Showing papers in "Journal of Human Resources in 1998"

Recent Advances in Quantile Regression Models: A Practical Guideline for Empirical Research

Abstract: This paper provides a guideline for the practical use of the semi-parametric technique of quantile regression, concentrating on cross-section applications. It summarizes the most important issues in quantile regression applications and fills some gaps in the literature. The paper (a) presents several alternative estimators for the covariance matrix of the quantile regression estimates; (b) reviews the results for a sequence of quantile regression estimates; and (c) discusses testing procedures for homoskedasticity and symmetry of the error distribution. The various results in the literature are incorporated into the generalized method of moments frame-work. The paper also provides an empirical example using data from the Current Population Survey, raising several important issues relevant to empirical applications of quantile regression. The paper concludes with an extension to the censored quantile regression model.

Estimating Models with Sample Selection Bias: A Survey

Abstract: This paper surveys the available methods for estimating models with sample selection bias. I initially examine the fully parameterized model proposed by Heckman (1979) before investigating departures in two directions. First, I consider the relaxation of distributional assumptions. In doing so I present the available semi-parametric procedures. Second, I investigate the ability to tackle different selection rules generating the selection bias. Finally, I discuss how the estimation procedures applied in the cross-sectional case can be extended to panel data.

An Analysis of Sample Attrition in Panel Data: The Michigan Panel Study of Income Dynamics

Abstract: By 1989 the Michigan Panel Study on Income Dynamics (PSID) had experienced approximately 50 percent sample loss from cumulative attrition from its initial 1968 membership. We study the effect of this attrition on the unconditional distributions of several socioeconomic variables and on the estimates of several sets of regression coefficients. We provide a statistical framework for conducting tests for attrition bias that draws a sharp distinction between selection on unobservables and on observables and that shows that weighted least squares can generate consistent parameter estimates when selection is based on observables, even when they are endogenous. Our empirical analysis shows that attrition is highly selective and is concentrated among lower socioeconomic status individuals. We also show that attrition is concentrated among those with more unstable earnings, marriage, and migration histories. Nevertheless, we find that these variables explain very little of the attrition in the sample, and that the selection that occurs is moderated by regression-to-the-mean effects from selection on transitory components that fade over time. Consequently, despite the large amount of attrition, we find no strong evidence that attrition has seriously distorted the representativeness of the PSID through 1989, and considerable evidence that its cross-sectional representativeness has remained roughly intact.

Using sibling data to estimate the impact of neighborhoods on children's educational outcomes

Abstract: Studies that attempt to measure the impact of neighborhoods on children's outcomes are susceptible to bias because families choose where to live. As a result, the effect of family unobservables, such as the importance parents place on their children's welfare, and other unobservables that are common to geographically clustered households, may be mistakenly attributed to neighborhood influences. Previous studies that attempt to correct for this selection bias have used questionable instrumental variables. This paper introduces an approach based on the observation that the latent factors associated with neighborhood choice do not vary across siblings. Therefore, family residential changes provide a source of neighborhood background variation that is free of the family-specific heterogeneity biases associated with neighborhood selection. Using a sample of multichild families whose children are separated in age by at least three years, I estimate family fixed effect equations of children's educational outcomes. The fixed effect results suggest that the impact of neighborhoods may exist even when family-specific unobservables are controlled. This finding is robust to many changes to estimation techniques, outcome measures, variable definitions, and samples but is sensitive to the exact formulation of the neighborhood measure. Daniel Aaronson is a researcher at the Federal Reserve Bank of Chicago. He thanks Joe Altonji, Becky Blank, Greg Duncan, Judy Hellerstein, Sandy Jencks, Bruce Meyer, John Karl Scholz, Lauren Sinai, and the reviewers for helpful suggestions. He takes responsibility for all errors and omissions. The views expressed in this paper are those of the author and are not necessarily those of the Federal Re

Instrumental variables methods for the correlated random coefficient model

Abstract: This paper considers the use of instrumental variables to identify a correlated random coefficients model in which coefficients are correlated with (or stochastically dependent on) the regressors. A correlated random coefficients model is central to the human capital earnings model. Conditions are given under which instrumental variables identify the average rate of return. These conditions are applied to David Card's version of Gary Becker's Woytinsky lecture.

Gender wage discrimination bias? A meta-regression analysis

Abstract: This study provides a quantitative review of the empirical literature on gender wage discrimination. Although there is considerable agreement that gender wage discrimination exists, estimates of its magnitude vary widely. Our meta-regression analysis (MRA) reveals that the estimated gender gap has been steadily declining and the wage rate calculation to be crucial. Large biases are likely when researchers omit experience or fail to correct for selection bias. Finally, there appears to be significant gender bias in gender research. However, it is a virtuous variety where researchers tend to compensate for potential bias implicit in their gender membership.

An Analysis of the Impact of Sample Attrition on the Second Generation of Respondents in the Michigan Panel Study of Income Dynamics

Abstract: Only a few studies have tried to estimate the trend in the elasticity of children's economic status with respect to parents' economic status, and these studies produce conflicting results. In an attempt to reconcile these findings, we use the Panel Study ...

An Analysis of Attrition in the Panel Study of Income Dynamics and the Survey of Income and Program Participation with an Application to a Model of Labor Market Behavior

Abstract: This paper analyzes attrition behavior in two major longitudinal surveys, the Panel Study of Income Dynamics (PSID) and the Survey of Income and Program Participation (SIPP) Significant indicators in a model of attrition include measures of mobility and variables that correspond to the interviewer and the interview process There is evidence that surveys with more waves or higher frequency interviews experience higher attrition rates The estimation results for a model of attrition and labor market behavior show little indication of bias due to attrition but there is evidence that the labor market behavior of attritors and nonattritors is different

Panel Attrition from the Panel Study of Income Dynamics Household Income, Marital Status, and Mortality

Abstract: This analysis is concerned with the determinants of panel attrition from the Panel Study of Income Dynamics and with its consequences for estimation of dynamic behavioral models which exploit the panel or longitudinal information-household income dynamics, marriage formation and dissolution, and adult mortality risk. We develop and estimate joint models of attrition and one or more of these substantive processes, and allow for correlation across the equations through random effects. Although we find evidence of significant selectivity in attrition behavior, the biases that are introduced by ignoring selective attrition are very mild.

Kernel regression in empirical microeconomics

Abstract: We consider the implementation of Kernel methods in empirical microeconomics with specific application to Engel curve estimation in the analysis of consumer behavior. A set of recently developed rests for the parametric null hypothesis against a nonparametric alternative are discussed and implemented for the Engel curve application. We also consider semi-parametric estimation in partially linear models and the case of endogenous regressors. Gauss-based software is available for each technique implemented in the paper.

A Cohort Analysis of Saving Behavior by U.S. Households

Abstract: The aim of this paper is to shed some light in the decline in personal saving rates in the United States in the 1980s. For a such a purpose the paper analyses the only U.S. data set containing information on consumption and income at the household level: the Consumer Expenditure Surveys (CEX) from 1980 to 1991. Because the CEX is not a panel, most of the analysis is conducted using average cohort techniques. The paper identifies a "typical age profile" for saving rates. Such a profile is "hump shaped" and peaks around age 57. The paper also argues that such a profile was "shifted down" for the cohorts born between 1920 and 1939 relative to the younger and older cohorts considered. These cohorts are the parents of the baby boom generation. The paper also argues that these "cohort effects" can account for a non-neglible proportion of the decline in aggregate saving because these cohorts were, during the 1980s, in the ages when saving rates are typically highest. The result is robust to the consideration of several controls and holds for several definitions of consumption. The only exception is when durable expenditure is considered as saving rather than consumption.

An evaluation of the National Longitudinal Survey on Youth.

Abstract: To investigate the nature and potential consequences of attrition in the NLSY [National Longitudinal Survey on Youth] this paper studies the reliability of these data by focusing on three issues. These are attrition patterns characteristics of attriters and returnees and comparability. "We begin in Section II by describing the structure of the NLSY sample. We then summarize the patterns of attrition and returning found in the NLSY.... In Section III we explore how attrition might corrupt the picture of labor market experiences portrayed in the NLSY by developing a characterization of [attriters and returnees].... In Section IV we carry out an extensive comparison of the NLSY with the CPS.... Finally Section V summarizes and interprets our findings." (EXCERPT)

Rank Regressions, Wage Distributions, and the Gender Gap.

Abstract: In this paper, we propose a novel rank-based procedure to decompose changes in the gender wage gap into three components: changes in the skill distribution, changes in the wage structure, and improvements in the position of women in a "distribution of reference." Using CPS data from 1979 and 1991, we perform the decomposition at each percentile of the wage distribution and show that the results are sensitive to the choice of distribution of reference (male versus overall wage distribution). We also find that relative wage gains of women may have been a source of increasing wage inequality among men.

The market work behavior and wages of women: 1975-94.

Abstract: Changes in the labor market participation and wages of women in the United States are analyzed over the period 1975-1994 using data from the Current Population Survey. "Women are organized into nine birth cohorts five schooling groups and each year of age from 25 to 60 years and their weekly and annual work hours their annual work weeks their employment-population ratio and their real average hourly earnings tabulated. Schooling differences in work behavior have become wider in recent cohorts as have their wages. The relationship between work and wages is estimated for women of different ages cohorts and marital status. The gap between the work of unmarried and married women has narrowed and the role of wages (both the wages of women and those of husbands) is examined to determine the extent to which changes in wages account for these movements." (EXCERPT)

Women helping women? role-model and mentoring effects on female ph.d. students in economics

Abstract: One potential method to increase the success of female graduate students in economics is to encourage mentoring relationships between these students and female faculty members, via increased hiring of female faculty, or having female faculty serve as dissertation chairs for female students. This paper examines whether either of these strategies results in more successful outcomes for female graduate students, using survey information on female graduate students and faculties of Ph.D.-producing economics departments. The empirical evidence provides virtually no support for the hypothesis that initial job placements for female graduate students are improved by adding female faculty members, or by having a female dissertation chair. However, female faculty members appear to reduce time spent in graduate school by female students.

Does the Gender Composition of Sibships Affect Women's Educational Attainment?.

Abstract: Data from the [U.S.] Panel Study of Income Dynamics the November 1989 Current Population Survey and the National Longitudinal Study of Women suggest that women with sisters may have completed less schooling than women without sisters.... There is relatively weak evidence for this hypothesis in the analysis on which the findings are based. Analyses of the effects of sibling gender composition on educational attainment among cohorts of women...offer no support for this hypothesis or for other related hypotheses about the effects of the gender composition of sibships. (EXCERPT)

Unions, Wages, and Skills

Abstract: Studies uniformly conclude that union wage effects are largest for workers with low measured skills. Longitudinal analysis using 1989/90-1994/95 CPS matched panels produces union premium estimates equivalent across skill groups, following appropriate sample restrictions and control for worker- specific skills. NLSY evidence on aptitude scores confirms that union workers with high measured skills have relatively low unmeasured skills. Differential selection by skill class and skill homogeneity in union workplaces result from employer and employee sorting in response to wage standardization, union organizing where skills are homogeneous, and unionized employers' reluctance to hire the most as well as least able workers.

The importance of sample attrition in life cycle labor supply estimation

Abstract: We examine the importance of possible nonrandom attrition to an econometric model of life cycle labor supply using both a Wald test comparing attriters to nonattriters and variable addition tests based on formal models of attrition. Estimates using the Panel Study of Income Dynamics show that nonrandom attrition is of little concern when estimating prime-age male labor supply because the effect of attrition is absorbed into fixed effects in labor supply. The wage measure and instrument set have much larger effects on the estimated labor supply function of prime-age men than how one adjusts for panel attrition.

Why do small establishment hire fewer blacks than large ones

Abstract: This paper shows that small establishments are much less likely to hire and employ blacks than are larger establishments. A number of possible explanations for this result are considered, such as differences across establishments in application rates from blacks, skill needs, locations, and recruiting behavior. Although these factors can account for some of the observed differences in race of new hires between small and large establishments, much remains unexplained. The results suggest that discrimination in hiring may be much more pervasive at smaller establishments than larger ones.

Elderly immigrants on welfare.

Abstract: The difference between immigrants and natives use of welfare programs [in the United States] is concentrated among the elderly This paper examines the determinants of immigrants welfare participation decisions to evaluate the consequences of changes in immigration and welfare policy An important finding for immigration policy is that immigrants who arrive after age 55 are significantly more likely to use welfare than the typical immigrant who arrives during prime working years Surprisingly this age-at-arrival effect is not explained by differences in social security benefits between young-arrivers and old-arrivers The problem of immigrant welfare use is not simply low incomes or poor labor market performance: decisions regarding take-up of benefits are an important explanation for the effect of age at arrival Finally the sharp rise in immigrants use of welfare during the 1980s was due mostly to higher welfare participation rates of new immigrants (EXCERPT)

Attrition in panel survey data and the estimation of multi-state labor market models

Abstract: In the analysis of labor market transitions, spells that are incomplete due to panel survey attrition are treated as spells that are subject to independent right-censoring. It is likely, however, that the unobserved characteristics affecting individual transitions on the labor market are related to those affecting the attitude toward survey participation. In that case, the transition rate estimates are inconsistent. In this paper we analyze the relation between durations spent in labor market states and the duration of panel survey participation, by explicitly modeling and estimating both stochastic processes. We use multi-state multi-spell models which allow for stochastically related unobserved determinants.

Experience, Tenure, and Wage Growth of Young Black and White Men

Abstract: This paper studies the source of differences in wage growth between young black and white workers. Focusing on "terminal" high school graduates from the National Longitudinal Survey of Youth, we estimate the returns to on-the-job tenure and general labor market experience using ordinary least squares, Altonji and Shakotko, and Topel estimators. Results from all three estimators indicate that returns to general experience for black workers trail those for white workers, but that black workers earn equal if not higher returns to tenure than do white workers.

Estimation of Semiparametric Censored Regression Models: An Application to Changes in Black-White Earnings Inequality during the 1960s

Abstract: Building on the work of Chay (1995), this study examines the impact of civil rights policies on black economic progress using individual-level panel data. Many earnings records are censored and the degree of censoring changed during the period of interest. Consequently, valid estimates of the program effects must account for this censoring. Maximum likelihood estimation can be used if the error terms of the model are identically normally distributed. We investigate the value of using weaker assumptions on the error process to estimate the laws impact. The analysis shows that there was significant black-white earnings convergence in the South during the 1960s. We also find that semiparametric estimation methods are informative in pinpointing which parts of the model are mis-specified.

Will Extending Medicaid to Two Parent Families Encourage Marriage

Abstract: Several welfare programs in the United States restrict eligibility to single-parent families. This paper asks whether eliminating this restriction for Medicaid encourages marriage. I identify Medicaid's effect through a series of health insurance reforms that were passed in the 1980s and 1990s targeting young children. These reforms were associated with an increase in the probability of marriage of 1.7 percentage points. While the expansions offered some incentives to become married, they also created other incentives to become divorced (known as the "independence effect"). After controlling for the outflows from marriage due to the independence effect, the estimated effect increases by 10 percent.

Survey Attrition and Schooling Choices

Abstract: We use data from three cohorts of the National Longitudinal Surveys of Labor Market Experience and from the Panel Study of Income Dynamics to study the effect of survey attrition on estimates of statistical models of schooling choices. We estimate regressions using data on people who always respond to the surveys (stayers) and on people who miss some surveys (attritors) and test whether the same statistical model describes the behavior of stayers and attritors. In general (with a few exceptions) we find that attrition either has no effect on the regression estimates or only affects the estimates of the intercept (and sometimes the coefficients of birth year dummies) and does not affect estimates of family background slope coefficients.

Bootstrap Methods for Covariance Structures

Abstract: The optimal minimum distance (OMD) estimator for models of covariance structures is asymptotically efficient but has much worse finite-sample properties than does the equally weighted minimum distance (EWMD) estimator. This paper shows how the bootstrap can be used to improve the finite-sample performance of the OMD estimator. The theory underlying the bootstrap's ability to reduce the bias of estimators and errors in the coverage probabilities of confidence intervals is summarized. The results of numerical experiments and an empirical example show that the bootstrap often essentially eliminates the bias of the OMD estimator. The finite-sample estimation efficiency of the bias-corrected OMD estimator often exceeds that of the EWMD estimator. Moreover, the true coverage probabilities of confidence intervals based on the OMD estimator with bootstrap-critical values are very close to the nominal coverage probabilities.

The Demand for Child Care Quality. An Hedonic Price Theory Approach.

Abstract: I use an hedonic price theory approach to estimate the demand for child care quality. Two complementary surveys, the National Child Care Survey, 1990 (NCCS) and the Profile of Child Care Settings Study (PCS), allow me to derive an implicit price for staff-to-child ratio. I use this price as an explanatory variable in a demand equation for this quality attribute. Direct purchase-of-service contracts or voucher programs, by subsidizing only those providers that satisfy state regulatory requirements, effectively lower the implicit price of regulated attributes, such as staff-to-child ratio. Results of this study suggest that such tied subsidies have almost no influence on the demand for child care quality.

On the Road Marriage and Mobility in Malaysia

Abstract: Migration choices of husbands and wives in a dynamic and developing country are studied in the context of an economic model of the household. Data are drawn from the second wave of the Malaysia Family Life Survey. Exploiting the retrospective histories, we compare moves that take place before marriage with those made during the marriage; among the latter, moves that are made with the spouse are distinguished from those made alone. The evidence indicates that male mobility is primarily economic in motivation and related to labor market factors. Moves by women, however, seem to be more closely related to fertility or family considerations. Migration is apparently not simply an individual decision; the attributes of the spouse are an important influence on mobility, albeit in an asymmetric manner. Moving toward a broader definition of the household, we find the characteristics of the parents, parents-in-law, and also the (relative) age and gender of siblings all influence mobility in a rich, if complex, way.

The Link between Ability and Specialization: An Explanation for Observed Correlations between Wages and Mobility Rates.

Abstract: Wage levels and turnover rates are negatively correlated across types of employment, and this fact is often interpreted as evidence that high-wage jobs are rationed. A simple training model illustrates, however, that this correlation may arise because able workers have an incentive to choose highly specialized jobs. In any job, the most able workers possess the most valuable stocks of specific skills and therefore face the highest mobility costs. Thus, able workers may have a comparative advantage in specialized employments. Data from the national Longitudinal Survey of Youth provide an opportunity to evaluate the merits of the training model developed here. Data on worker training and mobility provide support for several implications of the model. The model also provides new ways to interpret existing results in the literature on interindustry wage differentials.

The Effect of Family Planning Programs on the Fertility of Welfare Recipients Evidence from Medicaid Claims

Abstract: This paper explores the effectiveness of family planning (FP) services subsidized by the Medicaid program in lowering the fertility of female welfare recipients. Previous studies used single-equation estimates and produced conflicting and sometimes confounding results. Thus this study employed a bivariate probit model to estimate the effect of contraceptive acceptance on the individuals probability of giving birth without the bias that may result from single-equation estimates. A unique data set created from Maryland Medicaid claims records was also used. Results suggest that Medicaid FP programs are much more effective at lowering fertility than indicated by previous studies using single-equation estimates.