Race, Wealth and Incarceration: Results from the National Longitudinal Survey of Youth
Using the 1979 cohort of the National Longitudinal Study of Youth to explore the interwoven links between race, wealth and incarceration, this study examines the data on race and wealth status before and after incarceration. Data indicate that although higher levels of wealth were associated with lower rates of incarceration, the likelihood of future incarceration still was higher for blacks at every level of wealth compared to the white likelihood, as well as the Hispanic likelihood, which fell below the white likelihood for some levels of wealth. Further, we find that racial wealth gaps existed among those who would be incarcerated in the future and also among the previously incarcerated.
KeywordsRace Wealth Net worth Incarceration Imprisonment National Longitudinal Survey of Youth NLSY79
Department of Justice estimates indicate that one in three black men born in 2001 can expect to be incarcerated in the course of their lifetimes. In comparison, the rate for Hispanic males is one in six and for white males, one in seventeen (Bonczar 2003). The rate of incarceration in the USA quadrupled over the past three decades, contributing to the oft cited statistic that the USA has 5 % of the global population but holds 25 % of the world’s prisoners (Glaze and Parks 2014; Western and Pettit 2010; Mauer and King 2007).
Mass incarceration in the USA disproportionately affects African Americans, and Michelle Alexander (2012) has argued that imprisonment, de facto, is America’s new system of segregation. Pettit and Western (2004) estimated that “among men born between 1965 and 1969, 3 % of whites and 20 % of blacks had served time in prison by their early thirties.” The mark of incarceration has become a normal part of life for many black men, particularly, but not exclusively, among those living in impoverished neighborhoods (Freeman 1996; Irwin and Austin 1994).
Concurrently, blacks and Hispanics on average possess only a fraction of the net worth of whites (Tippett et al. 2014; Taylor et al. 2011; Hamilton and Darity 2009). Hamilton and Darity (2009) have argued that the racial wealth gap is the most acute indicator of racial inequality. Furthermore, racial discrimination in the justice system (Moore and Elkavich 2008; Petersilia 1983; Zatz 1987) compounds the wealth disadvantage that blacks and Hispanics already face. For example, Petersilia (1983) found that blacks and Hispanics convicted of felonies were more likely than whites to receive prison sentences. Even within the same groups, discrimination is harsher toward those with more Afrocentric features (Blair et al. 2004). Within the criminal justice system, even death-sentencing outcomes also have a key discriminatory dimension (Eberhardt et al. 2006; Unah 2011). With wealth and race being intricately linked, and race and incarceration clearly connected, we explore how racial wealth disparity relates to the racial incarceration disparity.
As early as birth, wealth can influence the likelihood of incarceration. Growing up with less family wealth means living in poorer neighborhoods with lower-quality education (Nguyen-Hoang and Yinger 2011) and a greater exposure to high “street” crime and high imprisonment areas (Pope and Pope 2012; Leventhal and Brooks-Gunn 2011). The presence of illegal drug trafficking, youth gangs and neighbors who have experienced high incarceration rates increases the likelihood of contact with the criminal justice enforcement system.
Lack of family wealth also could mean being unable to afford additional education and delaying entering the labor market, leading to higher risks of incarceration, risks that are highly stratified by education (Pettit and Western 2004). Once involved with the criminal justice system as a defendant, low family wealth means being unable to afford the resources to navigate the criminal justice system via high-quality lawyers or the capacity to post bond. Thus, a priori, low wealth can increase the likelihood of being incarcerated.
Imprisonment, in turn, can depress wealth accumulation through a variety factors. Involvement with the system of criminal justice increases legal debt (Harris et al. 2010), and incarceration means loss of income. During incarceration, being unable to make payments on mortgages or other debts can lead to an accumulation of interest obligations and penalties as well as a grossly diminished credit status. Incarceration also means household instability, placing an additional burden on the capacity to build assets (Zagorsky 2005).
A record of previous incarceration also has wide-ranging immediate and future consequences that affect asset-building capability over the lifetime. Such a record acts as a barrier to employment, thereby lowering earnings (Apel and Sweeten 2010; Pager 2003). Furthermore, the criminal “credential” (Pager 2007) associated with incarceration impedes wealth accumulation through stigmatization, poor credit access and lack of access to supportive social institutions (Brayne 2014; Maroto 2014; Wildeman 2014; Goffman 2009). With 1 out of 35 adults in the USA having been incarcerated previously, the effect on wealth outcomes is a major social concern.
Although the “mark” of incarceration has a negative impact on all, the impact of previous incarceration is not equivalent across all groups. For example, Pager (2003) uses evidence from a set of field experiments to demonstrate that when presented with a black male job applicant who reported prior incarceration, employers were twice as likely not to offer the applicant a call back for a job interview than a similarly qualified white male applicant who also reported prior incarceration. Moreover, the likelihood of a call back actually was slightly higher for white males reporting previous incarceration than black males reporting no incarceration record whatsoever.
Disparate patterns of incarceration across races also may exacerbate racial inequality. Indeed, Schneider and Turney (2015) found that higher incarceration rates depress black homeownership rates and widen black-white disparities in homeownership. Prior research has shown that previous incarceration is associated with substantially lower levels of wealth (Maroto 2014), estimating that respondents who had previously been incarcerated had an average of $42,000 less wealth (in 2010 dollars) than their peers who never had experienced incarceration. Although Maroto (2014) controlled for race in her model, the analysis did not examine differential effects due to race. In this paper, we explicitly look at racial wealth differences between those previously incarcerated and never incarcerated. Additionally, since the expectation is that low wealth increases the likelihood of being incarcerated, we hypothesize that those who experience incarceration had less wealth at the baseline and that the inverse also should hold; that is, those with less wealth at the baseline collectively experience higher rates of incarceration.
Using the National Longitudinal Survey of Youth 1979 (NLSY79), we present descriptive statistics on personal net worth stratified by race and incarceration status. We investigate the collective wealth positions of those who experience incarceration in the future (future “incarcerees”) compared to those who do not and separately calculate the likelihood of incarceration at various levels of initial wealth. Using the latest NLSY79 data—2012 wealth positions (Bureau of Labor Statistics, U.S. Department of Labor 2012), we examine the recent status of those previously and never incarcerated, in relation to race and other possible factors. In the “Discussion” section that concludes the paper, we speculate about reasons for the patterns of racial wealth trajectories with respect to incarceration, discuss limitations of our study and consider implications as well as future directions for investigation.
The NLSY79 is a large national panel dataset that contains both incarceration and prior wealth data at an individual or family level, enabling us to examine the net worth positions of individuals before and after incarceration. An ongoing longitudinal stratified probability survey, the NLSY79 contains a nationally representative cross-sectional group. The survey also oversamples respondents from economically deprived groups, called the supplemental subsample.
The survey cohort consists of 12,686 young men and women who were 14–22 years old when they were first interviewed in 1979. The respondents were interviewed annually from 1979 until 1994 and biennially thereafter. The NLSY79 had retention rates, excluding the deceased, exceeding 90 % for 1979–1994, exceeding 80 % for 1996–2010 and just below 80 % for 2012.
Over the years, there have been several changes in the eligibility criteria for the subsamples. Because the majority of the 1280 participants in the military supplement were dropped by the 1985 survey, the total number of eligible respondents at the beginning of the 1985 survey was 11,607. Additionally, after the 1990 survey, “non-black non-Hispanic” members of the supplemental subsample were no longer eligible, bringing the total to 9964 in 1991.
For this study, respondents were stratified racially based on their answers to questions about their only or primary racial and ethnic origins.1 Among those placed in the aggregate categories of American Indian, Asian/Pacific Islander, black, Hispanic, white, other and none, there are too few observations for analyses on group-specific levels of wealth except for black, Hispanic and white respondents. Therefore, we restrict our analyses in this study to those three groups.
Beginning in 1985, when all respondents had turned 20, the NLSY79 reported values of assets and debts. The family data collected in 1985 included the value of an owned home or apartment, monetary assets, farm, business, and real estate properties, vehicles, other owned items valued at more than $1000, and debt owed. In 1990, the wealth data also included the value of common stock and bonds, rights to estates and trusts, and the amount the respondent would have left if debts were paid off. Starting from 1994, the amount of money in IRA, 401Ks and CDs also were included in the data.
Using these recorded data, survey investigators in 2008 publically released respondent net worth variables, produced by summing all asset values and subtracting debts in each year, with missing asset and debt values imputed, providing an estimate of the wealth of each respondent. Although the estimate is generally of family net worth, since a respondent would include the home or assets co-owned with a spouse, the estimate would not be expected to include parental assets. Thus, a respondent’s net worth as reported would be individual wealth if unmarried and family wealth if married. For this study, all reported net worth values are adjusted to 2012 dollars.
Distributions of net worth by race and incarceration experience, 1985
# of Obs
Never previously incarcerated—black, Hispanic and white
Will be incarcerated 1986–1990
All incarcerated 1986–1990
Will be incarcerated 1991–2012
All incarcerated 1986–2008
Never incarcerated (to 2012)
Incarceration information was obtained in a somewhat indirect fashion. Survey participants only were asked extensive questions related to crime and sentencing history in 1980. However, the NLSY recorded the respondent’s type of residence at the time of each survey, including whether it was a jail or prison. In addition, being incarcerated as a “Reason for Non-Interview” was coded distinctly starting from 2004. Combining these three measures, we constructed a variable of whether a respondent was ever incarcerated. Because of the nature of these data, incarceration may be understated and likelihoods underestimated, since incarceration between surveys was not captured.
Incarceration for both genders are captured in the survey and analyzed in this paper. Although there are differences in incarceration outcomes by gender, patterns of wealth relationships between groups are essentially the same for both males and females. Females, who comprise approximately 10 % of those who were incarcerated by 1990 and 11 % of those who were incarcerated by 2012, generally had less wealth. Examining only males or females using the same analytical frame in this study leads to slightly higher wealth for males in most groups, but the net worth positions by race and incarceration outcomes are similar for the two genders. However, female outcomes for future incarceration projected from wealth prior to incarceration by race were qualitatively different from male outcomes in all cases where we had adequate numbers to draw conclusions. For these cases, we only present gender-specific results.2
In order to assess the wealth positions of the various groups demarcated by race and incarceration experience, we use wealth at the quartiles and at the 90th percentile for each group. Given the skewness in wealth distributions and the use of top-codes to protect the anonymity of respondents, we rely primarily on quartiles to compare group differences in wealth.3
In the following sections, we present the results split into three categories: wealth before incarceration, wealth in 1985 and likelihood of future incarceration and wealth after incarceration.
Wealth Before Incarceration
By 1985, when NLSY79 wealth data were first collected, 426 respondents, or 3.67 % of all eligible respondents, previously had been incarcerated. In order to examine wealth positions before the impact of incarceration, we limit our analyses in this section to only those who had not been incarcerated at that time. In Table 1, we report the distributions of NLSY respondent net worth in 1985 by race and incarceration experience. For each group, distributions are assessed by wealth values at the quartiles and at the 90th percentile. The median wealth of all blacks, Hispanics and whites with no known history of imprisonment was $7455 in 1985.
Stratifying by race, we separate incarceration experience into three categories: incarceration in the next 5 years (before a supplemental subsample was dropped), incarceration only in the next 6–27 years (maximum data) and never incarcerated. Since all respondents were in their 20 s in 1985, the next 5 years represent not only incarceration in the short term but also incarceration for the respondents in their young adulthood. The second group of future incarcerees, those who were incarcerated in the interval 1991–2012, would have experienced incarceration for the first time in the transition from young adulthood to middle age.
For each subgroup, we also assess the collective wealth distribution through the values at the quartiles and 90th percentile. Compared to their racial peers who never report having been incarcerated, future incarcerees had less wealth in 1985 at the median and at the 75th percentile. Those who faced incarceration sooner, within 5 years, also had less wealth at the median and at the 75th percentile, and at the 90th percentile compared with those who were incarcerated in the later period, 1991–2012. But these relationships do not hold across racial lines. For example, white eventual incarcerees had slightly more median wealth ($1491) than blacks who never were incarcerated ($1385).
Not all future incarcerees were economically disadvantaged in 1985. Of the whites who would be incarcerated in the far future, wealth at the 90th percentile was actually the top-code for the year, meaning that several white future incarcerees probably were in the top two percentiles of wealth at the time. Similarly in other years, although it was rare, future incarcerees from all three races were found with top-coded wealth before incarceration.4
After establishing differences in wealth between the groups even among future incarcerees, we stratify by race to determine whether future incarcerees started with lower net worth. Figure 1 displays the median wealth positions for males in each of the three groups, in every year in which wealth was recorded in the NLSY79 up to the year 2000.5 Although female future incarcerees similarly had lower net worth before incarceration compared to their never-incarcerated peers, the number of observations was too few for analysis. Males who would experience future incarceration consistently had lower median wealth, and this relationship remains true when we stratify by race. However, this does not hold true across racial lines: white future prisoners in some years had more wealth at the median than blacks who would never experience incarceration. For example, in 1985 and 1989, white male future incarcerees had $3621 and $2590 more net worth at the medians, respectively, compared against black males who would never be incarcerated.
Similar gaps also persisted at the 75th percentile of wealth when examining only males or males and females combined, with future incarcerees again consistently having lower wealth. At the 25th percentile, where wealth was typically close to zero or negative, future incarceree wealth was equal to or lower than that of their counterparts. However, again, there were disparities by race. In some cases, white future incarcerees had more wealth at the 25th and 75th percentile compared with blacks and Hispanics who never were incarcerated. Of course, in general, the evidence indicates that those who faced incarceration in the future began with lower wealth positions than those who would not.
Wealth Position in 1985 and Likelihood of Future Incarceration
Since those who do experience incarceration came from backgrounds with lower levels of wealth, we examine whether racial incarceration rate disparities can be explained by racial wealth disparities. To calculate the risk or likelihood of future incarceration in relation to wealth, we start with the wealth positions of a sample of persons who have never been incarcerated at the baseline. At the time of the 1985 survey, about 5 % of blacks, 4 % of Hispanics and 2.5 % of whites in the NLSY79 had experienced incarceration.
Selecting those who were not incarcerated by 1985 and their wealth levels as the starting point, we calculate the likelihood of incarceration by race within five and within 27 years. Using wealth at the deciles for blacks, Hispanics and whites combined, wealth levels in 1985 are split into ten groups. Additionally, because the rates of future incarceration were disparate between the genders, we analyze males and females separately. Likelihoods for each racial and wealth subgroup are equal to the percent of the respondents who were found incarcerated in future survey years. The number of observations for each race and wealth bracket is included in Appendix Table 5 tabulations to indicate robustness of the likelihoods.
The 27-year incarceration likelihoods for males, illustrated in Fig. 2b, which exclude the NLSY79 economically disadvantaged supplement dropped in 1990, present longer-term incarceration odds, including through the cohort’s middle age years. Black males had the highest rate of incarceration at all levels of wealth, and Hispanics had the second highest at most levels of wealth. The long-term black and Hispanic likelihoods decrease for those in the middle deciles, compared to those who had little or no wealth in 1985.
It is important to note that in these 27-year likelihood estimates, economically disadvantaged blacks and Hispanics are overrepresented, and thus, the differences in likelihoods also may be affected by other related economic factors such as income. In the short- and long-term likelihoods for males, Hispanic rates of incarceration are similar to white rates at higher levels of wealth at the baseline, but black rates of incarceration remain higher than those of whites. Even so, it is clear that for males of all races, higher wealth is associated with lower likelihoods of future incarceration. Although racial disparities in incarceration seem to converge for males in the top decile of wealth, given the small sample sizes, this finding is inconclusive.
For females, however, we find the convergence in racial incarceration likelihoods more pronounced. In the 5-year outcomes, shown in Fig. 2c, short-term black and Hispanic female likelihoods were highest in the little-debt to zero wealth category—similar to males. However, after the 60th percentile of wealth, no females were found incarcerated in the short term, indicating a convergence in racial incarceration likelihoods.
In the 27-year likelihoods for females, presented in Fig. 2d, convergence in racial rates was also evident, with females in four of the five top deciles of the wealth distribution experiencing no incarceration in the future. The data suggest that for females, after a threshold level of wealth is reached, racial likelihoods of incarceration may be similar. But, differences between incarceration outcomes for the wealthy and less wealthy are undeniably disparate.
Although more pronounced for females, at higher levels of wealth, the likelihood of incarceration falls for all racial groups. Yet, for comparable levels of wealth in 1985, when the respondents were entering young adulthood, blacks, Hispanics and whites still had disparate likelihoods of incarceration. As a result, we find that wealth is relevant to the prospect of incarceration, particularly when stratified by race, and that racial incarceration disparities persist even for individuals with similarly situated family wealth positions.
Wealth After Incarceration
Once incarcerated, in the NLSY79 surveys, a number of respondents were actually interviewed while in jail or prison and a number of those respondents reported their assets information. Due to the nature of the setting, reported data may be irregular, not the least due to the difficulty of attempting to recall financial details about outside assets. Still, as one of the few sources of data for wealth during imprisonment, we examine the results separately. While some respondents reported on their wealth while being incarcerated only once, there were respondents who reported wealth while under incarceration in several of the survey years. For some incarcerees who were interviewed for consecutive surveys, reported wealth levels did fluctuate. As expected, those in jail or prison typically had few or no assets, and their wealth levels were markedly low.
Compared with the cross-sectional median wealth of the entire NLSY cohort, incarcerees reported far less wealth, with a majority of the incarcerees in every year reporting zero or negative net worth. In later years, when the overall NLSY79 cohort became middle aged, they also had accumulated more wealth, whereas respondents under incarceration had zero or close-to-zero wealth. Among all of the concurrently incarcerated, with such low wealth reported each year, we observed no consistent racial wealth differences. In some years, relative racial wealth positions were reversed—that is, black incarcerees at the 50th or 75th percentile had the highest wealth compared to their Hispanic or white incarceree counterparts. In 1985, black incarcerees had $1278 at the median, whereas Hispanics and whites had $0; in 1990, whites interviewed in jail reported $933 at the 75th percentile, compared to their black counterparts who reported $3325. In most years, the median wealth for all races was at parity at zero. As such, observations from the NLSY79 suggest that wealth is reduced to little or nothing for all persons while in jail or prison, thereby eliminating racial differences in wealth during the period of incarceration.
Distribution of wealth of previously and never incarcerated
# of Obs
Previously incarcerated 1985
All previously incarcerated 1985
Never incarcerated 1985
All never incarcerated 1985
Previously incarcerated 2000
All previously incarcerated 2000
Never incarcerated 2000
All never incarcerated 2000
Previously incarcerated 2012
All previously incarcerated 2012
Never incarcerated 2012
All never incarcerated 2012
For instance, in 1985 the median wealth of white previous incarcerees was slightly higher than the median wealth of blacks who had never been incarcerated before. This phenomenon was observed in several years—1989 ($6845 vs $3515), 1996 ($9052 vs $7008) and 1998 ($11,280 vs $8460). Among the previously incarcerated, at each wealth quartile, blacks consistently reported having less wealth than Hispanics, while whites reported more wealth than Hispanics in most years.
Post-incarceration wealth by time of last incarceration and race, 2012
# of Obs
Last known incarceration 1991–2008 (4+ years ago)
Last known incarceration <1990 (22+ years ago)
Those who were incarcerated recently consistently had less wealth than those who were incarcerated more than two decades ago. Specifically, within the same race, those who were more recently incarcerated had less wealth at the median, the 75th percentile and the 90th percentile. Although the 80s incarcerees were clearly better off than the later incarcerated group, they were still worse off than those who were never been incarcerated. Black respondents in this group had the lowest median wealth compared to both white respondents, who had over three times as much, and Hispanic respondents—who were actually doing better than their white counterparts. Although the years since incarceration may have helped wealth trajectories recover, the previously incarcerated remain a disadvantaged group and have a disproportionately low share of wealth.
Our post-incarceration findings are consistent with previous research (especially Maroto 2014): A previous record of incarceration is associated with substantially lower wealth levels, and previously incarcerated blacks have significantly less wealth than previously incarcerated whites. Additionally, as the data in the NLSY79 demonstrate, low wealth is associated with an increase in the likelihood of incarceration, which in turn can depress wealth accumulation. Conversely, future incarcerees had less wealth at the baseline than those who would never be incarcerated. Although we do find that this is not true across races; in some years blacks who never had been incarcerated had less median wealth than whites who had been incarcerated. This may be because blacks and Hispanics generally have low wealth in comparison with whites (Tippett et al. 2014; Taylor et al. 2011; Hamilton and Darity 2009; Gittleman and Wolff 2004), or because whites are more dominant in terms of wealth such that even future incarcerees possessed more wealth to start with.
Separating by race, we showed that blacks had lower odds of incarceration in the short term, in their early twenties to thirties, when they possessed more wealth. The odds of such incarceration for black males who possessed as little as $2000 in 1985 were only half the odds for black males who possessed very little or no wealth and collectively had an incarceration rate of 10 % in the next 5 years. For black females, possessing more wealth meant virtually eliminating incarceration chances in the short term. In the long term as well as short term, having more wealth at the baseline was associated with lower incarceration rates in the future, for all races. Accordingly, we find that not only is wealth relevant to the prospect of incarceration, race and gender also interact with wealth prior to incarceration.
Between races, we find that at low levels of wealth both blacks and Hispanics still had a higher incarceration rate than whites. At higher levels of wealth at the baseline, although the black-white incarceration disparity was reduced for males, it was not eliminated. One explanation for the differential odds of incarceration between races may be that even while having similar wealth levels, individuals still may have disparate economic situations, through income, extended family wealth or differential exposure to discrimination. Personal and family human capital levels such as education, job experience and social connections also may differ greatly among those with similar wealth levels. Therefore, observed racial differences in male incarceration rates despite similar wealth levels may be explained once those factors are taken into account.
We do observe that for Hispanic males starting with higher levels of wealth, odds of incarceration are similar to those of white males at comparable wealth levels. Why Hispanic males experience this convergence but not black males, we leave to further study. For females, we find that after some point of wealth, the likelihood of incarceration changes little as wealth grows and is not different for the races. Our results suggest the same may be true for males at a much higher level of wealth, but additional data would be required for conclusive evidence, due to the lack of black male observations at the very high levels of net worth.
We acknowledge other limitations in our study that may be addressed by further study. Although we are able to examine the relationship between incarceration and personal wealth, we note that impact on extended family wealth is obviously salient, but we find insufficient data to examine this issue fully. In some cases, data regarding females with respect to wealth and incarceration were also insufficient, limiting our analyses, in these cases, only to males. In addition, we do not find consistent results of the impact of different lengths of time spent incarcerated, but increased length has been reported to be associated with lower wealth (Maroto 2014).
The reasons for which respondents were incarcerated also may be salient, but were not available in the NLSY79 data. These limitations invite further study through the collection and use of additional data sources, particularly for unpacking the economic impact of incarceration on the broader household as well as for studying the impact of disparate, and possibly racially relevant, reasons for incarceration.
With disparity strongly pronounced between blacks, Hispanics and whites with respect to wealth and incarceration, we have identified salient relationships to improve our understanding of the interacting forces. Additionally, considering the interaction of race with wealth and incarceration makes clear the disparities we present in Fig. 1, establishing that those who experience incarceration come from a lower wealth background compared to those who do not. Together, the results suggest that when it comes to wealth and incarceration outcomes, the disadvantages of being black or Hispanic compound the disadvantages of asset poverty.
NLSY79 respondents were separated into three major racial/ethnic groups by survey screeners: Hispanic, black and non-black, non-Hispanic. The “non-black, non-Hispanic” group includes American Indians, Asians, Asian Indians, Pacific Islanders and other non-white identities, including those who self-identified as black or Hispanic (Light and Nandi 2007).
For all results, male-only, female-only or combined estimates are available by request from the authors.
To protect the identities of those in the survey, wealth data in the NLSY79 were top-coded so that values in the top two percentiles would be assigned a capped value for each year. Until 1994, asset values were capped and top-coded, whereas afterward, the top two percent of valid values for net worth were averaged, and that average replaced all values in the top range. Although this change would allow for valid estimations of mean wealth for the total population, no distinctions were made based on race. With uneven numbers of observations from each racial group top-coded, and the true values lost, calculations of racial mean wealth based on the data would not be valid. However, racial wealth data at the quintiles and quartiles, including the median, were unaffected by this top-coding; thus, we examine group wealth distributions through those measures.
Although, of course, since we cannot know the nature of the crime, it may have been the criminal activity that led to the wealth accumulation and later imprisonment.
Although the gaps and trends did persist, data for 2004 and 2008 are omitted as the # of observations in some categories were too few.