Measuring the economic status of low-income individuals and families is a central focus of poverty scholars and is at the fore of much public policy debate. The stakes are substantial as changes in poverty (and poverty thresholds) influence the scale and scope of redistributive tax and transfer programs at all levels of government. In the United States, official poverty statistics are derived from the Annual Social and Economic Supplement to the Current Population Survey, a nationally representative survey of about 90,000 households. The main source of household income comes from labor-market earnings, but as highlighted in the work of UKCPR Affiliates Charles Hokeyem, Christopher Bollinger, and James Ziliak (DP2014-05), published in the Journal of the American Statistical Association, a challenge to the proper measurement of income is survey non-response. Using a restricted-access dataset that links the CPS ASEC to Social Security's Detailed Earnings Records, they show that survey non-response is not random and leads to a systematic downward bias in the official poverty rate of about 10% in an average year. Follow-up research that is forthcoming in 2019 at the Journal of Political Economy (DP2015-02), and conducted jointly with Barry Hirsch at Georgia State University, shows that earnings non-response affects broader measures of the income distribution, such as earnings inequality and gender wage gaps.


Poverty, race, and the contexts of achievement: Examining educational experiences of children in the American South

This paper reports findings of a study examining child-, classroom-, and school-level factors that effect academic achievement among public school children in the South. Using ECLS-K data, we compare and contrast the learning environments in high/low minority and high/low poverty schools. A sizeable minority of Southern children attend schools that are race and/or class segregated; on multiple dimensions these schools are less desirable than are schools attended by more privileged children, and children attending these schools have lower levels of academic achievement. Results from 3-level random intercepts models show that a range of child and family factors, as well as classrooms with less experienced teachers and with more low-level readers, and rural school location all contribute to lower reading gains during first grade. We find no “race effects” on achievement, net of other variables. Issues of “selection” are discussed, and implications for social work are explored.


Poverty rates of refugees and immigrants

Refugees are typically poorer than other immigrants and native born, although the average difference is smal, with changes in the unemployment rate explaining most of the difference in poverty rates. In times of recession, or in areas with particularly high unemployment rates, refugees will fare worse, perhaps due to concentrations of refugees in industries with higher cyclical variation in unemployment. We also find that refugees’ poverty rates start higher but fall more rapidly with time, suggesting that refugees assimilate more rapidly than other immigrants. We also find marked differences in the poverty rates of immigrants and refugees across regions. The evidence on poverty rates suggests that some regions are better than others at filling the poverty gap for immigrants than refugees.

Relative prices and substitution across wage, welfare, and disability income

In this paper I exploit the fact that the social and economic reforms over the past two decades differentially affected the opportunity costs of non-participation in work, welfare, and disability programs for single mothers across different birth-year and education cohorts. This cohort variation in after-tax wages and transfer benefits is used to identify own- and cross-price elasticities of demand for and substitution across wage, welfare, and disability income over 1979 to 2001 in the Current Population Survey. To estimate these key parameters I model household preferences with a conditional Almost Ideal Demand System that admits corner solutions, nonseparability, endogenous wages and incomes, and latent heterogeneity via cohort and state fixed effects. I match individual and family-level data in the CPS both with family-specific federal, state, and payroll tax rates, and with state-specific and time-varying benefit levels and effective tax rates in the AFDC and SSI programs. Using a two-limit Tobit instrumental variables estimator I find strong evidence of sizable own and cross-programmatic substitution effects. For example, the estimated elasticities imply that between 1979 and 1999 the increase in the generosity of SSI relative to AFDC accounts for about 40 percent of the average growth in SSI, while the increase in real wages accounts for about one-half of the average decline in AFDC shares over the past two decades. Simulations suggest that changes in relative after-tax wages and transfer-program benefits over the past two decades lead to a substantial “pull” out of cash welfare and into expanded reliance on employment and disability as a means of financial support among single mothers.

Does regional variation in multiple measures of health status differ across income levels?

This study examines whether regional variations in health status measures are consistent across the income gradient, or whether they are more pronounced at the lowest income levels. We use data from the Community Tracking Survey, a large randomized telephone survey of residents in 60 U.S. communities. Controlling for individual risk factors and county level income inequality, lowest income individuals have poorer scores on counts of chronic diseases, global health ratings, and the physical and mental components of the SF-12. Residents of the South have poorer scores on chronic disease counts, global health and physical health than residents of the Northeast, and poorer scores on physical and mental health than residents of the Midwest. Regional variations in the first three measures persist across the income gradient, and are more pronounced in the population group just above the poverty level. However, the lowest income group of residents of the South had poorer mental health scores than residents of all other regions, while the highest income group had better mental health scores than residents of all other regions. These findings suggest that a wide variety of community-level factors influence health status across the income gradient, while a separate set of community level factors may interact with income in communities to increase particularly mental health risks for a subset of the population.

The rise of low-skill immigration in the South

The 1990s witnessed a significant geographic redistribution of immigration away from the traditional immigrant-receiving states, mainly California, and towards other parts of the country, mainly the Southern states that have not historically been immigrant-receiving states. This paper documents the impact of this change in immigrant settlement patterns on the skill endowment of the workforce in Southern states. The empirical analysis indicates that the recent change in immigrant settlement patterns led to the rise of a sizable foreign-born low-skill workforce in the South, particularly outside Florida and Texas. This workforce developed both as a result of increased settlement of many newly arrived low-skill immigrants in those states, and increased internal migration of low-skill immigrants from the non-South to the South.

Prospects of agricultural entrepreneurship among resource limited farmers in the Central Appalachian tobacco belt

Agricultural entrepreneurship is receiving heightened attention as a potential means for economic revitalization of communities adversely affected by changes in the agricultural sector. In particular, resource limited farmers in the Appalachian region of the United States have been hit by major changes in the tobacco industry. Very little is known about resource limited farmers respond to changing industry conditions and policy attempts to remedy structural change. Recently, the Commonwealth of Kentucky has attempted to assist farmers in adopting new farmbased enterprises to expand their income base. However, it is unclear about the factors that drive entrepreneurial or diversification activities among resource limited farmers. In general, it is expected that resource limited farmers, most of whom work off-farm for a significant portion of their income, face a tradeoff between off-farm work constraints and potential new sources of income on-farm. This paper uses a survey of 765 farmers in Northeast Kentucky to explore factors correlating with agricultural entrepreneurship and understanding this tradeoff.


Filling the poverty gap, then and now

The extent to which means-tested transfers, social insurance, and tax credits fill the gap between family’s private resources and the poverty threshold is a periodic barometer of the social safety net. Using data on families from the Current Population Survey I examine how the level and composition of before- and after-tax and after-transfer poverty gaps changed in response to changes in the policy and economic landscapes over the past two decades. The estimates presented here indicate not only dramatic changes in the level and sources of income maintenance programs filling the poverty gap, but also dramatic changes in which demographic groups successfully fill the gap. From the peak-to-peak business-cycle years of 1979 to 1999, the fraction of the gap left unfilled among non-elderly families in poverty has expanded by 25 percent, while the unfilled gap has increased by 50 percent among single female-headed families, families headed by non-whites, and families residing in the Northeast. In a given year the poor in the South fill considerably less of the poverty gap with cash assistance, but make up for much of the shortfall with higher payments of food stamps, SSI, and SSDI. Over time the poor in all regions of the country have substituted SSI, SSDI, and the EITC for cash welfare. Indeed, by 1999 the unfilled gap for families with related children present would be one-fifth larger without the EITC. With the exception of married-couple families, this apparent rate of replacement of disability payments and tax credits for cash assistance is less than one for one, leaving most poor families, especially non-white families and single female-headed families, financially more vulnerable today than in previous decades.

The impact of welfare programs on poverty rates: Evidence from the American states

There is spirited debate between those who maintain that public assistance to the poor decreases poverty by raising their incomes (an income enhancement effect) and those who contend that welfare increases poverty by discouraging the poor from working (a work disincentive effect). Extant studies have been inconclusive because they have focused on the effect of welfare benefits on the poverty rate, but have not employed designs that allow researchers to sort out distinct income enhancement and work disincentive effects. We develop a model of poverty rates in the American states that permits estimation of these distinct effects – based on state-level time-series data observed annually for the years 1960-90 – and we find that welfare had both effects during our period of analysis. We also calculate the net impact on the poverty rate of an increase in welfare benefits (taking into account both income enhancement and work disincentives), and we conclude that it has varied across states and time. In general, however, the ability of marginal increases in welfare spending to reduce the poverty rate by enhancing incomes has declined since the 1970s.

Measuring underemployment at the county level

As labor markets tightened in the last half of the nineties, economic development and community leaders sought to identify more locally available workers than were indicated by published statistics. Using results from commissioned surveys, they pointed to large numbers of part-time workers who desired full-time work, and to full-time workers who were qualified for better jobs. These statistics were often used to negate low official unemployment rates that deterred firms, concerned by the ostensible shortage of workers, from locating in their counties. We have conducted a larger, statewide, survey of underemployment and linked it to the detailed demographic and labor force data from the 2000 Census. We used the results to identify variations in the number and type of underemployed persons around the state, with emphasis on the differences between urbanized and rural areas. Over a quarter of full-time workers reported underemployment, including a third of workers in exurban counties. However, forty to fifty percent of underemployment is reportedly by choice, with the highest rates in the small urban and exurban regions. Of those that are not underemployed by choice, over ninety percent of respondents in some regions cited lack of job opportunities. We find that between fourteen and forty percent of part-time workers prefer full-time work, with the highest rates in rural Appalachian counties. We provide some of the reasons underemployed people cite as constraints to better employment. Also, we used the survey results and the recent Census information to predict the number and type of underemployed persons in each county. The model can be used to update predictions as new local demographic and labor force estimates are released annually from the Census Bureau’s forthcoming American Community Surveys.

Obesity and the development of complications across the life span: Is there a relationship between obesity and poverty?

The purpose of this investigation was to examine the relationship between the development of obesity in children ages five to ten years, and poverty (the socio-economic status of the family). Because of the associated complications of obesity such as heart disease, stroke, diabetes and hypertension, this research aimed to determine if obesity, a precursor of these diseases, was related to poverty. The rate of the development of hypertension and diabetes in children and young adults has been steadily increasing over the past ten years (Hines, Fishman, Green, 1999). Therefore, there is an urgent need for continued investigation exploring the multiple variables associated with the development of this major health hazard. The goal of this research was to examine the following objectives: Investigate the relationship between poverty and obesity in children age 5 to 10 years, determine the genetic risk for subjects meeting the criteria for obesity, and communicate findings of this research via referred publications and presentations. The methods used in this research involved collection of secondary and primary data. Secondary data collection involved retrieving information on subjects assigned to the free lunch program, based on household income. Primary data collection included weight, height and Body Mass Index (BMI) of the subjects. Height, weight and BMI were recorded and compared with normal values for age and height, thus determining obesity status. The results of obesity status were correlated with household income. Thirty three subjects (N=33) were studied. The results of the study showed no significant relationship between obesity and poverty. However, the incidence of obesity in the sample according to body mass index was a finding that warrants further investigation. The results showed a higher level of obesity among the subjects who were not from households meeting the definition of poverty.