Where Do We Go When We Retire? A Broader Look at Retirement Destinations

Bookmark and Share

Ben Bolender, Population Division, U.S. Census Bureau

Where do you plan to live when you or a loved one reaches the golden years? Do you plan to stay where you are now? Do you plan to move to some place warm and sunny? Our new research from the 2014 population estimates looks at where people age 65 and over are choosing to move.

People moving from place to place are reflected in the measures of migration created every year for the Census Bureau’s population estimates. We use primarily Medicare enrollment and group quarters data to calculate the movement of the population age 65 and over. People move for a variety of reasons. The tendency to migrate is highest for young adults. However, in the ages following retirement (commonly around 65), we see another spike in migration. With the baby boomers being the largest cohort yet to enter these ages, the question of where people move when they retire is becoming increasingly important.

Sumter County, Fla., has been one of the fastest-growing counties in recent years, at least partially because of its planned retirement destination status. Based on our research into older age migration patterns, we see that Sumter attracted just over 11,000 older people from April 1, 2010 (Census Day) to July 1, 2014, from other parts of the U.S. Another county with a strong draw was Maricopa, Ariz., which gained almost 18,000 older people during this same period through migration from other parts of the U.S.

While it is true that these areas do attract the most older-age migrants, they do not reflect the wide variety of places that attract retirement-age people. Examining numeric change tells one story, but it is often useful to look at other measures of change. The picture changes when we group counties by population size and look at percent change in the older population from migration. For example, Sumter County ranks high in this measure, showing an increase of 27.7 percent in the 65-and-older population through just domestic migration from 2010 to 2014. The comparable number for Maricopa County was only 3.9 percent.

For counties with a total population of 50,000 or more in 2014, Sumter leads the list for fastest-growing older population through domestic migration. This was followed by Broomfield, Colo., with a 17.6 percent increase, and Rockwall, Texas (part of the Dallas metro area), with a 16.9 percent increase. Here we see that both midsized mountain towns and suburban areas saw their older populations rise rapidly through migration.

When we look at counties with populations between 20,000 and 50,000, Jasper, S.C., leads the way at 28.1 percent, followed by Kendall, Texas, at 14.9 percent. Many older people choose to move to suburbs and areas adjacent to larger cities. Both of these counties are on the outskirts of larger metropolitan statistical areas.

For counties with fewer than 20,000 people in 2014, looking at percent growth starts to pick up much smaller numeric increases. Stark, Ill., leads the list at 17.3 percent growth, followed by Lake, S.D., at 13.9 percent and Sequatchie, Tenn., at 12.6 percent. These three areas represent different flavors of rural America in their plains, forests, valleys, and fields. Even with the variety of places where you could choose to retire, with the aging of the baby boomers it is likely that when you arrive you will find others who made the same choice.

Posted in Uncategorized | 3 Comments

How Did People Experience Poverty from 2009 to 2012?

Bookmark and Share

Note: Census Bureau experts are presenting on a variety of topics at the Population Association of America annual conference. Follow the Research Matters blog or visit the press kit to learn more about their work.

Written by Ashley Edwards

Typical poverty statistics are a snapshot of the population in poverty at the time of the survey.  Are the poor young or old?  Single-parents, in large families, or childless?  New estimates  highlight an easy-to-overlook aspect of poverty:  It is not a constant and individuals move in and out of poverty.  Data from the Survey of Income and Program Participation allowed me to estimate the median length of time individuals spent in poverty between 2009 and 2012 as well as the number of times individuals shifted in and out of poverty.  The data highlight the diverse poverty experiences of different demographic groups in the early years of the current economic recovery.

Of individuals who entered poverty at some point between 2009 and 2012,  the median length of a poverty spell was 6.2 months. However, some individuals experienced repeated poverty spells, and measures based on spell length failed to account for the often cyclical nature of poverty. Over this four-year period, more than half (54.5 percent) of individuals who entered poverty experienced only a single spell, while 45.5 percent had two or more poverty spells.

My research shows children under the age of 18 and Hispanics tended to experience more poverty spells than the overall population experiencing poverty from 2009 to 2012. Individuals age 65 and older experienced fewer and longer poverty spells, but their median total time in poverty was not statistically different from the overall ever-poor population.

Blacks and individuals in female-householder families did not experience more frequent poverty spells, but their median spell lengths were longer. In contrast, Hispanics experienced more poverty spells, but their individual spell lengths were not any longer than the ever-poor.

Title of Graph – Selected Family and Demographic Characteristics by Poverty Spell Frequency and Duration: 2009 to 2012

poverty1Initiated in 1983, the Survey of Income and Program Participation provides a wealth of information to analyze the economic situation of people in the United States. It offers detailed information on cash and noncash income, while also collecting data on taxes, assets, liabilities and participation in government transfer programs. The data allow the government to evaluate the effectiveness of federal, state and local programs.

The 1982 annual poverty thresholds were used as the base to calculate monthly poverty thresholds during the reference period. These annual thresholds were then divided by 12 and adjusted using the consumer price index for a given reference month to provide a monthly poverty threshold by family size. For example, in January 2009, the monthly poverty threshold for a family of four with two children was $1,784, and in December 2012, the monthly poverty threshold for the same family was $1,940.

Note: The Census Bureau’s Trudi Renwick has conducted related poverty research on the supplemental poverty measure. A poster presented at the Population Association of America annual conference displays the results of using data from the Current Population Survey Annual Social and Economic Supplement to estimate the value of variables not included in the American Community Survey but critical to the production of supplemental poverty measure estimates from the American Community Survey. Her poster examines the effect of government noncash benefits and necessary expenditures on state level supplemental poverty measure rates.  See the results of this research at:

Posted in Uncategorized | 2 Comments

Talkin’ ‘Bout Our Generations: Will Millennials Have a Similar Impact on America’s Institutions as the Baby Boomers?

Bookmark and Share

Note: Census Bureau experts are presenting on a variety of topics at the Population Association of America annual conference. Follow the Research Matters blog or visit the press kit to learn more about their work.

Written by Sandra Colby, Population Division

Over the next several years, baby boomers will continue transitioning into retirement and old age while millennials, many of whom are children of the baby boomers, pass through the traditional benchmarks of adulthood (e.g., finishing school, finding jobs and buying homes). Researchers and reporters, among others, have drawn comparisons between the experiences and behaviors of these two generations.

However, these comparisons often overlook one important difference between the generations — their memberships are not defined by the same metric. That is, while birth cohorts included in the baby boomer generation are associated with a rise in fertility, no similar demographic event can be used to distinguish the birth cohorts included in the millennial generation. Notably, shared experiences rather than demographics define the millennials.

Because of this, the birth years for this generation are not as distinguishable as those of the baby boomers. The Census Bureau does not provide guidance on which years are included in the millennial generation, and many definitions are used by the public. For the purposes of this blog, I use the term “millennials” to encompass those born between 1982 and 2000.

Figure 1 shows births for the years 1909 through 2013. A large increase in the number of births between 1945 and 1946 marks the start of the baby boom generation, but there is no corresponding increase to establish the start of the millennial generation. Gens1

Although the birth cohorts comprising the millennial generation were as large, and in some cases larger, than those of the baby boomer generation, the millennial generation differs from the baby boomers because these large birth cohorts are part of a broader trend that started in the previous generation and is continuing into the next.  In other words, the fertility trends associated with the beginning of the millennial generation are not exceptional.

One consequence of this difference relates to the impact that these generations have on societal institutions. Despite the similar size of the millennial generation relative to the baby boomers, their transition through life thus far has not introduced the same level of shock as the baby boomers caused.

To illustrate this point, I use the example of school enrollment for 5 and 6 year olds. Figure 2 shows the annual change in the number of 5 and 6 year olds enrolled in school. For the majority of years, the number of students remained relatively stable. One exception was the 1.4 million increase in students between 1953 and 1954. The growth in 1954 corresponds to children born between 1947 and 1949 (the first of the baby boomers) enrolling in school. This increase is nearly triple that experienced at any point during the millennial generation’s transition into school and represents an increasing demand on the education system not replicated by the millennials.


The goal here is not to minimize the millennial generation’s significance, only to highlight important differences in the origin of this generation and its impact on societal institutions as its members transition through the course of life.

Posted in Uncategorized | Leave a comment

When Do Mothers Earn More? A Look at Fertility Timing and Occupation

Bookmark and Share

Note: Census Bureau experts are presenting on a variety of topics at the Population Association of America annual conference. Follow the Research Matters blog or visit the press kit to learn more about their work.

Written By: Liana Christin Landivar, Sociologist, Industry and Occupation Statistics Branch

Researchers have highlighted a consistent motherhood “wage penalty” of 6 to 7 percent for mothers of one child and 12 to 13 percent for mothers of multiple children.  Another  examination of the data, however, shows that there are certain circumstances where mothers outearn nonmothers.

Using data collected in the 2013 American Community Survey, I look at the earnings gap between mothers and nonmothers, and then examine how the earnings gap varies by occupation and age of children. My research shows that while mothers were more likely to be out of the labor force or work part time than nonmothers, mothers earned more than nonmothers among full-time, year-round workers. One explanation provided in prior research is that mothers with higher earnings potential may be more likely to remain employed full time.

Median earnings of women ages 18-50 employed full time, year-round, 2013


Fertility delay has been linked to higher earnings for mothers. The mean age at first birth in 2013 was 26 years old and a growing number of women are postponing children. When looking at earnings by age, we see that younger mothers earned less than nonmothers in the same age group, while older mothers earned more than nonmothers in the same age group. Mothers in the youngest age group, 18-29, experienced the largest earnings penalty relative to nonmothers. Mothers in the oldest age group, 40-50, experienced the largest earnings premium, particularly among mothers of preschool-age children (ages 0-5).

Median earnings by age of mother and age and presence of children, 2013


The earnings gap also varied by occupation. Mothers in the managerial and professional occupation group experienced the largest earnings penalties for young motherhood, but also the largest earnings premium for delayed fertility. In managerial and professional occupations, mothers of preschoolers earned $11,000 more than nonmothers of the same age if they had children between the ages of 40 and 50. Women in managerial and professional occupations with the earliest fertility, that is, mothers ages 18 to 29 with school-age children (ages 6 to 17), earned $9,000 less than nonmothers of the same age. Having children at older ages did not translate into an earnings premium for women in construction, production, agriculture, health care support, cleaning and maintenance, or food preparation, where mothers earned less than nonmothers or there is no statistical difference in their earnings.

Mothers earned more than nonmothers when they had children at older ages and they worked in managerial and professional, protective service or sales occupations. From an earnings point of view, delaying fertility (that is, putting off when they have their first child) may be particularly important for women in occupations requiring advanced degrees or longer tenure for career advancement, because younger mothers may never catch up to the earnings of women who wait to have children or who never have children.  In other occupations, however, delaying fertility is not associated with higher earnings for mothers.

Earnings penalty or premium by occupation and age among mothers of preschool children, 2013*

The estimates presented are based on responses from a sample of the population. As with all surveys, estimates may vary from the actual values because of sampling variation or other factors. For information on confidentiality protection, sampling error, nonsampling error, and definitions, see: https://www.census.gov/acs/www/Downloads/data_documentation/Accuracy/ACS_Accuracy_of_Data_2013.pdf

Posted in Uncategorized | 1 Comment

China Replaces Mexico as the Top Sending Country for Immigrants to the United States

Bookmark and Share

Note: Census Bureau experts are presenting on a variety of topics at the Population Association of America annual conference. Follow the Research Matters blog or visit the press kit to learn more about their work.

Written by Eric Jensen

Based on my research, in 2013, China replaced Mexico as the top sending country for immigrants to the United States. This followed a decade where immigration from China and India increased while immigration from Mexico decreased. Other top immigrant-sending countries in 2013 from Asia included Korea, the Philippines and Japan. This new pattern in the national origins of recent immigrants is a notable change from recent decades.

The racial and ethnic composition of immigration flows to the United States has also been shifting. In 2000, nearly half of all foreign-born immigrants, 41.2 percent, were Hispanic, compared with 23.6 percent for the non-Hispanic Asian alone population. Since 2009, a greater proportion of foreign-born immigrants have been non-Hispanic Asian alone (34.7 percent) than Hispanic (30.1 percent). By 2013, the percentage of non-Hispanic Asian alone had increased to 40.2 percent of the total immigration flow, while the percentage Hispanic had dropped to 25.5 percent.

The U.S. Census Bureau’s Population Estimates Program measures net international migration, including the foreign-born population whose residence one year ago was abroad. According to the 2013 American Community Survey, there were 1,201,000 immigrants. China was the top sending country with 147,000, followed by India with 129,000, and Mexico with 125,000. The numbers of immigrants from India and Mexico were not statistically different from each other. In 2012, the American Community Survey showed that Mexico and China were the top two sending countries with 125,000 and 124,000, respectively (which were not significantly different from each other).

Change in the racial and ethnic composition of immigrant flows contributes to the overall racial and ethnic makeup of the United States. While Hispanics are still the largest racial or ethnic minority group, a larger percentage of the Asian population was foreign-born (65.4) compared with the Hispanic population (35.2) in 2013. Given the numbers above, it is likely that the contribution of immigration to overall population growth will be greater for Asians than for Hispanics.

Historically, the national origins of immigrant flows have changed dramatically. The earliest waves of immigrants originated in Northern and Western Europe. Immigrants from Southern and Eastern Europe later predominated. The most recent wave of immigrants has largely been from Latin America, and to a lesser extent, Asia. Whether these recent trends signal a new and distinct wave of immigration is yet to be seen.

The figure below shows the foreign-born population whose residence one year ago was abroad for China, India and Mexico from 2000 to 2013.

Foreign-Born Population Whose Residence One Year Ago Was Abroad by Selected Places of Birth: 2000-2013



Posted in Uncategorized | Leave a comment

Disparities in Health Insurance Coverage

Bookmark and Share

Note: Census Bureau experts are presenting on a variety of topics at the Population Association of America annual conference.  Follow the Research Matters blog or visit the press kit to learn more about their work.

Written by Amy Steinweg and Carla Medalia, Statisticians, Social, Economic, and Housing Statistics Division

Rates of health insurance coverage vary across population groups. Two recent projects by researchers from the U.S. Census Bureau explore some of these disparities in health insurance coverage.

The first project, Disparities in Private Health Insurance Loss in the Wake of the Great Recession, examines which population groups lost their private health insurance coverage after the 2007-2009 recession. Data from the 2008 panel of the Survey of Income and Program Participation show that many groups who were less likely to have coverage during the recession were also more likely to lose that coverage by December of 2011.

Disparities across many demographic and socioeconomic groups may have grown over the 2008-2011 period. For example, in August 2008, 79.4 percent of non-Hispanic Whites had private health insurance compared with only 59.1 percent of non-Hispanic Blacks (see Figure 1). Of those who had private health insurance initially, rates of coverage loss were nearly twice as high for non-Hispanic Blacks as for non-Hispanic Whites. In other words, during this time, non-Hispanic Blacks were less likely to have private health insurance initially, and then more likely to lose that coverage subsequently.

Rates of Private Health Insurance at Baseline, and Subsequesn Loss, By Selected Characteristics

A second project explores whether the Patient Protection and Affordable Care Act has the potential to reduce health insurance coverage disparities across demographic groups. This project, Health Insurance Disparities and the Affordable Care Act: Where Could Inequality Decline?, uses data from the Current Population Survey Annual Social and Economic Supplement to examine current and potential future gaps in the uninsured rate.

The results suggest that disparities may decrease with the change in the law and may decrease more in states that expanded Medicaid eligibility. For example, in 2013, in states that would go on to expand Medicaid, 12.2 percent of non-Hispanic Whites and 18.9 percent of non-Hispanic Blacks were uninsured (see Figure 2). If all eligible uninsured individuals took advantage of new coverage options through the Affordable Care Act, the uninsured rate could decrease for both groups, and the coverage gap between non-Hispanic Blacks and non-Hispanic Whites could be largely reduced. At the same time, in non-expansion states, while both groups could benefit from the Affordable Care Act, the coverage gap could be reduced but not closed.

Observed and Potential Uninusred Rates for the Popoulation Ages 19 to 64 by Characteristic


This research establishes a benchmark to evaluate how closely future changes in the uninsured rate associated with the Affordable Care Act meet the potential for bridging disparities.

To learn more about these projects, visit http://www.census.gov/hhes/www/hlthins/

Posted in Uncategorized | Leave a comment

Housing Crisis and Family Well-Being: Examining the Effects of Foreclosure on Families

Bookmark and Share

Note: Census Bureau experts are presenting on a variety of topics at the Population Association of America annual conference. Follow the Research Matters blog or visit the press kit to learn more about their work.

Written by Laryssa Mykyta, Statistician, Social, Economic, and Housing Statistics Division

After the housing bubble popped in the mid-2000s, foreclosure rates increased fivefold. Many families had trouble paying their mortgages and faced losing their homes to foreclosure.

While we have information about the characteristics of families who lost their homes to foreclosure from earlier studies, we do not have much information about what happens to families throughout the foreclosure process or after losing their homes.

In this paper, I use a unique data set linking the 2008 panel of the Survey of Income and Program Participation (covering the time period May 2008 through November 2013) with foreclosure event data for 2005 through 2011 from RealtyTrac, a company that maintains a database of foreclosure events based on government records. Using these data, I look at how families at risk of losing their homes or who lost their homes have fared.

I examine how foreclosure affects family well-being, including family income, use of government assistance programs, doubling up or sharing a household, food insecurity, and support from others. Not surprisingly, families experiencing foreclosure had a harder time paying their mortgage or other bills than families who did not experience foreclosure. Families facing foreclosure also saw their earnings fall more than those families who did not experience foreclosure, suggesting that losing a job could trigger and accelerate housing hardship.

Families at risk of losing their home were more likely to turn to government assistance programs for support than other families. Families experiencing foreclosure were also more likely to double up or share their home.

On average, families that experienced foreclosure received less support from family and friends to pay housing costs than other families, however upon receiving a notice of foreclosure the likelihood of receiving support for housing costs from family or friends increased.

The stage of foreclosure also affects well-being.

Families that defaulted on mortgage payments had a harder time meeting their expenses, including housing costs, than other families. However, families in default were less likely to receive help in paying their mortgage, even from family or friends.

Families with homes listed for sheriff’s sale were more likely to double up than other families, by either sharing their home or moving in with others. In an effort to slow down or prevent foreclosure, some families doubled up.

Families who lost their homes to foreclosure had lower earnings than other families. These families were also more likely to access public safety net programs and help from sources other than family or friends, suggesting that they had fewer of their own resources to save their home or find a new home.

In general, families facing foreclosure were worse off than their counterparts, and experienced declines in well-being in terms of income, the ability to meet their expenses, and support from family and friends. You can read more about how foreclosure affects family well-being in the paper.

Note: Initiated in 1983, the Survey of Income and Program Participation provides a wealth of information to analyze the economic situation of people in the United States. It offers detailed information on cash and noncash income, while also collecting data on taxes, assets, liabilities and participation in government transfer programs. The data allow the government to evaluate the effectiveness of federal, state and local programs.

Posted in Uncategorized | Leave a comment

Visit Us at the 2015 Population Association of America Meetings in San Diego

Bookmark and Share

U.S. Census Bureau

Along with over a thousand demographers, sociologists and other professionals, Census Bureau staff will be participating in the Population Association of America (PAA) annual meetings in San Diego. The meetings, taking place from April 30 through May 2, allows researchers from across the country and different disciplines to present testing and research results on many interesting and noteworthy topics.

This year, Census Bureau staffers will present research findings on a spectrum of topics, such as:

  • Examining health insurance coverage inequalities and showing potential impact of the Affordable Care Act.
  • Evaluating the risk of health insurance loss between 2008 and 2011.
  • Improving estimates of same-sex married couples.
  • Changing methods for estimating foreign-born emigration.
  • Examining how income inequality of men and women persist over time.
  • Measuring supplemental poverty and its impact on programs and policies at the state level.
  • Examining the effect of housing foreclosure on families.
  • Analyzing how age of children affects women’s earnings.
  • Assessing the quality of S. vital statistics at the county level using the sex ratio at birth.
  • Changing risks of living alone or as a boarder since 1850.

The Population Association of America offers a forum for Census Bureau staffers to present their research for professional discussion. It is a major setting for ensuring that Census Bureau research and testing protocols remain relevant.

Attendees present and hear about advances in demographic population projection and estimation procedures, changing family dynamics and new advances in statistical sampling, estimation and modeling. PAA offers professional development courses, career placement services, and opportunities to meet and collaborate with individuals conducting similar research. For example, there will be a workshop demonstrating how to use the IPUMS-Current Population Survey and the Integrated Health Interview Series. The IPUMS-Current Population Survey currently includes the March Annual Social and Economic Supplement data from 1962 to 2014 and Current Population Survey basic monthly samples from 1989 to 2013. In addition to the basic monthly data, 13 supplements, including food security, veterans, fertility, tobacco use and voter surveys, are available.

On Wednesday, April 29, the Census Bureau, the Population Association of America’s Committee on Population Statistics and the Triangle Census Research Network sponsored a workshop on the Survey of Income and Program Participation. It provided an introduction to the survey’s 2014 redesign and included demonstrations on ways to access and use these data.

The annual conference also features training sessions on data from the decennial census and the American Community Survey. We look forward to sharing ideas with you at this year’s conference. For a listing of Census Bureau presentations, see census.gov.

A copy of the final program is available at http://paa2015.princeton.edu.

Follow the Research Matters blog this week for a closer look at many of the research topics presented at the Population Association of America’s annual meetings.

Posted in Uncategorized | Leave a comment

Using the Survey of Income and Program Participation to Study Residential Mobility

Bookmark and Share

Written By: Peter Mateyka, Analyst, Journey to Work and Migration Branch

Considerable research has been devoted to understanding why households move. The generally accepted model of residential mobility suggests that moving is a process that begins with a mismatch between current housing and housing needs and aspirations. This mismatch leads to residential dissatisfaction, which in turn is associated with the desire to move and an actual move. However, despite wide acceptance of this model, our research and outside empirical tests have provided mixed results. In other words, households frequently do not move, even after expressing residential dissatisfaction and the desire to move.

The main difficulty with studying residential mobility is that few data sets combine longitudinal data on moves with measures of residential satisfaction and detailed characteristics of respondents and their residences. In a new report, titled Desire to Move and Residential Mobility: 2010 to 2011, I demonstrate how the Survey of Income and Program Participation, a current source of longitudinal data on migration, can be used to study residential mobility.

I combine longitudinal data on the demographic and socio-economic characteristics of householders, including demographic events and changes in job status, with topical module data (asked at only one or two time points) on residential satisfaction, self-reported home equity, and disability status to study residential mobility. I also utilize internal data on respondents’ census track of residence to look at relationships between neighborhood characteristics and residential mobility.

I first looked at how many people desire to move because of dissatisfaction with their current residence and the relationships between the various sets of predictor variables and householders’ reports of desiring to move. I found that in 2010, nearly one in 10 American households (9.6 percent) reported that they were dissatisfied enough with their current housing, neighborhood, local safety or public services that they desired to move.

My research observed many expected relationships between the demographic and socio-economic predictors. Those who were younger and had lower income were more likely to desire to move. Renters were also more likely to desire to move than homeowners were.

Thanks to the richness of survey data, I was able to move beyond exploring relationships between basic demographics and desire to move. I found that demographic events, including having a child and getting married during the prior year, were associated with desiring to move. Disability status, in particular having a mental disability, was associated with desiring to move, as were neighborhood poverty levels and racial composition.


I next used the longitudinal design of the survey to see how many householders who reported desiring to move actually moved in the following year. I found that the majority of those who desired to move did not move within the next year, but their rate of moving was higher than that of the general population (18.3 percent compared with 9.6 percent). Consistent with past research on tenure status and mobility, renters who desired to move, moved at significantly higher rates than homeowners who desired to move. However, there was also some evidence the relationship between desiring to move and moving was stronger for homeowners. Homeowners who desired to move were almost twice as likely to move as the average homeowner (8.1 percent vs. 4.1 percent), but renters who desired to move were only about 1.2 times as likely to move as the average renter (25.4 percent vs. 20.8 percent). Also noteworthy, only 89,000 (5.3 percent) of the roughly 1.7 million homeowners age 55 and older who reported desiring to move actually moved in the following year.


In the final portion of the report, I focused on how householders’ reports of residential satisfaction and desiring to move changed from 2010 to 2011. On average, householders who desired to move and moved from 2010 to 2011 reported greater residential satisfaction at their new residences compared with their old. However, the results also made it clear why a report of desiring to move may not always lead to a move, as many householders who desired to move in 2010 and did not move no longer reported desiring to move in 2011. This suggests another possible answer for why many empirical studies find that many people who report the desire or intention to move do not move; they change their minds. The reasons for these changes are for future analyses to untangle. However, this does not necessarily mean that these householders suddenly became happier with their residential circumstances. They may still be dissatisfied, just not to the extent that they desire to move.

Posted in Uncategorized | Leave a comment

Synthetic Data: Public-Use Micro Data for a Big Data World

Bookmark and Share

Written by: Ron S. Jarmin, Assistant Director, Research and Methodology Directorate
Thomas A. Louis, Associate Director, Research and Methodology Directorate
Javier Miranda, Principal Economist, Center for Economic Studies

Businesses, households and policymakers need timely and accurate data to make informed decisions. National statistical offices around the world have a wealth of information from survey and administrative sources to meet these needs. However, they are constrained in their ability to release these data because of the confidentiality pledge to data respondents.

Synthetic data offer a way to expand the amount of information that national statistical offices can publically release while maintaining respondent confidentiality. In synthetic datasets, some or all data values are simulated (synthesized) using statistical models designed to mimic the (joint) distributions of the underlying data.

Researchers at the Census Bureau, in partnership with academic economists and statisticians through the Census Bureau’s secure research data centers, recently produced two synthetic public micro datasets. The SIPP-Synthetic Beta product combines survey data from the Survey of Income and Program Participation with administrative records from the Internal Revenue Service and the Social Security Administration (see Benedetto, Stinson and Abowd 2013). The Synthetic Longitudinal Business Database is the first business establishment-level public-use micro dataset made available by a U.S. statistical agency (see Kinney et. al. 2011).

Research findings on the development and use of synthetic data and future usage of these data were presented in a session of the World Statistical Congress in August 2013 held in Hong Kong. These articles are accessible in the Statistical Journal of the International Society of Official Statistics.

While synthetic data are exciting and hold great promise, there are challenges to expanding their development and use. Creating synthetic data requires significant technical expertise that is not widely available within many statistical agencies. Census Bureau progress on synthetic data has relied on robust collaboration with academic experts. Users also confront challenges. Synthetic microdata are still experimental and not as straightforward to use as conventional microdata. Because users may not understand what is involved in developing apps and online tools constructed using synthetic data, such as OnTheMap, they may understate the variance of estimates supplied by such tools.

Synthetic data are one way for national statistical organizations to take the lead in making high quality and reliable official statistics more accessible and relevant. However, creating and supporting synthetic data requires staffing and resources beyond what are generally available to them. The Census Bureau’s “two-way-street” strategy of developing partnerships with academic and funding institutions offers a way to move forward.

Posted in Uncategorized | Leave a comment