4.122)
Five_Pct_M_Vo -2.596***
(0.490)
-3.976***
(1.050)
-1.109***
(0.288)
-1.811***
(0.568)
-0.611**
(0.239)
-0.996**
(0.486)
-0.251**
(0.101)
-0.525**
(0.228)
Five_Emp_Pop_Ratio 2.679***
(0.355)
2.499***
(0.449)
0.716***
(0.186)
0.709***
(0.239)
0.737***
(0.176)
0.717***
(0.232)
0.111
(0.068)
0.068
(0.127)
Five_PUMA_Self_Emp_Pop_Ratio -0.347
(0.457)
1.454**
(0.587)
-0.297
(0.238)
0.293
(0.397)
-0.572**
(0.246)
-0.096
(0.422)
-0.100
(0.181)
-0.228
(0.331)
Five_PUMA_Mttw -1.858***
(0.555)
-3.494***
(0.953)
-1.119***
(0.288)
-2.061***
(0.685)
-0.676***
(0.209)
-1.220**
(0.480)
0.078
(0.083)
-0.031
(0.157)
New_England 7.684
(15.710)
8.105
(10.227)
-0.872
(5.474)
1.059
(1.827)

Middle_Atlantic -16.319**
(7.285)
-5.351**
(2.676)
-6.045**
(3.028)
0.367
(1.805)

East_North_Central 8.399
(11.149)
3.603
(4.419)
-2.364
(3.434)
-1.741
(1.829)

West_North_Central -14.899*
(8.262)
-7.320*
(3.846)
-8.487**
(3.941)
-2.989
(1.953)

South_Atlantic 2.613
(8.594)
0.216
(4.373)
-3.657
(3.348)
-2.423
(1.704)

East_South_Central 3.598
(7.060)
-1.404
(3.030)
-8.140**
(3.340)
-3.081*
(1.795)

West_South_Central -0.730
(6.954)
-1.838
(2.736)
-5.321*
(2.962)
-2.258
(1.650)

Mountain -14.633*
(7.541)
-8.240***
(2.987)
-8.795**
(3.806)
-2.331
(2.267)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 392 392 325 325 284 284 164 164
N 1,557 1,557 1,226 1,226 1,120 1,120 465 465
R
2
0.2795 0.4392 0.2413 0.4033 0.1729 0.3372 0.1010 0.2788
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

127

Appendix F.2. Regression Results for the Migration Models – Group B (Age 35-44) (2006)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc_B Six_MAPLUS_Inmig_Pc_B Six_SCC_Inmig_Pc_B Six_BOH_Inmig_Pc_B
Constant -445.537***
(86.613)
-385.550***
(69.028)
-270.577***
(52.273)
18.592
(38.448)
-164.749***
(30.963)
-261.094***
(23.720)
-67.135**
(29.003)
38.738
(31.115)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den /
Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.004**
(0.002)
0.004**
(0.002)
0.005***
(0.002)
0.005**
(0.002)
0.005**
(0.002)
0.004*
(0.002)
0.005***
(0.001)
0.005***
(0.001)
Five_MSA_Inc_BEA_Ln 36.847***
(9.055)
34.640***
(8.530)
24.760***
(4.982)
-1.452
(4.691)
16.270***
(3.191)
26.885***
(3.247)
7.824***
(2.583)
-2.423
(1.505)
Five_Pct_M_Vo -0.718***
(0.254)
-0.852
(0.576)
-0.413***
(0.154)
-0.542
(0.362)
-0.007
(0.117)
-0.329
(0.234)
-0.018
(0.090)
-0.088
(0.254)
Five_Emp_Pop_Ratio 1.386***
(0.204)
1.395***
(0.280)
0.568***
(0.132)
0.572***
(0.157)
0.225**
(0.091)
0.218*
(0.120)
-0.068
(0.107)
-0.056
(0.238)
Five_PUMA_Self_Emp_Pop_Ratio 0.388
(0.248)
1.127***
(0.372)
-0.051
(0.187)
0.385*
(0.208)
0.055
(0.139)
0.114
(0.232)
0.038
(0.109)
0.179
(0.205)
Five_PUMA_Mttw -0.075
(0.286)
-0.892**
(0.429)
-0.257**
(0.127)
-0.616**
(0.238)
-0.293**
(0.115)
-0.438**
(0.186)
-0.082
(0.087)
-0.081
(0.135)
New_England -1.136
(4.656)
-1.063
(3.382)
-0.919
(1.616)
-1.563
(1.311)

Middle_Atlantic -9.305***
(3.031)
-3.559*
(2.125)
-2.859*
(1.701)
0.630
(1.829)

East_North_Central 3.124
(5.169)
1.160
(2.685)
0.242
(2.116)
0.589
(1.642)

West_North_Central -8.495**
(3.335)
-3.397
(2.227)
-2.414
(2.594)
1.447
(2.031)

South_Atlantic 5.326
(3.490)
0.492
(2.598)
-0.516
(1.687)
0.198
(1.324)

East_South_Central 6.762*
(3.948)
3.637
(2.632)
-1.311
(2.035)
-1.683
(1.496)

West_South_Central 5.295
(3.343)
0.507
(1.759)
0.047
(1.722)
0.627
(1.625)

Mountain -3.910
(4.762)
-5.635**
(2.484)
-2.544
(2.068)
2.142
(2.397)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 362 362 295 295 235 235 116 116
N 1,412 1,412 1,056 1,056 797 797 253 253
R
2
0.2284 0.3940 0.1890 0.4045 0.0946 0.3217 0.0866 0.4932
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

128

Appendix F.3. Regression Results for the Migration Models – Group C (Age 45-54) (2006)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc_C Six_MAPLUS_Inmig_Pc_C Six_SCC_Inmig_Pc_C Six_BOH_Inmig_Pc_C
Constant -108.789***
(32.073)
-437.829***
(48.781)
-77.135***
(25.343)
-219.346***
(40.363)
-83.666***
(20.001)
-108.707***
(40.873)
-14.629
(18.429)
173.034**
(74.287)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den /
Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.002***
(0.001)
0.002***
(0.001)
0.003***
(0.001)
0.003***
(0.001)
0.002*
(0.001)
0.001
(0.001)
0.004***
(0.001)
0.003*
(0.002)
Five_MSA_Inc_BEA_Ln 9.025***
(3.262)
43.325***
(5.752)
7.940***
(2.543)
22.860***
(4.756)
8.066***
(2.059)
11.299**
(4.761)
1.883
(1.783)
-15.210**
(6.958)
Five_Pct_M_Vo -0.322**
(0.125)
-0.447
(0.295)
-0.196**
(0.084)
-0.424**
(0.195)
-0.083
(0.074)
-0.362**
(0.160)
-0.128
(0.095)
-0.285
(0.307)
Five_Emp_Pop_Ratio 0.490***
(0.112)
0.299*
(0.176)
0.122
(0.079)
0.038
(0.128)
0.181***
(0.066)
0.126
(0.122)
0.069
(0.072)
0.018
(0.191)
Five_PUMA_Self_Emp_Pop_Ratio 0.432***
(0.158)
0.905***
(0.263)
0.280**
(0.121)
0.537***
(0.163)
0.118
(0.102)
0.278
(0.182)
0.107
(0.115)
0.011
(0.278)
Five_PUMA_Mttw -0.110
(0.123)
-0.464**
(0.179)
-0.190**
(0.089)
-0.422**
(0.169)
-0.167**
(0.077)
-0.304**
(0.151)
-0.153**
(0.067)
-0.286*
(0.148)
New_England 1.224
(2.310)
3.597**
(1.790)
-0.269
(1.550)
-0.162
(1.037)

Middle_Atlantic -2.793*
(1.485)
-0.096
(0.911)
-1.416
(0.968)
1.138
(0.858)

East_North_Central 1.115
(2.231)
2.776*
(1.502)
0.763
(1.550)
4.475**
(1.883)

West_North_Central 0.249
(3.284)
0.815
(2.083)
-0.631
(1.498)
4.465***
(1.182)

South_Atlantic 4.339**
(1.911)
1.813
(1.489)
-0.445
(0.914)
0.201
(0.845)

East_South_Central 5.807***
(2.114)
2.781*
(1.530)
4.027***
(1.424)
9.526***
(3.224)

West_South_Central 4.644**
(2.056)
2.474**
(1.019)
1.530
(1.084)
1.197
(1.222)

Mountain 2.484
(2.308)
1.487
(1.441)
-0.022
(1.452)
-0.206
(1.123)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 349 349 284 284 222 222 87 87
N 1,284 1,284 887 887 615 615 172 172
R
2
0.1220 0.3555 0.1101 0.3951 0.0779 0.3518 0.2122 0.4685
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

129

Appendix F.4. Regression Results for the Migration Models – Group A (Age 25-34) (2009)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc_A Nine_MAPLUS_Inmig_Pc_A Nine_SCC_Inmig_Pc_A Nine_BOH_Inmig_Pc_A
Constant -691.662***
(181.264)
-2368.239***
(206.812)
-382.723***
(92.989)
-446.467***
(115.062)
-310.024***
(69.249)
-140.502***
(46.061)
-66.369*
(34.571)
-571.327***
(37.249)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
0.019*
(0.010)
0.017
(0.011)
0.016*
(0.009)
0.015
(0.009)
0.020*
(0.011)
0.017
(0.012)
0.014***
(0.003)
0.012***
(0.003)
Eight_MSA_Inc_BEA_Ln 68.703***
(19.659)
237.340***
(21.947)
39.773***
(9.616)
50.148***
(13.508)
32.370***
(7.176)
19.847***
(5.812)
8.060**
(3.778)
58.051***
(4.076)
Eight_Pct_M_Vo -3.219***
(0.751)
-5.468***
(1.693)
-1.468***
(0.441)
-2.380**
(0.938)
-1.001***
(0.343)
-1.995***
(0.713)
-0.330*
(0.172)
-1.186**
(0.461)
Eight_Emp_Pop_Ratio 1.559***
(0.475)
1.503**
(0.743)
0.381*
(0.207)
0.350
(0.342)
0.310**
(0.145)
0.268
(0.250)
-0.063
(0.126)
0.052
(0.293)
Eight_PUMA_Self_Emp_Pop_Ratio -1.148
(0.734)
0.068
(1.007)
-0.732**
(0.321)
-0.232
(0.563)
-0.671**
(0.285)
-0.389
(0.488)
0.032
(0.226)
-0.308
(0.389)
Eight_PUMA_Mttw -1.909***
(0.633)
-4.172***
(0.977)
-0.876***
(0.299)
-2.034***
(0.531)
-0.687**
(0.265)
-1.493***
(0.382)
-0.037
(0.158)
-0.399
(0.280)
New_England 20.786
(18.153)
7.222
(7.323)
6.508
(6.661)
2.466
(2.135)

Middle_Atlantic -20.968***
(7.589)
-6.101
(3.757)
-9.672***
(3.076)
-2.301
(1.773)

East_North_Central 21.653
(14.048)
13.894**
(6.544)
7.286
(5.225)
3.564
(2.510)

West_North_Central -6.881
(8.315)
-4.016
(4.487)
-5.447*
(3.100)
1.332
(2.443)

South_Atlantic 3.341
(8.114)
1.597
(3.978)
-2.562
(2.893)
-1.577
(2.321)

East_South_Central 9.926*
(5.952)
4.538
(2.809)
-3.424
(2.359)
1.111
(2.352)

West_South_Central -4.400
(5.962)
-0.560
(3.159)
-6.134***
(2.115)
-2.129
(1.716)

Mountain -18.360***
(6.773)
-6.720**
(3.006)
-7.324***
(2.727)
-2.053
(2.419)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 384 384 317 317 285 285 151 151
N 1,614 1,614 1,251 1,251 1,129 1,129 424 424
R
2
0.2349 0.3677 0.2124 0.3435 0.2053 0.3530 0.1622 0.3860
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

130

Appendix F.5. Regression Results for the Migration Models – Group B (Age 35-44) (2009)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc_B Nine_MAPLUS_Inmig_Pc_B Nine_SCC_Inmig_Pc_B Nine_BOH_Inmig_Pc_B
Constant -348.300***
(65.567)
-826.019***
(72.076)
-209.210***
(47.617)
-527.988***
(53.954)
-122.927***
(32.494)
-123.553***
(23.694)
-39.008
(29.594)
99.873***
(28.547)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
0.003**
(0.001)
0.003*
(0.002)
0.004**
(0.002)
0.003**
(0.002)
0.004***
(0.001)
0.004***
(0.001)
0.002***
(0.001)
0.002*
(0.001)
Eight_MSA_Inc_BEA_Ln 32.978***
(6.499)
82.054***
(7.991)
21.781***
(4.748)
55.138***
(6.510)
12.909***
(3.387)
15.336***
(2.948)
5.241*
(3.018)
-8.216***
(3.092)
Eight_Pct_M_Vo -0.943***
(0.207)
-1.102**
(0.473)
-0.616***
(0.156)
-1.037***
(0.339)
-0.438***
(0.112)
-0.575**
(0.226)
-0.017
(0.117)
-0.108
(0.262)
Eight_Emp_Pop_Ratio 0.600***
(0.157)
0.548***
(0.203)
0.129
(0.111)
-0.001
(0.157)
0.121
(0.093)
0.072
(0.120)
-0.130
(0.113)
-0.068
(0.204)
Eight_PUMA_Self_Emp_Pop_Ratio 0.604*
(0.321)
1.485***
(0.459)
0.171
(0.208)
0.567**
(0.257)
0.070
(0.153)
0.106
(0.240)
0.205
(0.147)
0.204
(0.267)
Eight_PUMA_Mttw -0.542**
(0.210)
-1.330***
(0.342)
-0.455***
(0.139)
-0.980***
(0.271)
-0.275***
(0.082)
-0.716***
(0.184)
-0.071
(0.112)
-0.081
(0.339)
New_England 5.033
(5.951)
4.746
(4.722)
1.233
(3.111)
3.247
(2.888)

Middle_Atlantic -6.025*
(3.064)
-0.995
(2.259)
-1.002
(1.619)
0.615
(1.512)

East_North_Central 3.909
(4.193)
4.896**
(2.400)
6.282***
(2.122)
8.195***
(2.508)

West_North_Central -5.704
(3.777)
-2.669
(2.351)
1.279
(2.435)
6.644**
(2.678)

South_Atlantic 5.422
(4.008)
2.740
(2.636)
1.187
(1.508)
0.460
(1.415)

East_South_Central 10.130***
(3.413)
4.001*
(2.414)
2.594
(1.589)
3.383
(2.763)

West_South_Central 2.872
(2.772)
0.805
(1.973)
-0.349
(1.392)
1.642
(1.655)

Mountain -5.260*
(2.756)
-5.055***
(1.822)
-1.285
(1.735)
1.923
(3.976)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 345 345 278 278 234 234 105 105
N 1,395 1,395 1,014 1,014 805 805 231 231
R
2
0.1856 0.3655 0.1632 0.3778 0.1249 0.3624 0.1001 0.4959
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

131

Appendix F.6. Regression Results for the Migration Models – Group C (Age 45-54) (2009)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc_C Nine_MAPLUS_Inmig_Pc_C Nine_SCC_Inmig_Pc_C Nine_BOH_Inmig_Pc_C
Constant -155.602***
(36.818)
-298.911***
(27.376)
-91.225***
(32.958)
-105.060***
(17.877)
-54.227**
(24.344)
-141.335**
(60.105)
16.643
(33.802)
94.396
(69.838)
Eight_PUMA_BAPLUS_Pop_Den
/
Eight_PUMA_MAPLUS_Pop_Den
/
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
12.059×10
-4
***
(3.458×10
-4
)
12.213×10
-4
***
(4.042×10
-4
)
0.003***
(0.001)
0.002***
(0.001)
0.276×10
-4

(2.498×10
-4
)
-1.819×10
-4

(3.070×10
-4
)
4.891×10
-4

(6.429×10
-4
)
3.845×10
-4

(11.235×10
-4
)
Eight_MSA_Inc_BEA_Ln 16.239***
(3.715)
30.903***
(3.044)
10.920***
(3.235)
13.134***
(2.090)
6.997***
(2.560)
16.740**
(6.675)
0.606
(3.506)
-6.912
(7.147)
Eight_Pct_M_Vo -0.395***
(0.112)
-0.447*
(0.263)
-0.193
(0.126)
-0.471*
(0.245)
-0.128
(0.118)
-0.329
(0.300)
0.089
(0.086)
-0.112
(0.241)
Eight_Emp_Pop_Ratio 0.074
(0.088)
0.011
(0.106)
-0.046
(0.070)
-0.074
(0.090)
-0.102
(0.097)
-0.189
(0.155)
-0.249***
(0.083)
-0.091
(0.332)
Eight_PUMA_Self_Emp_Pop_Rati
o
0.616***
(0.148)
0.963***
(0.253)
0.052
(0.099)
0.124
(0.185)
0.087
(0.105)
0.104
(0.218)
0.101
(0.123)
-0.095
(0.313)
Eight_PUMA_Mttw -0.355***
(0.109)
-0.498**
(0.200)
-0.406***
(0.067)
-0.509***
(0.115)
-0.150*
(0.080)
-0.326*
(0.185)
-0.048
(0.132)
-0.143
(0.393)
New_England 3.103
(2.643)
0.509
(1.741)
-0.157
(0.957)
2.449
(2.106)

Middle_Atlantic -0.704
(2.044)
0.759
(1.600)
0.833
(1.357)
3.053*
(1.832)

East_North_Central 5.270**
(2.173)
3.006*
(1.544)
3.159**
(1.264)
3.231**
(1.530)

West_North_Central 4.950***
(1.791)
3.660**
(1.800)
3.166
(2.068)
5.629*
(3.024)

South_Atlantic 4.554***
(1.571)
2.273*
(1.355)
0.456
(0.845)
0.244
(1.172)

East_South_Central 2.352
(2.388)
-0.189
(1.703)
-1.136
(0.958)
-2.128
(1.540)

West_South_Central 2.102
(1.683)
0.227
(1.268)
0.654
(1.139)
2.297
(1.903)

Mountain 1.943
(1.721)
-0.333
(1.389)
0.538
(1.387)
3.533*
(1.943)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 322 322 243 243 193 193 87 87
N 1,236 1,236 799 799 546 546 139 139
R
2
0.1263 0.3307 0.1164 0.3791 0.0454 0.3235 0.1654 0.7377
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

132

Appendix G. Regression Results for the Migration Models – PUMA “Talented” People Population Density (with amenity variable)
Appendix G.1. Regression Results for the Migration Models – Group A (Age 25-34) (2006)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc_A Six_MAPLUS_Inmig_Pc_A Six_SCC_Inmig_Pc_A Six_BOH_Inmig_Pc_A
Constant -529.396***
(148.366)
-2174.067***
(747.615)
-297.493***
(93.741)
-7254.063***
(653.903)
-188.306***
(69.670)
-4555.187**
(2125.333)
79.556*
(47.351)
1093.303***
(182.944)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den /
Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.013***
(0.004)
0.012***
(0.004)
0.014***
(0.005)
0.013***
(0.005)
0.018***
(0.006)
0.016***
(0.006)
0.008***
(0.001)
0.008***
(0.001)
Five_MSA_Pop_BEA_M 16.691***
(5.209)
-50.999***
(15.844)
5.936**
(2.986)
-98.656***
(11.879)
3.503
(2.363)
-60.025***
(15.361)
2.939**
(1.256)
45.630***
(8.325)
Five_MSA_Pop_BEA_L 8.578
(8.463)
-81.745**
(34.501)
8.628
(5.813)
-166.015***
(24.934)
4.824
(3.627)
-185.565***
(58.818)
3.267*
(1.736)
92.381***
(16.858)
Five_MSA_Inc_BEA_Ln 53.365***
(15.128)
201.355***
(70.676)
33.133***
(9.601)
681.676***
(61.379)
19.668***
(6.872)
420.509**
(205.316)
-5.725
(4.346)
-96.765***
(16.482)
Five_Pct_M_Vo -2.679***
(0.458)
-3.072***
(0.782)
-1.240***
(0.319)
-1.368***
(0.522)
-0.586**
(0.266)
-0.740
(0.491)
-0.240**
(0.115)
-0.500**
(0.247)
Five_Emp_Pop_Ratio 2.372***
(0.299)
2.217***
(0.348)
0.774***
(0.186)
0.738***
(0.253)
0.763***
(0.176)
0.734***
(0.222)
0.007
(0.074)
0.025
(0.132)
Five_PUMA_Self_Emp_Pop_Ratio 0.219
(0.465)
1.761***
(0.615)
-0.229
(0.258)
0.487
(0.412)
-0.510*
(0.299)
-0.009
(0.446)
-0.081
(0.208)
-0.224
(0.345)
Five_PUMA_Mttw -2.573***
(0.481)
-4.135***
(0.946)
-1.519***
(0.433)
-2.308***
(0.858)
-0.949***
(0.296)
-1.468**
(0.609)
-0.004
(0.112)
-0.013
(0.165)
January_Average_Temperature -0.743**
(0.295)
-2.736*
(1.617)
-0.313*
(0.167)
11.715***
(1.587)
-0.246**
(0.114)
13.555***
(1.322)
-0.114**
(0.053)
-8.720***
(1.502)
Five_Violent_Crime -0.010
(0.011)
0.558***
(0.179)
-0.009
(0.006)
-0.751***
(0.172)
-0.003
(0.006)
0.151
(0.187)
-0.001
(0.003)
-0.136***
(0.022)
Five_Property_Crime -0.003
(0.003)
0.002
(0.004)
-0.002
(0.001)
0.033***
(0.003)
1.737×10
-4

(0.002)
-0.098***
(0.026)
-5.960×10
-5
(0.001)
0.077***
(0.014)
New_England -3.818
(18.357)
2.506
(12.367)
-4.760
(6.673)
-0.332
(2.254)

Middle_Atlantic -28.520**
(11.417)
-15.109***
(5.713)
-10.355**
(3.997)
-1.399
(2.533)

East_North_Central -26.683***
(8.935)
-12.839**
(4.930)
-13.772***
(4.388)
-5.298***
(1.958)

West_North_Central -39.703***
(10.079)
-18.768***
(6.129)
-15.901***
(5.721)
-4.886**
(2.285)

South_Atlantic 2.801
(6.754)
-0.035
(3.799)
-4.103
(3.010)
-2.612
(1.908)

East_South_Central -7.875
(7.029)
-4.551
(3.063)
-13.016***
(3.709)
-5.284**
(2.157)

West_South_Central -0.089
(7.625)
-2.650
(3.545)
-7.140*
(3.666)
-2.508
(2.065)

Mountain -25.982***
(9.185)
-11.930***
(3.923)
-12.344**
(5.016)
-2.741
(2.236)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 276 276 241 241 218 218 136 136
N 1,273 1,273 1,019 1,019 947 947 406 406
R
2
0.3032 0.4446 0.2662 0.4027 0.1890 0.3418 0.1102 0.2668
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.
133

Appendix G.2. Regression Results for the Migration Models – Group B (Age 35-44) (2006)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc_B Six_MAPLUS_Inmig_Pc_B Six_SCC_Inmig_Pc_B Six_BOH_Inmig_Pc_B
Constant -437.131***
(98.949)
1319.841**
(585.873)
-262.209***
(56.118)
2172.691***
(447.501)
-168.058***
(40.634)
2257.484***
(269.487)
-69.413**
(29.769)
115.244
(643.503)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den /
Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.004***
(0.001)
0.003***
(0.001)
0.005***
(0.001)
0.004***
(0.001)
0.004***
(0.001)
0.003**
(0.001)
0.004***
(0.001)
0.005***
(0.001)
Five_MSA_Pop_BEA_M 5.427**
(2.733)
-40.141***
(10.523)
2.563
(1.906)
35.946***
(8.555)
3.159**
(1.420)
14.511***
(1.531)
1.337
(1.280)
1.725
(10.674)
Five_MSA_Pop_BEA_L 2.588
(4.741)
-58.008***
(19.141)
0.903
(3.642)
117.070***
(19.150)
1.060
(1.885)
47.426***
(4.320)
0.940
(1.830)
7.119
(34.867)
Five_MSA_Inc_BEA_Ln 37.721***
(9.944)
-130.457**
(53.357)
25.015***
(5.368)
-204.983***
(41.469)
16.282***
(3.853)
-229.104***
(27.175)
7.870***
(2.652)
-10.959
(64.609)
Five_Pct_M_Vo -0.526**
(0.245)
-0.328
(0.384)
-0.395**
(0.177)
-0.336
(0.314)
0.028
(0.108)
-0.255
(0.191)
0.016
(0.103)
0.032
(0.206)
Five_Emp_Pop_Ratio 1.244***
(0.206)
1.292***
(0.300)
0.544***
(0.138)
0.656***
(0.176)
0.295***
(0.092)
0.338**
(0.137)
0.050
(0.118)
0.042
(0.296)
Five_PUMA_Self_Emp_Pop_Ratio 0.720**
(0.309)
1.323***
(0.352)
0.151
(0.192)
0.423*
(0.217)
0.186
(0.168)
0.179
(0.234)
0.089
(0.119)
0.193
(0.202)
Five_PUMA_Mttw -0.408*
(0.242)
-1.218***
(0.371)
-0.342**
(0.163)
-0.781***
(0.268)
-0.395***
(0.111)
-0.564***
(0.171)
-0.218**
(0.098)
-0.162
(0.140)
January_Average_Temperature -0.181
(0.143)
-9.393***
(1.212)
-0.253***
(0.094)
-4.417***
(0.723)
-0.041
(0.056)
8.356***
(0.643)
-0.055
(0.065)
0.017
(0.174)
Five_Violent_Crime 0.001
(0.006)
1.127***
(0.107)
0.003
(0.005)
0.027
(0.036)
-0.005
(0.003)
0.412***
(0.043)
0.002
(0.003)
0.053**
(0.025)
Five_Property_Crime 2.212×10
-4
(0.001)
-0.022***
(0.003)
8.140×10
-5

(0.001)
0.027***
(0.002)
4.818×10
-4

(0.001)
-0.097***
(0.009)
-0.001
(0.001)
-0.005
(0.005)
New_England -4.412
(6.856)
-6.309
(4.721)
-2.202
(1.805)
-4.280**
(1.886)

Middle_Atlantic -9.887*
(5.837)
-7.735*
(4.076)
-1.941
(2.061)
-1.086
(2.093)

East_North_Central -7.126
(4.472)
-6.545**
(3.215)
-2.136
(2.258)
-1.273
(2.243)

West_North_Central -12.142**
(5.109)
-10.387***
(3.438)
-2.661
(2.956)
-0.693
(3.846)

South_Atlantic 6.340*
(3.566)
-0.150
(2.193)
0.592
(1.549)
0.631
(1.173)

East_South_Central 3.937
(4.352)
-0.358
(2.963)
-1.000
(2.335)
-2.065
(1.794)

West_South_Central 4.386
(3.757)
-0.747
(2.146)
0.087
(1.851)
0.108
(1.611)

Mountain -7.299
(5.248)
-9.156***
(2.725)
-4.209*
(2.257)
1.945
(2.381)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 260 260 226 226 183 183 94 94
N 1,165 1,165 886 886 669 669 220 220
R
2
0.2314 0.3923 0.2125 0.4133 0.1220 0.3108 0.1273 0.4544
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

134

Appendix G.3. Regression Results for the Migration Models – Group C (Age 45-54) (2006)
Dependent variable
Independent variable Six_BAPLUS_Inmig_Pc_C Six_MAPLUS_Inmig_Pc_C Six_SCC_Inmig_Pc_C Six_BOH_Inmig_Pc_C
Constant -114.248***
(36.072)
-3523.634***
(330.078)
-31.587
(23.536)
2134.916**
(931.795)
-69.361***
(25.956)
23.112
(267.787)
3.266
(20.689)
607.558**
(300.451)
Five_PUMA_BAPLUS_Pop_Den /
Five_PUMA_MAPLUS_Pop_Den /
Five_PUMA_SCC_Pop_Den /
Five_PUMA_BOH_Pop_Den
0.002***
(1.612×10
-4
)
0.002***
(1.888×10
-4
)
0.003***
(3.205×10
-4
)
0.002***
(3.083×10
-4
)
0.001
(0.001)
3.529×10
-4

(9.137×10
-4
)
0.005***
(0.001)
0.004***
(0.001)
Five_MSA_Pop_BEA_M 3.230**
(1.511)
-35.143***
(5.526)
3.429***
(1.250)
35.727***
(12.688)
2.120**
(0.850)
0.138
(2.194)
0.343
(0.904)
8.506**
(4.044)
Five_MSA_Pop_BEA_L 1.064
(2.259)
-93.999***
(12.983)
4.178**
(1.850)
98.262***
(35.648)
1.171
(1.248)
-0.793
(9.314)
1.072
(0.838)
33.181**
(14.650)
Five_MSA_Inc_BEA_Ln 10.503***
(3.548)
337.229***
(32.464)
5.099**
(2.325)
-196.727**
(85.766)
7.268***
(2.437)
-1.953
(26.568)
0.662
(1.824)
-59.202*
(29.584)
Five_Pct_M_Vo -0.257**
(0.110)
-0.252
(0.207)
-0.221**
(0.085)
-0.351**
(0.176)
-0.054
(0.072)
-0.278*
(0.158)
-0.060
(0.119)
-0.093
(0.289)
Five_Emp_Pop_Ratio 0.441***
(0.132)
0.352*
(0.200)
0.040
(0.093)
-0.035
(0.142)
0.150*
(0.080)
0.113
(0.142)
0.054
(0.116)
0.118
(0.189)
Five_PUMA_Self_Emp_Pop_Ratio 0.605***
(0.196)
0.965***
(0.283)
0.453***
(0.127)
0.637***
(0.157)
0.215*
(0.113)
0.417**
(0.189)
-0.049
(0.112)
-0.051
(0.222)
Five_PUMA_Mttw -0.231*
(0.134)
-0.610***
(0.148)
-0.318***
(0.120)
-0.503***
(0.177)
-0.261***
(0.098)
-0.322
(0.195)
-0.260**
(0.098)
-0.327
(0.282)
January_Average_Temperature -0.095
(0.079)
4.480***
(0.352)
-0.143**
(0.061)
-4.270**
(2.005)
-0.070*
(0.038)
0.488**
(0.244)
0.052
(0.037)
0.065
(0.194)
Five_Violent_Crime -0.005
(0.003)
-0.420***
(0.032)
-0.006**
(0.003)
0.046***
(0.014)
0.001
(0.002)
0.030*
(0.017)
-0.004
(0.003)
-0.010
(0.014)
Five_Property_Crime -2.535×10
-4

(0.001)
0.014***
(0.001)
-1.834×10
-4

(0.001)
0.018*
(0.010)
-2.850×10
-4

(0.001)
-0.007***
(0.001)
-1.589×10
-4

(0.001)
0.003**
(0.002)
New_England -4.310
(2.928)
0.453
(2.768)
-3.043
(1.945)
-0.716
(1.422)

Middle_Atlantic -3.594
(2.771)
-3.073
(1.961)
-2.060
(1.495)
1.443
(1.340)

East_North_Central -2.008
(2.439)
-0.179
(1.982)
-2.308
(1.452)
1.392
(1.360)

West_North_Central -1.072
(3.619)
-3.868
(2.391)
-3.234*
(1.836)
4.047**
(1.912)

South_Atlantic 6.070***
(1.686)
2.851**
(1.441)
-0.450
(0.974)
0.208
(0.857)

East_South_Central 5.357**
(2.430)
1.526
(1.749)
2.214
(1.594)
15.813***
(1.748)

West_South_Central 5.689**
(2.338)
2.251
(1.450)
1.383
(1.215)
0.749
(1.497)

Mountain 2.524
(2.794)
0.710
(1.571)
-0.254
(1.690)
0.217
(1.311)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 247 247 208 208 170 170 65 65
N 1,057 1,057 737 737 512 512 140 140
R
2
0.1298 0.3175 0.1382 0.3545 0.0962 0.3238 0.3433 0.5150
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. ***significant at 1% level; ** significant at 5% level; * significant at 10% level.

135

Appendix G.4. Regression Results for the Migration Models – Group A (Age 25-34) (2009)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc_A Nine_MAPLUS_Inmig_Pc_A Nine_SCC_Inmig_Pc_A Nine_BOH_Inmig_Pc_A
Constant -436.015**
(193.476)
-151080.100***
(47703.820)
-284.661***
(89.669)
-1102.229
(1227.850)
-247.199***
(79.239)
15238.320***
(4014.809)
-19.008
(45.040)
-92.950
(160.459)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den / Eight_PUMA_BOH_Pop_Den
0.013***
(0.005)
0.012**
(0.005)
0.012***
(0.005)
0.011**
(0.005)
0.015**
(0.006)
0.013*
(0.007)
0.012***
(0.001)
0.010***
(0.001)
Eight_MSA_Pop_BEA_M 21.999***
(5.739)
-2788.786***
(882.283)
9.373***
(2.829)
-21.483
(22.778)
6.180***
(2.341)
229.984***
(61.621)
4.370**
(1.705)
-7.774***
(2.277)
Eight_MSA_Pop_BEA_L 25.928***
(9.668)
-2860.818***
(911.645)
10.418**
(4.089)
10.500
(17.729)
7.437**
(3.126)
247.749***
(58.784)
6.865**
(2.790)
21.182***
(2.495)
Eight_MSA_Inc_BEA_Ln 55.866***
(18.316)
13946.340***
(4402.641)
34.908***
(9.088)
114.384
(119.969)
29.101***
(7.422)
-1401.234***
(370.561)
5.571
(4.258)
15.518
(15.508)
Eight_Pct_M_Vo -3.092***
(0.575)
-4.248***
(0.943)
-1.393***
(0.362)
-1.682***
(0.439)
-0.881***
(0.227)
-1.340***
(0.291)
-0.147
(0.126)
-0.769***
(0.261)
Eight_Emp_Pop_Ratio 0.811**
(0.311)
0.810**
(0.365)
0.164
(0.192)
0.138
(0.245)
0.199
(0.121)
0.104
(0.134)
-0.244*
(0.142)
-0.222
(0.261)
Eight_PUMA_Self_Emp_Pop_Ratio 0.129
(0.731)
0.788
(0.827)
-0.241
(0.256)
0.124
(0.352)
-0.312
(0.229)
-0.021
(0.274)
0.245
(0.205)
0.012
(0.291)
Eight_PUMA_Mttw -3.078***
(0.629)
-4.635***
(1.012)
-1.424***
(0.354)
-2.218***
(0.635)
-1.157***
(0.229)
-1.697***
(0.326)
-0.414***
(0.131)
-0.708***
(0.208)
January_Average_Temperature -1.058***
(0.298)
184.204***
(58.866)
-0.462***
(0.155)
0.543
(1.553)
-0.215*
(0.113)
-21.626***
(5.637)
-0.144
(0.106)
-0.233
(0.240)
Eight_Violent_Crime -0.023*
(0.012)
-7.217***
(2.369)
-0.008
(0.006)
-0.050
(0.092)
-0.011**
(0.005)
0.752***
(0.180)
-0.004
(0.006)
-0.033***
(0.012)
Eight_Property_Crime -0.001
(0.003)
0.590***
(0.191)
-0.001
(0.002)
-0.005
(0.004)
-4.417×10
-4

(0.001)
-0.011***
(0.001)
0.001
(0.001)
-1.496×10
-4

(0.001)
New_England 8.774
(25.322)
0.258
(10.738)
4.874
(9.215)
3.366
(2.755)

Middle_Atlantic -35.541***
(10.206)
-12.696**
(5.856)
-12.078***
(3.779)
-3.289
(2.776)

East_North_Central -17.376*
(9.865)
-5.616
(5.465)
-6.187
(4.311)
-5.805**
(2.911)

West_North_Central -32.865**
(14.463)
-14.288**
(5.649)
-12.785***
(4.294)
-4.850*
(2.820)

South_Atlantic 3.837
(7.084)
1.638
(3.600)
-1.502
(3.199)
-2.770
(2.034)

East_South_Central -0.296
(9.196)
-2.012
(4.091)
-5.984*
(3.216)
-4.219*
(2.508)

West_South_Central -8.021
(7.948)
-2.469
(3.753)
-7.102**
(2.928)
-5.752**
(2.831)

Mountain -31.755***
(10.048)
-12.848***
(4.887)
-9.980***
(3.568)
-3.919
(2.575)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 241 241 209 209 196 196 110 110
N 1,244 1,244 973 973 892 892 335 335
R
2
0.2755 0.4036 0.2430 0.3626 0.2376 0.3754 0.2102 0.3637
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

136

Appendix G.5. Regression Results for the Migration Models – Group B (Age 35-44) (2009)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc_B Nine_MAPLUS_Inmig_Pc_B Nine_SCC_Inmig_Pc_B Nine_BOH_Inmig_Pc_B
Constant -292.247***
(84.749)
1675.738***
(547.086)
-179.796***
(57.458)
11293.290***
(1884.671)
-117.367**
(53.578)
205.549**
(80.344)
-50.630
(33.392)
232.056
(201.861)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
0.002***
(0.001)
0.002***
(0.001)
0.003***
(0.001)
0.003***
(0.001)
0.004***
(0.001)
0.003***
(4.284×10
-4
)
0.002***
(0.001)
0.002**
(0.001)
Eight_MSA_Pop_BEA_M 5.415**
(2.574)
40.330***
(9.937)
2.200
(1.896)
199.911***
(32.322)
1.938
(1.514)
7.366***
(1.552)
-0.035
(1.677)
10.414
(7.034)
Eight_MSA_Pop_BEA_L 4.753
(4.796)
56.650***
(13.556)
1.846
(3.355)
192.147***
(30.956)
2.305
(2.553)
33.767***
(1.891)
-1.073
(1.881)
11.461*
(6.490)
Eight_MSA_Inc_BEA_Ln 29.595***
(8.203)
-153.421***
(46.858)
20.649***
(5.610)
-1049.667***
(175.264)
12.945**
(5.071)
-9.315
(8.465)
5.423*
(3.122)
-20.641
(19.464)
Eight_Pct_M_Vo -0.930***
(0.213)
-0.702**
(0.322)
-0.726***
(0.166)
-0.757***
(0.243)
-0.431***
(0.104)
-0.431***
(0.140)
-0.040
(0.147)
-0.096
(0.306)
Eight_Emp_Pop_Ratio 0.551***
(0.190)
0.591**
(0.235)
0.052
(0.132)
-0.010
(0.170)
0.074
(0.115)
1.976×10
-4

(0.135)
-0.103
(0.136)
0.056
(0.230)
Eight_PUMA_Self_Emp_Pop_Ratio 0.988***
(0.354)
1.600***
(0.474)
0.368*
(0.190)
0.633**
(0.259)
0.195
(0.179)
0.260
(0.223)
0.249**
(0.122)
0.296
(0.210)
Eight_PUMA_Mttw -0.682**
(0.277)
-1.364***
(0.415)
-0.534**
(0.230)
-1.028***
(0.320)
-0.348***
(0.124)
-0.717***
(0.202)
0.056
(0.123)
0.076
(0.306)
January_Average_Temperature -0.324**
(0.150)
-2.912***
(0.937)
-0.146
(0.094)
-14.624***
(2.381)
-0.092*
(0.055)
-0.722***
(0.089)
-0.004
(0.069)
-0.217
(0.288)
Eight_Violent_Crime -0.013**
(0.006)
0.245***
(0.042)
-0.014***
(0.004)
0.954***
(0.157)
-0.007**
(0.003)
-0.145***
(0.009)
-1.698×10
-4

(0.005)
0.042
(0.029)
Eight_Property_Crime 0.001
(0.002)
-0.018***
(0.005)
3.126×10
-4

(0.001)
-0.045***
(0.007)
0.001
(0.001)
0.001
(0.001)
0.002
(0.001)
-0.010
(0.006)
New_England 0.651
(7.934)
3.536
(6.026)
0.166
(4.273)
1.803
(2.703)

Middle_Atlantic -11.468**
(4.681)
-2.911
(3.471)
-1.677
(1.931)
1.634
(1.724)

East_North_Central -7.927*
(4.628)
1.031
(3.633)
1.982
(3.100)
8.458*
(4.544)

West_North_Central -13.407**
(6.391)
-4.683
(4.149)
0.880
(4.151)
7.564**
(3.202)

South_Atlantic 5.462*
(3.082)
4.344*
(2.310)
0.739
(1.714)
-0.333
(1.713)

East_South_Central 9.630**
(4.594)
6.726**
(3.134)
1.809
(2.461)
3.076
(3.508)

West_South_Central 2.140
(3.473)
1.386
(2.397)
-1.463
(1.949)
0.940
(1.980)

Mountain -10.472***
(3.771)
-6.740***
(2.443)
-4.079**
(1.710)
-2.207
(2.190)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 234 234 198 198 165 165 84 84
N 1,093 1,093 811 811 644 644 195 195
R
2
0.2058 0.3731 0.1897 0.3867 0.1356 0.3713 0.1415 0.4879
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

137

Appendix G.6. Regression Results for the Migration Models – Group C (Age 45-54) (2009)
Dependent variable
Independent variable Nine_BAPLUS_Inmig_Pc_C Nine_MAPLUS_Inmig_Pc_C Nine_SCC_Inmig_Pc_C Nine_BOH_Inmig_Pc_C
Constant -95.027**
(39.568)
43.735
(48.957)
-75.815*
(40.779)
-35816.760***
(2841.781)
-17.752
(25.344)
1081.691**
(452.040)
52.208
(37.606)
0.484
(250.419)
Eight_PUMA_BAPLUS_Pop_Den /
Eight_PUMA_MAPLUS_Pop_Den /
Eight_PUMA_SCC_Pop_Den /
Eight_PUMA_BOH_Pop_Den
0.001***
(1.081×10
-4
)
0.001***
(1.503×10
-4
)

0.002***
(2.624×10
-4
)
0.002***
(2.646×10
-4
)
1.456×10
-4

(2.265×10
-4
)
3.870×10
-6
(3.128×10
-4
)
0.001*
(4.061×10
-4
)
0.001
(0.001)
Eight_MSA_Pop_BEA_M 2.103*
(1.144)
13.552***
(1.399)
0.158
(0.917)
-554.484***
(43.819)
1.967**
(0.944)
11.412***
(2.932)
-0.641
(1.072)
16.368***
(0.771)
Eight_MSA_Pop_BEA_L 3.768**
(1.449)
14.977***
(1.307)
0.167
(1.370)
-524.683***
(40.955)
1.852*
(1.037)
17.219***
(3.761)
0.665
(1.319)
-7.304
(13.496)
Eight_MSA_Inc_BEA_Ln 10.935***
(3.969)
0.949
(4.881)
9.879**
(3.916)
3319.729***
(263.189)
3.555
(2.534)
-99.041**
(42.293)
-3.692
(3.773)
-1.199
(24.939)
Eight_Pct_M_Vo -0.306***
(0.111)
-0.158
(0.196)
-0.130
(0.133)
-0.280
(0.183)
0.009
(0.134)
-0.191
(0.303)
0.144
(0.089)
0.004
(0.280)
Eight_Emp_Pop_Ratio 0.105
(0.094)
0.075
(0.109)
-0.002
(0.070)
-0.028
(0.093)
-0.093
(0.099)
-0.142
(0.167)
-0.128*
(0.071)
-0.255
(0.397)
Eight_PUMA_Self_Emp_Pop_Ratio 0.775***
(0.177)
1.011***
(0.251)
0.156
(0.120)
0.177
(0.188)
0.244*
(0.134)
0.119
(0.227)
0.166
(0.117)
0.071
(0.344)
Eight_PUMA_Mttw -0.469***
(0.129)
-0.521**
(0.238)
-0.394***
(0.088)
-0.510***
(0.150)
-0.136
(0.112)
-0.225
(0.203)
0.033
(0.155)
0.071
(0.358)
January_Average_Temperature -0.175***
(0.055)
-0.662***
(0.118)
-0.203***
(0.049)
48.800***
(3.945)
-0.112***
(0.039)
-0.459***
(0.158)
-0.067
(0.046)
-4.165***
(0.866)
Eight_Violent_Crime -0.001
(0.003)
0.002
(0.024)
3.440×10
-4

(0.003)
-1.757***
(0.132)
-0.001
(0.002)
0.032
(0.024)
0.001
(0.002)
0.353***
(0.077)
Eight_Property_Crime 2.972×10
-4

(0.001)
-0.005***
(0.001)
2.747×10
-4

(0.001)
0.015***
(0.001)
7.800×10
-5
(4.335×10
-4
)
-0.005
(0.003)
3.633×10
-4

(0.001)
0.009***
(0.002)
New_England 4.033
(2.587)
-3.627
(2.554)
-1.849
(1.244)
1.828
(3.492)

Middle_Atlantic -3.412
(2.140)
-3.372*
(1.908)
-1.117
(1.269)
0.922
(1.330)

East_North_Central -1.527
(2.301)
-3.945**
(1.941)
0.678
(1.575)
-0.367
(2.429)

West_North_Central 1.941
(2.248)
-3.963*
(2.307)
3.135
(3.429)
-2.051
(2.271)

South_Atlantic 5.042***
(1.552)
1.241
(1.292)
0.947
(0.953)
-0.645
(1.330)

East_South_Central 1.338
(2.604)
-3.235
(2.063)
-2.448*
(1.391)
-3.230*
(1.900)

West_South_Central 1.288
(1.691)
-1.606
(1.316)
0.115
(1.096)
-0.204
(1.570)

Mountain 0.992
(1.799)
-3.026*
(1.685)
-0.024
(1.449)
1.455
(2.050)

Census Division Fixed Effect Yes No Yes No Yes No Yes No
MSA Fixed Effect No Yes No Yes No Yes No Yes
Number of MSA Clusters 219 219 172 172 143 143 67 67
N 967 967 643 643 442 442 109 109
R
2
0.1557 0.3324 0.1384 0.3664 0.0561 0.2938 0.1973 0.7352
Notes: 1. Cluster-robust standard errors are in parentheses (MSA as the cluster variable).
2. *** significant at 1% level; ** significant at 5% level; * significant at 10% level.

Abstract (if available)

Abstract This dissertation sought to examine whether urban population density matters for talented migrants. Most researchers in the fields of urban planning and urban economics utilize population or employment density based on large geographic units. However these density measures may mask considerable variation across localities and thus just plain “density” is much too general and much too vague to yield useful findings. To overcome this issue, this study constructs various measures of urban population density – MSA population density, PUMA population density, and PUMA “talented” people density – and tests whether they are statistically significant in explaining the in-migration of talented individuals in the continental United States. The dissertation addresses three research questions. First, are denser places especially attractive to talented people? Second, does density become more or less important for the talented migrants in the face of the recent economic downturn? Third, does the density-talent migration nexus vary with age, educational or occupational groups? ❧ For the first research question, estimation results from multivariate tests show that, although the effects of various density measures are small when PUMA densities are utilized, talented individuals do prefer denser places. This finding corroborates what the theories of the agglomeration economies suggest in terms of density preferences for talented individuals. ❧ Turning to the second question, correlation coefficients do show that PUMA population density is relatively stronger in the “bust” year (2009). Multivariate tests using MSA population density show that although the density variable exerts mixed effects on various talented groups in 2006, it has positive effects on the talented in 2009. This suggests that density has become more important for talented migrants in the “bust” year (2009). When PUMA population density and PUMA “talented” people population density are tested however, the difference between the “boom” and “bust” years of density effects is not obvious. Talented movers do not systematically change their density preferences over the recent business cycle. ❧ As for the third research question, although theories suggest that younger people may prefer denser places whereas older people may prefer places with lower densities, the estimation results of this study do not support this prediction for the older talented groups. Various talented groups all prefer denser places. Nonetheless, both the correlation coefficients and multivariate tests do show that the effect of population densities wanes as we move up the age stratum. Density is more important for the youngest (25-34) migrant group. Another point worth mentioning is that by and large PUMA population densities do not matter for the Super-Creative Core and Bohemian migrants belonging to the 45-54 age group. This may suggest that the more senior members of the creative class may be idiosyncratic when making their migration decisions.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Housing market crisis and underwater homeowners

PDF

Essays on congestion, agglomeration, and urban spatial structure

PDF

Urban universities' campus expansion projects in the 21st century: a case study of the University of Southern Calfornia's "Village at USC" project and its potential economic and social impacts on...

PDF

Where there is discretion, should law enforcement officers at the local level be involved in enforcing federal immigration law? a study for consideration

PDF

Essays on the empirics of risk and time preferences in Indonesia

PDF

Three papers on dynamic externalities and regional growth

PDF

The built environment, tour complexity, and active travel

PDF

Urban spatial structure, commuting, and growth in U.S. metropolitan areas

PDF

Reforming developmental education in math: exploring the promise of self-placement and alternative delivery models

PDF

Urbanization beyond the metropolis: three papers on urbanization patterns and their planning implications in the contemporary global south

PDF

The demand for reliable travel: evidence from Los Angeles, and implications for public transit policy

PDF

The functions of the middleman: how intermediary nonprofit organizations support the sector and society

PDF

Collaboration: is it worth it? The Magnolia Community Initiative from the perspective of initiative partner participants

PDF

The economic and political impacts of U.S. federal carbon emissions trading policy across households, sectors and states

PDF

English language development materials in Texas: a study of effectiveness and selection

PDF

Intradepartmental collaboration in the public organizations: implications to practice in an era of resource scarcity and economic uncertainty

PDF

Network-based simulation of air pollution emissions associated with truck operations

PDF

Does collaborative R&D policy work? The effect of collaborative R&D policy on innovative activities of firms in Korea

PDF

Avoiding middle-class planning 2.0: media arts and the future of urban planning

PDF

Healthy mobility: untangling the relationships between the built environment, travel behavior, and environmental health

Asset Metadata

Creator Lin, Cheng-Yi (author)

Core Title Talent migration: does urban density matter?

School School of Policy, Planning and Development

Degree Doctor of Philosophy

Degree Program Policy, Planning, and Development

Publication Date 07/17/2015

Defense Date 08/13/2013

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

Tag density,migration,OAI-PMH Harvest,talent

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Gordon, Peter (committee chair), Moore, James Elliott, II (committee member), Painter, Gary Dean (committee member)

Creator Email chengyil@usc.edu,cylinmike@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-291633

Unique identifier UC11287972

Identifier etd-LinChengYi-1787.pdf (filename),usctheses-c3-291633 (legacy record id)

Legacy Identifier etd-LinChengYi-1787.pdf

Dmrecord 291633

Document Type Dissertation

Rights Lin, Cheng-Yi

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA