retirement migration in the 2000 census - aarpretirement migration one of the myths of retirement in...

Retirement Migration in the 2000 Census

AARP Knowledge ManagementApril 2005

2

AcknowledgementsAARP Knowledge Management prepared this report. We would like to acknowledge the contributions of Ed Evans, lead analyst and author. We would further like to acknowledge Robert Prisuta, the project lead, who provided analytical and editorial consultation, and Carlos Figueiredo of the AARP Public Policy Institute, who provided invaluable technical support in preparing the 2000 Census PUMS 5% data set that made this analysis possible.

3

Table of Contents

47Summary: Retirement Migration in the 2000 Census37Differentiating Movers from Non-Movers Age 60+34Characteristics of Migrants Age 60+ 1995-200032Characteristics of Top Retirement Destination Counties26Characteristics of Retirement Destination Counties21County Migration Flows Age 60+21County-to-County Migration Age 60+ 1995-200010Geographic Variation in Migration Age 60+9Retirement Migration Trends7Retirement Migration 7Background7Retirement Migration in the 2000 Census4Executive Summary

4

One of the myths of retirement in the United States is found in the widely held perception that Americans often move when they retire. In fact, most older Americans prefer not to move if they can avoid it, and most of us tend to become more residentially stable and less mobile as we age.

Migration among the 60+ population has remained very stable over recent decades. The movements of the 60+ population as measured in the 1990 and 2000 Censuses are almost unchanged. In both decades, about nine out of ten Americans age 60+ stayed in the same home or a different home in the same county in the five years before the Census.

The 2000 Census found some 46 million persons ages 60+ in the United States in 2000, up from 41 million in 1990 – an increase of about 9 percent over 10 years.

About 41.5 million or about 90 percent of the 60+ population either stayed in the same house, or moved within the same county, between 1995 and 2000. This dominant group can be thought of as the residentially stable older population.

Nationally, 9.4 percent or 4.4 million of the 60+ population changed communities –moving across state or county borders – between 1995 and 2000. This group can be thought of as the residentially mobile older population.

Executive Summary

5

States that rank highest in terms of the percent of population age 60+ who moved across state borders between 1995 and 2000 were among the fastest growing states in terms of older (60+) population.

Western states such as Nevada (22%) and Arizona (16%) had the highest rates of 60+ in-migration and were also those with the fastest growing senior populations.

Other states including Florida, Delaware, New Mexico, Idaho, New Hampshire Wyoming, Alaska, and Colorado experienced among the fastest rates of in-migration age 60+ between 1995 and 2000, and also demonstrated above average growth in their 60+ populations.

The top 10 percent of U.S. counties in terms of percent 60+ in-migrants (about 300 counties) received over 1 million new 60+ residents between 1995 and 2000 –representing about 20 percent of their total population age 60+. Many of these retirement destination counties are found in the South and West, and often they are smaller counties where more moderate absolute migration volumes represented relatively large shares of total population.

Top destination counties tend to have higher average temperatures, lower crime, lower property and state income tax rates, and higher percentages of recreation/entertainment employment, and are often non-metro counties with high percentages of federal lands – areas that tend to be rich in natural amenities, mostly in the western half of the country.

6

While most older Americans do not move, about 25 percent of those ages 60-64 moved to a different home between 1995 and 2000. This share decreased to about 20 percent of those in their 70’s, and then increased again to 30 percent of those ages 85+.

This bimodal distribution is largely a function of the two dynamics that motivate retirement migration. Amenity migration is represented by the relative spike in moves that is found among the majority of migrants in their early 60’s without disabilities.

As age increases, migration among those without disabilities appears to decrease, but migration among the expanding share who have multiple disabilities increases. The relatively large rates of migration among those with multiple disabilities regardless of age is representative of dependency migration.

The desire to move among those without disabilities appears to decrease as age increases, suggesting that once “settled in,” older Americans would prefer to stay where they are if they can.

Markedly higher rates of migration found among those older Americans with multiple disabilities – regardless of age – suggests that many older Americans may need to relocate in order to more comfortably adjust to the demands of the changing physical capabilities and declining independence that tend to accompany aging.

7

Background

There has been much study and discussion centering on the aging of the baby boom generation – the nearly 80 million Americans born between 1946 and 1964. Currently in 2005, just about half of all boomers have passed the age of 50, and the first boomers will begin turning 60 next year in 2006. By 2025, all 65 million of the boomers then remaining will be over age 60, and the total population age 60+ is projected to be 85 million.

The potential implications for American society are numerous and complex as this large generation moves towards traditional retirement age. While the nature of retirement may indeed change substantially as the boomers pass through it, it is reasonable to expect that many of the trends and patterns that have been part of the contemporary American retirement experience will continue in some form for the baby boom.

Retirement Migration


Retirement Migration in the 2000 Census

8

AL

AK

AZ AR

CA

CO

CT

DE

FL

GA

HI

ID

ILIN

IA

KSKY

LA

MEMI

MN

MS

MO

MT

NE

NV

NH

NM

NY

NC

ND

OH

OK

OR

PA RI

SC

SD

TN

TX

UT

VA

WA

WV

WIWY

Percent 60+ Same House

80% to 82%78% to 80%76% to 78%70% to 76%57% to 70%

Source: Special Tabulation of Census 2000 from AOA and AARP Analysis of PUMS 5% Sample. Prepared by AARP Knowledge Management 2005.

Percent of 60+ Population Living in the Same Home 1995 - 2000

9

Data from recent Censuses bear this out. The Census long form asks Americans if they have moved in the last five years. A special tabulation of the 2000 Census conducted for the Administration on Aging found that 76 percent of Americans age 60+ lived in the same house in 2000 as in 1995. About 2.1 million Americans age 60+ or 4.5 percent of the 60+ total moved across state lines between 1995 and 2000. An additional 350,000 (0.8%) moved to the U.S. from another country between 1995-2000. By comparison, 54 percent of all Americans ages 5+ stayed in the same house between 1995 and 2000, and about 8 percent moved between states in the 5 years before the 2000 Census.

The 2000 Census found some 46 million persons ages 60+ in the United States in 2000, up from 41 million in 1990 – an increase of about 9 percent over 10 years. This increase ismoderate relative to the growth in total population, which grew by 13 percent to 281 million over the decade of the nineties. Of course, overall growth, and growth of the 60+ population varied significantly from state to state.

Retirement Migration Trends

Migration of the 60+ population has remained very stable over recent decades, going back as far as data available from 1960. The movements of the 60+ population as measured in the 1990 and 2000 Censuses are almost unchanged. In both decades, about 90 percent of the 60+ population stayed in the same home or a different home in the same county, and 9 percent moved across county or state borders in the five years prior to the Census.

10

Geographic Variation in Migration Age 60+

Following the general “snow belt to sun belt” trend over recent decades, the 60+ population grew fastest in the West and South in states like Nevada (67%), Alaska (50%), Arizona (38%), and Georgia (20%) and grew least or declined in the Northeast and Midwest, including Pennsylvania (-0.2%), New York (0.5%), and Iowa (0.2%).

60+ PopulationResidence 5 years ago # % # % Same house 5 years ago 32,289,336 77.2% 35,332,950 76.2%Different house in United States 5 years agoSame county 5,446,971 13.0% 6,218,269 13.4%Different county: 3,862,027 9.2% 4,358,768 9.4%

Same State 1,962,472 4.7% 2,273,660 4.9%Different State: 1,899,555 4.5% 2,085,108 4.5%

Northeast 541,192 1.3% 514,341 1.1%Midwest 452,674 1.1% 458,948 1.0%South 523,324 1.3% 643,904 1.4%West 382,365 0.9% 467,915 1.0%

Puerto Rico 11,919 0.0% 18,702 0.0%Outlying Area or Foreign Country 220,784 0.5% 351,736 0.8%Total Age 60+ 41,831,037 100.0% 46,352,613 100.0%

1990 Census (1985-1990) 2000 Census (1995-2000)

Source: 1990 and 2000 Census Special Tabulation on Aging, Administration on Aging. 2000 data includes Puerto Rico.

Migration of the Population Age 60+ in 1990 and 2000 Censuses

11Source: U.S. Census. Special Tabulation of Census 2000 from AOA and AARP Analysis of PUMS 5% Sample. Prepared by AARP Knowledge Management 2005.

Population Age 60+ 2000 and Population Growth Ages 60+: 1990-2000

AL

AK

AZ AR

CA

CO

CT

DEDC

FL

GA

HI

ID

ILIN

IA

KSKY

LA

ME

MD

MAMI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PA

SC

SD

TN

TX

UT

VT

VA

WA

WV

WI

WY

Population Age 60+ 2000

5M

2.5M500K

Population Growth Age 60+1990-2000

20% to 70%10% to 20%5% to 10%0% to 5%

-12% to 0%

12

Growth of the older population in a given place is a function of several key demographic influences and trends, including the size, growth, and age structure of the overall population, mortality rates, international immigration, and internal migration. However, even though a relatively small share of the older population moved across state borders in the 1990’s, this migration was meaningfully related to variations in growth of the 60+ population from state to state.

Percent Change 60+ Population 1990-2000

806040200-20

Perc

ent 6

0+ In

-Mig

rant

s 19

95 -

2000

30

20

10

0

DC

AK

NYIL

CTNJ

MINDIA OH

MNINMA

MD

PA WINELA

RI KS

WY

WVSD

KYCA

VT

MO

VA

NH

OKME

CO

WAHI

MSAL

ORUT

TX

AR

ID

MT

TNGA

NM

NC

DE

SC

FL

AZ

NV

Migration and Growth of the Population Age 60+

13

Indeed, states that rank highest in terms of the percent of population age 60+ who moved across state borders between 1995 and 2000 were among the fastest growing states in terms of older (60+) population.

Western states such as Nevada (22%) and Arizona (16%) had the highest rates of 60+ in-migration and were also those with the fastest growing senior populations.

Other states including Florida, Delaware, New Mexico, Idaho, New Hampshire Wyoming, Alaska, and Colorado experienced among the fastest rates of in-migration age 60+ between 1995 and 2000, and also demonstrated above average growth in their 60+ populations.

We can also think about state-to-state flows in terms of net migration, that is the relationship between the number of people moving in, and the number moving out. Analysis of net migration ages 60+ by state between 1995 and 2000 (using the Census 5% PUMS file which provides state-level data on out-migration) again reinforces the movement of older Americans from the Northeast and Midwest to the South and West. As we might expect, Florida topped the nation with an estimated 275,000 net in-migrants between 1995 and 2000. While there were about 448,000 new 60+ migrants to Florida, there was an offsetting out-migration flow about 173,000. At the other end of the spectrum was New York, which had about 93,000 arrivals age 60+, but lost some 211,000 older residents for a net migration loss of about 119,000 persons age 60+ between 1995 and 2000.

14Source: Special Tabulation of Census 2000 from AOA and AARP Analysis of PUMS 5% Sample. Prepared by AARP Knowledge Management 2005.

Net In-Migration of the 60+ Population 1995 - 2000

MA-9,387

-12,100

AL

AK

AZ AR

CA

CO

CT

DEDC

FL

GA

HI

ID

ILIN

IA

KSKY

LA

ME

MD

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PA

SC

SD

TN

TX

UT

VT

VA

WA

WV

WI

WY

10,185

-2,837

90,155 8,940

13,226

6,065

275,348

27,577

2,549

3,610

-49,948

-8,517

-6,861

-4801,277

-1,688

2,297

-26,770

-7,794

5,951

6,739

3,415

-1,022

38,382

1,658

-27,223

8,161

-118,459

40,308

-1,478

-22,073

5,747

8,507

-11,956

24,488

125

21,319

47,139

3,795

562

8,938

10,361

307

-3,477

37

Net Migration Age 60+ 1995 2000

25,000 to 276,00010,000 to 25,000

0 to 10,000-10,000 to 0

-119,000 to -10,000

15

Other states that were recipients of large net in-migration flows include Arizona (90K), Texas (47K), North Carolina (40K), Nevada (38K), Georgia (27K), South Carolina (24K), and Tennessee (21K).

Notably, the largest state – California – received about 208,000 residents age 60+ from outside the state, but also lost some 195,000 older residents for a net positive flow of just 13,000 against a total 60+ population of 4.7 million in 2000. Were it not for some 79,000 immigrants age 60+ to California from outside the US, California would have shown net negative migration age 60+.

How does the migration of older Americans in the ’90’s compare to the movement of those who were younger? Many of the destination states for older migrants in the South and West were also drawing large numbers of younger Americans. Older migration should perhaps be considered as a component within the larger overall patterns of migration –both helping to create, and responding to the “pushes and pulls” that influence the changing social geography of the United States.

Between 1995 and 2000 older Americans were about half as likely to move from state to state as those who were younger: about 4.5 percent of those ages 60+ moved from state to state compared to about 8.4 percent ages 5+, and 9.4 percent ages 5-59.

16

AL

AK

AZ AR

CA

CO

CT

DEDC

FL

GA

HI

ID

ILIN

IA

KSKY

LA

ME

MD

MAMI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PA

SC

SD

TN

TX

UT

VT

VA

WA

WV

WI

WY

Ages 5-59 Different State

1M

500K100K

% Age 5-59 Different State

18% or more15% to 18%12% to 15%9% to 12%0% or 9%


Number and Percent Population Age 5-59 In-Migrants* 1995-2000

* Population Ages 5-59 in 2000 arriving from another state between 1995 and 2000, excluding immigration.

17

AL

AK

AZ AR

CA

CO

CT

DEDC

FL

GA

HI

ID

ILIN

IA

KSKY

LA

ME

MD

MAMI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PA

SC

SD

TN

TX

UT

VT

VA

WA

WV

WI

WY

60+ Different State

200,000

100,00020,000

% 60+ Different State

15% or more8% to 15%6% to 8%4% to 6%1% to 4%


Number and Percent 60+ Population In-Migrants* 1995 - 2000

*60+ Population in 2000 arriving from another state between 1995 and 2000, excluding immigration.

18

AL

AK

AZ AR

CA

CO

CT

DEDC

FL

GA

HI

ID

ILIN

IA

KSKY

LA

ME

MD

MAMI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PA

SC

SD

TN

TX

UT

VT

VA

WA

WV

WI

WY

State In-Migration by Age Index

1.7 to 1.91.1 to 1.70.9 to 1.10.7 to 0.90 to 0.7


State In-Migration by Age Index: 1995 - 2000

Nationally, the 60+ population was about half as likely (4.5%) to migrate to a new state as the younger population (9.4%). The 60+ State Migration Index compares the state-specific migration by age ratio to the national proportion (48%).

19

There are several States where the older state-to-state in-migration rate is quite high, and disproportionately high relative to the younger rate, based on the national comparison.

In Arizona, about 15 percent of the 60+ population moved from a different state as compared to 17 percent of those ages 5-59.

In Florida, about 11 percent of the older population moved across state lines as compared to 13 percent of the younger population.

And in Nevada – the state with the highest in-migration rate regardless of age – 21 percent of the population 60+ moved from another state between 1995-2000 as compared to 27 percent of those ages 5-59.

Other states where older in-migration appears strong relative to younger migration include New Mexico, Delaware, South Carolina, Arkansas, and Maine.

Older state-to-state migration is often intra-regional – those who move across state borders frequently go to nearby states. In most southern states, the majority of older migrants came from other southern states, followed by those arriving from the Northeast. However new arrivals to Florida were predominantly from the Northeast – and particularly from New York which sent 82,000 residents ages 60+ between 1995 and 2000, followed by other populous states in the Northeast and Midwest including New Jersey, Massachusetts, Michigan, and Illinois. Notably, over 10,000 Californians ages 60+ moved to Florida during the last five years of the ’90’s.

20

In the West, the plurality of in-migration age 60+ came from other states in the West, followed by arrivals from the Midwest and South. Looking at the major western destination of Arizona, of the roughly 130,000 in-migrants 60+ from other states, the largest single source was California (26K), followed by Illinois (11K), Washington (10K), Colorado (6K), Michigan (6K) and Texas, Wisconsin, New York, Oregon, and Minnesota which each sent about 5,000 residents age 60+ to Arizona between 1995 and 2000.

While migration age 60+ to the Northeast and Midwest was light both in terms of volume and share when compared to the South and West, of the migration to the North that was not intra-regional, a significant portion came from the South, particularly with regard to the flows to the larger states including New York, Pennsylvania, Ohio, Michigan, and Illinois. While the data presented here are not conclusive in this regard, this pattern is suggestive of “return migration” – that is the return to “home” states of older migrants in order to be closer to family or other support.

21

Data on residential mobility between 1995 and 2000 at the county level allows us to look at migration of the older population at the community level.

Of course, the counties that attracted the greatest share of new 60+ residents tend to be in the major destination states in the South and West, including Florida, Arizona, and Nevada. But analysis at the county level also reveals patterns of intra-state migration by county.

County Migration Flows Age 60+

Nationally, 2.3 million or 4.9 percent of the total population age 60+ moved to a different county in the same state – a half percentage point more than the proportion that crossed state borders. Interstate and county-to-county migration taken together finds that 9.4 percent of the 60+ population changed communities between 1995 and 2000. Together, this group can be thought of as the residentially mobile older population.

Adding together those that stayed in the same house or moved within the same county finds that 41.5 million or about 90 percent of the 60+ population either stayed in the same house, or moved within the same county, between 1995 and 2000. This dominant group can be thought of as the residentially stable older population.

County-to-County Migration Age 60+ 1995 - 2000

22

Percent of 60+ Population Living in Different County or State 1995 - 2000

Source: Special Tabulation of Census 2000 from AOA. Prepared by AARP Knowledge Management 2005.

AL

AZAR

CA

CO

CT

DEDC

FL

GA

ID

IL IN

IA

KSKY

LA

ME

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PARI

SC

SD

TN

TX

UT

VA

WA

WV

WI

WY

Percent 60+ Different County/ State 1995 2000

20% to 50%15% to 20%10% to 15%5% to 10%0% to 5%

AK

HI

23

AL

AZAR

CA

CO

CT

DEDC

FL

GA

ID

IL IN

IA

KSKY

LA

ME

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PARI

SC

SD

TN

TX

UT

VA

WA

WV

WI

WY

Percent 60+ Same House/County 1995 - 2000

95% or more90% to 95%85% to 90%80% to 85%0% to 85%

HI

AK

Percent of 60+ Living in a Same Home or Same County 1995 - 2000

Source: Special Tabulation of Census 2000 from AOA. Prepared by AARP Knowledge Management 2005.

24

In about 150 of the nation’s 3,141 counties, over 20 percent of the 60+ population moved from another county or state between 1995 and 2000. With a number of notable exceptions these retirement destination counties tend to be in the South and West, and often they are smaller counties where more moderate absolute migration volumes represented relatively large shares of total population.

Sumter County, Florida tops the list with 49 percent of it’s 60+ population of 19,465 arriving between 1995 and 2000. Some of the larger counties near the top of the list include Douglas County, Colorado; Washington County, Utah; Williamson County, Texas; and Pinal County, Arizona.

State CountyTotal Age 60+ 2000

# % Florida Sumter County 19465 9605 49%Utah Daggett County 180 77 43%Nevada Nye County 8495 3500 41%Colorado Douglas County 11330 3975 35%Nevada Storey County 700 235 34%South Carolina McCormick County 2315 770 33%Colorado Archuleta County 1645 539 33%Utah Washington County 19535 6300 32%Virginia James City County 10755 3455 32%Texas Williamson County 24720 7930 32%Arizona Pinal County 38665 12130 31%

Different County or State 1995 - 2000

Top Counties – Percent In-Migrants Age 60+ 1995-2000

25

However, there are also counties that received substantially larger numbers of older residents from other states and counties that were more populous counties initially. Twenty counties in the United States received more than 20,000 in-migrants age 60+ from another county or state between 1995 and 2000. This list includes some very large counties where in-migration represented small shares of total 60+ population, including Los Angeles and San Diego Counties, California and Cook County, Illinois. Eleven counties received more than 20,000 60+ in-migrants that represented 16 percent or more of total 60+ population –approaching twice the national share. This list is topped by Maricopa County, Arizona with over 75,000 new residents age 60+. Seven of these counties are in Florida:

Top Counties – Volume and Percent In-Migrants Age 60+: 1995-2000

State CountyTotal Age 60+ 2000

# %Arizona Maricopa County 465850 76295 16%Florida Palm Beach County 313195 57735 18%California Riverside County 249240 48645 20%Nevada Clark County 205025 48135 23%Florida Lee County 139595 32255 23%New Jersey Ocean County 135950 29580 22%Florida Sarasota County 123235 25230 20%Florida Pasco County 111595 22300 20%Florida Collier County 78130 20955 27%Florida Volusia County 121265 20820 17%Florida Brevard County 120515 20745 17%

Different County or State 1995 - 2000

26

Characteristics of Retirement Destination Counties

What are the characteristics of retirement destination counties? Do they share attributes that can help to explain what brings new older residents to these communities?

A model, using data measured for counties from the U.S. Census Bureau and other public and private sources, has been developed to help frame a picture of county traits that tend to draw retirement age movers. Uniformly measured indicators for climate, employment, economic focus, cost of living, and other relevant data have been considered to identify those measures that may help describe the communities where 60+ in-migration between 1995 and 2000 most contributed to growth.

Separate models have been developed for metropolitan and non-metropolitan counties in order to help distinguish between factors affecting larger, urban counties where total numbers may be larger but percentages are smaller, and smaller rural counties where smaller absolute numbers of older in-migrants represent substantial growth. In addition to modeling the percent of 60+ population who are in-migrants, the model has also been applied to help identify characteristics of counties that attracted the largest absolute numbers of new residents age 60+.

27

The models reveal patterns that are largely in keeping with the conventional wisdom. A report describing retirement migration during the 1980’s by Charles Longino noted that older migrants can be classified into two groups: dependency migrants and amenity migrants:

Dependency migrants are typically forced to move due to deterioration of health or financial resources or the death of a spouse. Amenity migrants are looking for settings that will afford a new and better lifestyle. Communities located on or near lakes, beaches and mountains, and those in temperate climates, have an advantage in attracting this type of migrant, who tends to be recently retired, younger, usually married, and economically better-off than many other retirees (Retirement Migration in America, Charles Longino. 1995)

Longino discussed a number of characteristics of communities found to be appealing to amenity migrants including low crime, mild climate, low taxes, proximity to a major city, low cost of living and housing, good hospitals, and an active social and cultural environment.

While the data included in models considered here do not fully capture all of these dimensions, the models below describing retirement migration in the 1990’s reflect many of these same general tendencies.

28

All CountiesMetro

CountiesNon-Metro

Counties All Counties Percent Percent Percent Number

Model R2 0.23 0.37 0.28 0.25County CharacteristicAverage Temperature 0.22 0.34 0.17 0.17Median Value Owner Occupied Homes 1998 0.15 0.11 0.26 0.33Percent Commuting Times Less Than 15 Minutes 1990 -0.15 -0.10 -0.19 0.01Total Crime Index 1998 -0.12 -0.18 -0.13 -0.06Estimated County Property Tax Revenue (1997) per Dollar Value of Owner Occupied Homes (1998) -0.06 -0.05 -0.03 0.05Highest State Income Tax Rate 1997 -0.08 -0.09 -0.09 -0.03Percent of Total Employment Recreation and Entertainment 1998 0.11 0.11 0.10 0.02Percent of Total Employment Retail 1998 0.22 0.25 0.19 -0.01Percent of Total Employment Health and Related Professional 1998 -0.04 -0.18 0.02 0.06Percent of Total Employment Public Sector 1998 0.04 -0.09 0.09 -0.02Housing Unit Density per Square Mile (Deciles) -0.27 -0.33 -0.25 0.19Non-Metro Counties more than 30% Federal Lands 1993 0.20 na 0.22 0.03

Standardized Regression Coefficients

In-Migrants Age 60+ from Another County or State

A county-level regression model for the percent of total 60+ population who migrated from another county or state identified a number of county characteristics that are related to higher proportions of new 60+ residents by county.

Higher percentages of older in-migrants tend to be found in counties with higher average temperatures, lower crime, lower property and state income tax rates, and higher percentages of recreation/entertainment employment – characteristics very consistent with those described above.

Standardized Regression Coefficients: County In-Migration Age 60+

29

Migration destination counties also tend to be less densely settled and tend to have the longer commuting times associated with urban areas – whether or not they are “Metro” or Non-Metro” counties.*

Non-Metro counties with large percentages of federally owned land* also tended to attract larger percentages of in-migrants age 60+.

Together these factors suggest that migration as a percent of total tends to be stronger in less densely settled non-metro and metro counties, as we should expect. Lower densities and the prevalence of retail employment and longer commuting times all suggest “commuter” or “bedroom” counties, typically on the urban periphery. Non-metro counties that attract higher percentages of in-migrants age 60+ also display a tendency for higher percentages of public sector employment, relative to the balance of non-metro counties. This may suggest a greater availability of government services, relative to less attractive areas.

During the 1990’s there were about 270 counties identified with large proportions (>30%) of Federal land by the USDA. These counties are most often in western states and are often associated with natural amenities including forests, mountains, bodies of water, and the like. While some portion of the pull of these areas for older migrants can be attributed to the broader, ongoing “east to west” migration trend, the propensity of these areas to attract older migrants, along with the tendency for stronger in-migration in counties with relatively high shares of recreation or entertainment related employment, supports the notion that amenity migration is drawn by these attributes.* 1993 USDA Economic Research Service County Classification System

30

AL

AZAR

CA

CO

CT

DEDC

FL

GA

ID

IL IN

IA

KSKY

LA

ME

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PARI

SC

SD

TN

TX

UT

VA

WA

WV

WI

WY

Median Value Owned Homes 1998

100,000 to 501,00075,000 to 100,00050,000 to 75,00045,000 to 50,000

0 to 45,000

AL

AZAR

CA

CO

CT

DEDC

FL

GA

ID

IL IN

IA

KSKY

LA

ME

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PARI

SC

SD

TN

TX

UT

VA

WA

WV

WI

WY

Housing Unit Density 1998

200 or more100 to 20050 to 10010 to 500 to 10

AL

AZAR

CA

CO

CT

DEDC

FL

GA

ID

IL IN

IA

KSKY

LA

ME

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PARI

SC

SD

TN

TX

UT

VA

WA

WV

WI

WY

USDA Metro and Federal Lands Counties

Non-MetroNon-Metro "Federal Lands"Metro

AL

AZAR

CA

CO

CT

DEDC

FL

GA

ID

IL IN

IA

KSKY

LA

ME

MI

MN

MS

MO

MT

NE

NV

NH

NJ

NM

NY

NC

ND

OH

OK

OR

PARI

SC

SD

TN

TX

UT

VA

WA

WV

WI

WY

Percent 60+ Different County/ State 1995 2000

20% to 50%15% to 20%10% to 15%

5% to 10%0% to 5%

Percent County In-Migration Age 60+ Metro, Non-Metro & Federal Lands Counties

Housing Density per Square Mile 1998 Median Value of Owned Homes 1998

31

The models suggests higher home prices in areas that are attracting retirement migrants – even in non-metro counties. Part of this follows from the tendency for the lowest home costs to be found in rural agricultural counties – areas that are not attracting migrants. But beyond this, the reported tendency of amenity migrants to be affluent would indicate a general ability to afford relatively expensive homes – that may be less costly than the homes left behind by these migrants.

Counties that attracted the largest absolute numbers of in-migrants age 60+ also tend to be in warmer areas of the south and west, but because these counties are more typically larger metropolitan counties, home values, home densities, and property taxes tend to be relatively high. Because these larger urban counties also had more diversified economies, employment was not as focused on retail or entertainment/recreation. These larger urban destination counties did however offer greater percentages of professional health related employment, which suggests they may offer more well developed health care services, relative to counties that attract lower volumes of new residents age 60+.

32

The top 10 percent of U.S. Counties in terms of percent 60+ in-migrants (about 300 counties) received over 1 million new 60+ residents between 1995 and 2000 –representing about 20 percent of their total population age 60+.

81 of these top destination counties, or about 40 percent, are non-metro counties with high percentages of Federal Lands (> 30%) – areas that tend to be rich in natural amenities, mostly in the western half of the country. Nationally all 270 “federal lands”counties represent fewer than 10 percent of all counties.

Average temperature in the top 300 retirement destination counties is about 58 degrees – 3 degrees warmer than the national county average and about 6 degrees warmer than the 300 counties with the lowest percentages of new 60+ residents.

Median home value of owned homes in the top 300 counties was about $83,000 in 1997/1998 – some 40 percent higher than the national average.

Characteristics of Top Retirement Destination Counties Age 60+ 1995-2000

33

The highest crime rate in 1998 among the top 300 destination counties was about half the highest rate found in 300 counties with the lowest percentages of new 60+ residents.

On average, entertainment and recreation employment was about double the national average in the top 300 destination counties (average of 2 percent in the top 300 destination counties as compared to about 1 percent across all counties nationwide in 1998).

Retail employment was about 18 percent higher in the 300 retirement destination counties as compared to the nation (county average of 20 percent versus national county average of 17 percent).

Based on the estimates used here, property taxes were some 25 percent lower, and state income taxes were 15 percent lower, in the top 300 60+ destination counties, compared to the nation overall in the 1990s.

34

Measuring Individual Characteristics of Migrants Age 60+ in the Census

The 2000 Census PUMS data can provide us with another perspective on migration of the population over age 60. This data set contains the detailed demographic information available from the Census long form questionnaire for each person in the sample. This information allows us to compare the demographic profile of older migrants to the residentially stable older population.

However, while the Census provides information on residential mobility over a five year period, it is not a longitudinal study. That is, it does not provide “before and after”information for all of the other characteristics measured in the Census. So while the data allows us to compare those who moved in the previous five years to non-movers based on their profiles at the end of the period (2000), it does not allow us to fully identify the degree to which the all of various demographic traits measured in the Census may be associated with the propensity to move.

For example, those who are institutionalized or live with relatives are more likely to have moved in the previous five years. But clearly this tendency is often the result of moving, and can in no way be considered a “predictor” of the propensity to move based on the data in the Census.

Characteristics of Migrants Age 60+: 1995-2000

35

Despite this limitation, the data are instructive. While the data may not help us to fully understand the demographic factors associated with migration, the data provide some insight in this regard, and they certainly can help us to understand how those who have moved are different from those who have not.

The Census 5% PUMS file, while quite large, is restricted geographically for reasons both of confidentiality and statistical precision. For purposes of this analysis we have used the “Super PUMA” geography. This system is comprised of counties and multi-county areas with populations of at least 400,000 persons in 2000. Each sample record in the PUMA data are assigned a geographic code, and those that have migrated are also assigned a code for their previous residence, although there are some further confidentially restrictions on the Migration Super PUMA areas (The Super PUMA areas may be referred to as PUMA1s, PUMAs or areas in the balance of this report.)

For purposes of this part of the study, the 60+ population is divided into three key groupings: those who remained in the same house between 1995 and 2000; those who moved to a different house in the same Super PUMA area, and those who moved to a different Super PUMA area.

36

Because information at this geographic level is less precise than the county data –there are about 540 Super PUMAs as compared to 3,141 counties – the overall migration percentages differ somewhat, but are very similar.

About 77 percent of the 60+ population stayed in the same home between 1995 and 2000, while almost 9 percent of the 60+ population moved from one PUMA region to another. The balance (excluding international immigrants) of about 15 percent moved within their area of residence in 1995.

Notably, the 60+ population who resided in the West Census Region in 1995 appeared the most likely to have moved – either within or without of their PUMA area of residence in 1995.

PUMA Area Migration of the 60+ Population: 1995 - 2000

TotalRegion in 1995 # (000) % # (000) % # (000) % # (000)Northeast 7,581 78.6% 1,246 12.9% 823 8.5% 9,650Midwest 8,381 77.0% 1,588 14.6% 914 8.4% 10,883South 12,403 77.3% 2,292 14.3% 1,348 8.4% 16,043West 6,503 73.5% 1,480 16.7% 865 9.8% 8,848Total 35,347 76.9% 6,662 14.5% 3,965 8.6% 45,974

Same House in 1995Different House

Same PUMADifferent House Different PUMA

37

Region in 2000 TotalRegion in 1995 # (000) % # (000) % # (000) % # (000) % # (000)Northeast 9,266 96.1% 28 0.3% 295 3.1% 54 0.6% 9,650Midwest 17 0.2% 10,545 96.9% 223 2.0% 98 0.9% 10,883South 85 0.5% 93 0.6% 15,768 98.3% 94 0.6% 16,043West 22 0.3% 56 0.6% 97 1.1% 8,672 98.0% 8,848Total 9,390 20.7% 10,721 23.6% 16,383 36.1% 8,919 19.6% 45,974

Northeast Midwest South West

However, not surprisingly, residents age 60+ of the South (98.3%) and West (98%) were more likely to have stayed in their source region than those in the Northeast (96.1%) and Midwest (96.9%).

Differentiating Movers from Non-Movers Age 60+

How do the personal and household characteristics of movers age 60+ differ from those of the residentially stable population age 60+? To begin to explore this question using the 2000 Census PUMS data, a logistic regression model was developed that uses selected data in the PUMS files to distinguish between those who stayed in the same PUMA area (same house or different house) and those who moved to a different PUMA. This construct is arguably a reasonable approximation of the distinction between residential mobility and residential stability in the older population.

PUMA Migration of the 60+ Population by Region: 1995-2000

38

Variable Mean B Wald Sig. Exp(B) R Statistic Owned or mortgaged home 0.78 -0.86 38726.30 0.00 0.42 -0.12Parent of householder 0.05 1.45 32463.33 0.00 4.24 0.11Estimated monthly rent or mortgage payments 311.46 0.00 17666.07 0.00 1.00 0.08Insitutionalized in group quarters 0.04 1.01 16477.01 0.00 2.75 0.08Age 72.05 -0.03 14169.03 0.00 0.97 -0.07Educational attainment 3.11 0.18 12328.09 0.00 1.19 0.07White alone 0.82 0.53 10040.12 0.00 1.70 0.06Not working 0.80 0.49 8623.47 0.00 1.63 0.06Average 60+ estimated property tax rate for PUMA1 0.01 -27.63 6627.70 0.00 0.00 -0.05Other relative of householder 0.01 1.36 6148.88 0.00 3.91 0.05Number of persons in household 2.06 -0.08 1740.37 0.00 0.92 -0.03Number of disabilities 1.03 0.04 895.51 0.00 1.04 0.02Married 0.59 0.11 732.16 0.00 1.12 0.02Gender (male) 0.43 0.09 588.00 0.00 1.09 0.01Child of householder 0.00 0.66 583.29 0.00 1.93 0.01Has Retirement income 0.34 0.09 527.21 0.00 1.09 0.01Has SSI income 0.05 -0.18 488.98 0.00 0.84 -0.01Has Public assistance income 0.02 -0.22 256.30 0.00 0.81 -0.01Has Social Security Income income 0.72 0.04 76.42 0.00 1.04 0.01Household total income in 1999 48354.53 0.00 51.72 0.00 1.00 0.00Has interest income 0.38 0.01 13.39 0.00 1.01 0.00Constant -1.05 3010.75 0.00 0.35

A number of factors were considered as potential discriminators between movers and non-movers, including age, education, disability status, household status, home ownership and costs, gender, marital status, employment status, and sources of income. The table above ranks the factors considered in terms of the strength of the estimated partial correlation they contribute to differentiating those ages 60+ who stayed within the same PUMA area, and those who changed PUMA areas between 1995 and 2000.

Logistic Regression Model for PUMA Migration of the 60+ Population: 1995-2000

39

The strongest predictors of having moved are as we might expect, characteristics that may have changed for those who moved. Movers are less likely to live in an owned home, more likely to have higher monthly housing costs, and substantially more likely to be other than the householder or spouse in their households – more frequently living with a child or other relative. This strong association is a likely reflection of dependency migration, although as mentioned the data here are inconclusive in this regard.

Other demographic characteristics that are most strongly tied to residential mobility include age, education, race, employment status, household size, and disability status. Movers overall tend to be younger, more highly educated, less diverse, have smaller households, are less likely to be working, and more often report disabilities.

Younger, highly educated, white retirees with smaller households (fewer children) are those that we might most closely associate with amenity migration. However some of these traits may also fit the “dependency migrant” profile: Those who are disabled older retirees who live alone may be most prone to move in with relatives or to an institution, in order to cope with declining independence.

40

76%87%

57%77%

14%8%

30%

15%10% 5%

13% 8%

0%10%20%30%40%50%60%70%80%90%

100%

Owned w ith amortgage

Owned without amortgage

Rented for cashrent

Occupiedwithout payment

of cash rent

Same House Moved, Same PUMA Moved, Different PUMA

82%72% 69% 68% 66%

11%20% 19% 18% 19%

7% 9% 12% 14% 15%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

$0 $1-$499 $500-$999 $1000-$1499 $1500+


PUMA Migration by Monthly Housing Cost

PUMA Migration by Tenure

As noted above, renters are most likely to be migrants who changed areas (13%). This holds to an even greater degree for those who moved within the same area (30%). This may relate to the “cashing out”of home equity for some movers, or the relative flexibility afforded by renting.

Movers may pay more for housing as a group because they might be giving up the low cost of “old”mortgages, or even mortgage free ownership to purchase or rent a home which may be relatively expensive, particularly if the new home is in a convenient or “amenity rich” location.

41

72% 76% 77% 78% 76% 76%

21% 18% 16% 13% 14% 13%

7% 6% 7% 8% 10% 11%

0%10%20%30%40%50%60%70%80%90%

100%

none grades P-6 grades 7-12

HSgraduate

somecollege

collegegrad


68%80% 80% 80% 79% 78% 78%

22%13% 12% 12% 12% 13% 13%

10% 7% 8% 8% 8% 9% 10%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

<$15K $15-25K $25-35K $35-50K $50-75K $75-100K >$100K


PUMA Migration by Income

PUMA Migration by EducationThose who are more highly educated are more likely to have changed PUMA areas between 1995 and 2000 (11% of college graduates), while those with the least education are most likely to have moved within areas (21%). This may relate to generally higher transience among the less affluent, or may relate indirectly to age because education levels have increased over time.

Similarly, overall migration is highest among those with the lowest incomes (32% less than $15K), but area-to-area migration appears very slightly higher at either end of the income scale. Recall that income did not demonstrate a strong effect in the multivariate model for area-to-area movement.

42

77% 78% 78% 72% 71% 73% 76%

14% 16% 14% 19% 16% 15%17%

9% 6% 8% 9% 13% 12% 6%

0%10%

20%30%

40%50%60%

70%80%

90%100%

White B lack N at iveA merican

A sian P acif icIs lander

Other/ 2o r mo re

H ispanicany race


81%

57% 59%

31%

12%

24% 25%

45%

7%19% 16%

24%

0%10%20%30%40%50%60%70%80%90%

100%

Householder /spouse

parent Other Relative Institutionalgroup quarters


PUMA Migration by Household Relationship

PUMA Migration by Race/EthnicityAs the multivariate model suggests, Whites are more likely to have moved from area to area (8%) than Blacks and Hispanics (6%). Notably, Asians, Pacific Islanders, and those who reported multiple or other races are the most likely to move – perhaps because of an enhanced tendency for elders to live with children in these groups.

As noted previously, those living with children, other relatives, or in institutional group quarters in 2000, were substantially more likely to have moved since 1995 – because the individual would have very often moved into this new situation from that of householder or spouse.

43

75% 78% 80% 79% 77%69%

15% 13% 13% 13% 15%20%

10% 8% 8% 7% 8% 10%

0%10%

20%30%

40%50%

60%70%

80%90%

100%

60-64 65-69 70-74 75-79 80-84 85+


80% 77%70%

57%

12% 15%20%

28%

8% 8% 11% 15%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

none one or two three or four five or six


PUMA Migration by Disability

PUMA Migration by Age

Migration among the 60+ population -- whether within or between areas -- is most prevalent among those at either end of the age scale (25% under 65, and 30% of 85+). Movement within a given area increases notably after age 85, suggesting the increased frequency of dependency migration in this age group we should expect.

Both long and short distance migration appear to increase steadily with the number of disabilities reported in the Census (20% with zero disabilities to 43% with 5 or 6 disabilities), suggesting the dependency migration that is motivated by the functional limitations associated with aging.

44

% MovedPopulation

(000) % MovedPopulation




(000)60-64 25.0% 10,865 23.8% 7,504 26.0% 2,537 31.8% 668 35.8% 15565-69 21.9% 9,272 20.6% 6,011 22.6% 2,437 27.3% 648 37.7% 17670-74 20.1% 9,000 18.2% 5,459 20.6% 2,528 27.1% 760 37.6% 25375-79 20.9% 7,372 17.3% 3,731 20.4% 2,370 29.2% 886 40.1% 38580-84 23.3% 5,171 16.9% 2,108 21.6% 1,720 30.3% 885 45.0% 45885+ 30.7% 4,294 18.8% 950 24.3% 1,263 34.5% 1,125 46.7% 955Total 23.1% 45,974 20.2% 25,764 22.5% 12,856 30.4% 4,972 43.0% 2,382

Total

Age

No Disabilities 1 or 2 Disabilities 3 or 4 Disabilities 5 or 6 Disabilities

While the data considered here do not allow us to fully distinguish between amenity and dependency migrants, a look at total residential movement by age and disability status provides evidence that supports the expected patterns.

Overall, about 44 percent of the 60+ population reported one or more of the six disabilities measured in the 2000 Census. This percentage increases dramatically with age from about 30 percent among those ages 60-64 to nearly four out of five (78%) ages 85+.

As noted earlier, residential movement among the older population appears to decline with age through the sixties and early seventies, and then increase again through the late seventies and eighties.

Moved to a Different House Ages 60+ by Age and Disability Status: 1995-2000

45

Percent of the 60+ Population Moving to a Different House by Age and Disability Status

25%22%

20% 21%23%

24%21%

18% 17% 17%19%

36%38% 38%

40%

45% 47%

31%

10%

15%

20%

25%

30%

35%

40%

45%

50%

60-64 65-69 70-74 75-79 80-84 85+

Total No Disabilities 1 or 2 Disabilities3 or 4 Disabilities 5 or 6 Disabilities

1,7901,236 991

647357 1790

500

1,000

1,500

2,000

2,500

3,000

60-64 65-69 70-74 75-79 80-84 85+

Thou

sand

s

No Disabilities 1or 2 Disabilities 3 or 4 Disabilities 5 or 6 Disabilities

2,717

2,0291,813

1,543

1,203 1,320

60+ Population Moving to a Different Houseby Age and Disability Status

Overall, the percent of the population age 60+ living in a different house after five years increased with age to 30 percent among those age 85+. However, among those with no disabilities, moving tended to decline gradually with age and did not increase nearly as sharply after age 80.

Conversely, moving increased substantially with the number of disabilities regardless of age, and tended to increase sharply with age among those with higher numbers of disabilities.

46

These data may provide some additional insight into the dynamics that undergird the residential migration of older Americans. While most older Americans do not move, about 25 percent of those ages 60-64 moved between 1995 and 2000. This share decreased to about 20 percent of those in their 70’s, and then increased again to 30 percent of those ages 85+.

This bimodal distribution is largely a function of the two dynamics that motivate retirement migration. Amenity migration is represented by the relative spike in moves that is found among the majority of migrants in their early 60’s without disabilities. As age increases, migration among those without disabilities appears to decrease, but migration among the expanding share who have multiple disabilities increases. The relatively large rates of migration among those with multiple disabilities regardless of age is representative of dependency migration.

These data indicate that the desire to move among those without disabilities appears to decrease as age increases, suggesting that once “settled in,” older Americans would prefer to stay where they are if they can. But the markedly higher rates of migration found among those older Americans with multiple disabilities – regardless of age --suggests that many older Americans may need to relocate in order to more comfortably adjust to the demands of the changing physical capabilities and declining independence that tend to accompany aging.

47


The 2000 Census found some 46 million persons ages 60+ in the United States in 2000, up from 41 million in 1990 -- an increase of about 9 percent over 10 years. This increase is moderate relative to the growth in total population, which grew by 13 percent to 281 million over the decade of the nineties. Of course, overall growth, and growth of the 60+ population varied significantly from state to state.

Migration among the 60+ population has remained very stable over recent decades. The movements of the 60+ population as measured in the 1990 and 2000 Censuses are almost unchanged. In both decades, about 90 percent of the 60+ population stayed in the same home or a different home in the same county, and 9 percent moved across county or state borders in the five years prior to the Census.

About 2.1 million Americans age 60+ or 4.5 percent of the 60+ total moved across state lines between 1995 and 2000. An additional 350,000 (0.8%) moved to the U.S. from another country between 1995-2000. By comparison, 54 percent of all Americans ages 5+ stayed in the same house between 1995 and 2000, and about 8 percent moved between states in the 5 years before the 2000 Census.

Summary: Retirement Migration in the 2000 Census

48

Nationally, 2.3 million or 4.9 percent of the total population age 60+ moved to a different county in the same state – a half percentage point more than the proportion that crossed state borders. Interstate and county-to-county migration taken together finds that 9.4 percent of the 60+ population changed communities between 1995 and 2000. Together, this group can be thought of as the residentially mobile older population.

Adding together those that stayed in the same house or moved within the same county finds that 41.5 million or about 90 percent of the 60+ population either stayed in the same house, or moved within the same county between 1995 and 2000. This dominant group can be thought of as the residentially stable older population.

The top 10 percent of U.S. counties in terms of percent 60+ in-migrants (about 300 counties) received over 1 million new 60+ residents between 1995 and 2000 –representing about 20 percent of their total population age 60+. Many of these retirement destination counties are found in the South and West, and often they are smaller counties where more moderate absolute migration volumes represented relatively large shares of total population.

81 of these top destination counties, or about 40 percent, are non-metro counties with high percentages of federal land – areas that tend to be rich in natural amenities, mostly in the western half of the country. Nationally all 270 “federal lands” counties represent fewer than 10 percent of all counties.

49

A county-level regression model for the percent of total 60+ population who migrated from another county or state finds that higher percentages of older in-migrants tend to be found in counties with higher average temperatures, lower crime, lower property and state income tax rates, and higher percentages of recreation/entertainment employment.

The 2000 Census PUMS data can provide us with another perspective on migration of the population over age 60. This data set contains the detailed demographic information available from the Census long form questionnaire for each person in the sample. This information allows us to compare the demographic profile of older migrants to the residentially stable older population.

Movers are less likely to live in an owned home than non-movers, more likely to have higher monthly housing costs, and substantially more likely to be other than the householder or spouse in their households – more frequently living with a child or other relative. Movers overall tend to be younger, more highly educated, less diverse, have smaller households, are less likely to be working, and more often report disabilities.

While most older Americans do not move, about 25 percent of those ages 60-64 moved between 1995 and 2000. This share decreased to about 20 percent of those in their 70’s, and then increased again to 30 percent of those ages 85+.

50

This bimodal distribution is largely a function of the two dynamics that motivate retirement migration. Amenity migration is represented by the relative spike in moves that is found among the majority of migrants in their early 60’s without disabilities. As age increases, migration among those without disabilities appears to decrease, but migration among the expanding share who have multiple disabilities increases. The relatively large rates of migration among those with multiple disabilities regardless of age is representative of dependency migration.

These data indicate that the desire to move among those without disabilities appears to decrease as age increases, suggesting that once “settled in,” older Americans would prefer to stay where they are if they can. But the markedly higher rates of migration found among those older Americans with multiple disabilities –regardless of age -- suggests that many older Americans may need to relocate in order to more comfortably adjust to the demands of the changing physical capabilities and declining independence that tend to accompany aging.

retirement migration in the 2000 census - aarpretirement migration one of the myths of retirement in...

Documents