Studies on Economic Well-Being: Essays in the Honor of John P. Formby: Volume 12

Research on Economic Inequality Volume 12 is the outgrowth of University of Alabama Poverty and Inequality Conference, May 22–25, 2003. The motivation for the conference was to honor John P. Formby upon his retirement. The conference, funded by the University, was designed to bring together three groups of people; first, some of the most recognized scholars in the field, second, current and former colleagues of John Formby’s working in this field, and third, Dr. Formby’s former Ph.D. and post-doctoral students. Seventeen papers were presented, 11 of which are authored or co-authored by Dr. Formby’s former students. Peter Lambert and Yoram Amiel also participated in the conference. Dan Slottje, John Creedy, Shlomo Yitzhaki and Quentin Wodon did not attend but contributed papers.

Labor markets for unskilled and low-wage workers in the United States stagnated in the last quarter of the 20th century. The collapse in the low-wage labor market has been well documented¹ and numerous research initiatives have investigated the causes. Despite some geographical mismatches between buyers and sellers low-paying jobs are generally available but forces are at work on both the demand and supply sides of unskilled labor markets that make it increasingly difficult for working families at or near the bottom of the income distribution to earn enough to meet basic needs. Welfare reform effectively increased the supply of unskilled workers, which placed added pressures on wages and earnings of low-income families.

This study presents estimates of the effect of changes in the real minimum wage on the employment ratio of three groups believed to be most vulnerable to changes in the minimum wage: teenagers, young adults, and adult high-school dropouts. It also examines the effect of the minimum wage on three sub-groups within each of these larger groups: males, females, and nonwhites. The data set was obtained from the monthly outgoing rotation groups of the Current Population Survey (CPS), Three Budgets for Urban Families, and the CPI-W for various urban areas. The sample period is 1979–1999.

The leaky bucket and the transfer principle are tested under conditions of individual uncertainty, behind a veil of ignorance and when positions are known. We find that choices under individual uncertainty are slightly more risk seeking than behind a veil of ignorance indicating that the conventional practice of modeling inequality aversion as risk aversion does not lead to serious error. However, our subjects can not be said to be risk seeking or risk averse but rather protect against downside risks and seek upside gain. As in previous experiments, we find that choices with positions known are quite insensitive to inefficiency and exhibit considerable antipathy to returns that accrue to others, whether richer or poorer. Richer American males are least likely to support leaky-bucket transfers that reduce inequality once positions are known. Lottery players, but not smokers show greater risk preference given individual uncertainty.

This paper analyses the principles underlying the theories of risk and inequality, and the connections between the two. Using two experimental designs, we investigate the structure of individuals’ rankings of uncertain prospects in terms of risk and inequality. We examine these individual perceptions in the light of the conventional principles underlying risk and inequality. We show that, although the principle of mean-preserving spreads and the principle of transfers are often rejected a weaker principle, “lowest-to-highest” is usually supported.

This paper examines the notion of intermediate inequality and its measurement. Specifically, we investigate whether the intermediateness of an intermediate measure can be preserved through repeated (affine) inequality-neutral income transformation. For all existent intermediate measures of inequality, we show that the intermediateness cannot be preserved through the transformation; each intermediate measure tends to either a relative measure or an absolute measure. This observation is then generalized to the class of unit-consistent inequality measures. An inequality measure is unit-consistent if inequality rankings by the measure are not affected by the measuring units in which incomes are expressed. We show that the unit-consistent class of intermediate measure of inequality consists of generalizations of an existent intermediate measure and, hence, the intermediateness also cannot be retained in the limit through transformations.

We use recently developed methods to perform decompositions of the Lorenz curve in the United States by race, region, and marital status. The decomposed Lorenz ordinates are used to construct interdistributional Lorenz curves (ILCs), which allow us to identify an economic advantage by one subgroup over another or changes in economic advantage over time. We propose asymptotically distribution-free estimators for the ILCs and apply these estimators to data from the Current Population Survey for 1977 and 1997. As one might expect, there are economic advantages by race, region, and marital status, even in 1997. Economic advantage is greatest for marital status and smallest for region in both years. We find significant convergence (i.e. a smaller economic advantage) over time by race and region, though not by marital status.

Mobility implies initial and final distributions and a transition process linking the observations of these two distributions. An inequality index describes properties of the intitial or final distribution. A mobility index describes the transition. In most cases, mobility indices have been developed using properties of transition matrices independently of the concepts of inequality and equity that may also be used in the analysis. This paper presents a new tool – the Gini index of mobility – that provides an overall consistent framework for the analysis of mobility, inequality, and horizontal equity. The theoretical concepts are illustrated empirically using panel data from rural Mexico.

Recent studies have examined tax policy issues using labour supply models characterised by a discretised budget set. Microsimulation modelling using a discrete hours approach is probabilistic. This makes analysis of the distribution of income difficult as even for a small sample with a modest range of labour supply points the range of possible labour supply combinations over the sample is extremely large. This paper proposes a method of approximating measures of income distribution and compares the performance of this method to alternative approaches in a microsimulation context. In this approach a pseudo income distribution is constructed, which uses the probability of a particular labour supply value occurring (standardised by the population size) to refer to a particular position in the pseudo income distribution. This approach is compared to using an expected income level for each individual and to a simulated approach, in which labour supply values are drawn from each individual’s hours distribution and summary statistics of the distribution of income are calculated by taking the average over each set of draws. The paper shows that the outcomes of various distributional measures using the pseudo method converge quickly to their true values as the sample size increases. The expected income approach results in a less accurate approximation. To illustrate the method, we simulate the distributional implications of a tax reform using the Melbourne Institute Tax and Transfer Simulator.

Tax microsimulation models are based on large-scale cross-sectional survey data. Each individual or household has a sample weight provided by the statistical agency responsible for collecting the data. The typical starting point is to use weights that are inversely related to the probability of selecting the individual in a random sample, with some adjustment for non-response. It has become common for agencies, using “minimal” adjustments, to produce revised weights to ensure that, for example, the estimated population age/gender distributions match population totals obtained from other sources, in particular census data. Such calibration methods appear to be well known among survey statisticians, a highly influential paper being that by Deville and Särndal (1992).²

This paper has two goals. First it determines the respective impacts of variations in the tax rates and in the distribution of pre-tax incomes on changes in tax progressivity in the United Kingdom during the period 1960–2001. Second it checks whether macroeconomic variables or the political cycle influenced the degree of tax progressivity. The results of the empirical analysis show that the significant decrease in tax progressivity observed between 1960 and 1982 was essentially the result of a variation in the distribution of pre-tax incomes. During the later period (1982–2001) the data indicate that there was no significant change in overall progressivity and in the components of its change. The second part of this study indicates that in the long run both inflation and unemployment negatively affect tax progressivity. The impact of the political cycle on tax progressivity is not clear and the results depend on the tax progressivity index that is used. It is interesting to note that in the cases where a political effect is found it indicates that under the Labour party, tax progressivity increased, for a given level of inflation and unemployment. The econometric analysis also shows that in the short run, only unemployment has a significant effect on tax progressivity.

The impacts of median income and other variables on the Sen index of poverty in the United States are investigated using panel data with fixed time period and cross sectional effects. Estimates for the Sen index and its decomposed components – the headcount ratio, poverty gap ratio, and Gini coefficient among the poor reveal that median income among state/regions and across time systematically influences the Sen index and each of its components. However, the results reveal that labor market and demographic control variables have quite different effects on the distinct components of the Sen index.

In an economy with a developed financial system, wages and salaries are not the only source of income for households. Capital markets allow households to invest their saving and earn interest and return. An accumulation of wealth is greater for households with higher incomes, who then earn more interest and returns from their wealth, leading to more inequality of income among households. The recent financial crisis experienced in Japan, characterized by a substantial decrease in stock and real estate prices, should have had a reversing effect on the income distribution among Japanese households. Time-series data of quintile income shares and a measure of income inequality in Japan are used to analyze the effects of the financial bubble of the late 1980s and the financial crisis of the 1990s on the income distribution in Japan. The result reveals a significant de-equalizing effect of rising asset prices on income distribution in Japan. However, the equalizing effect of the falling prices of stocks and real estate was partially offset by the de-equalizing effect of rising unemployment in the late 1990s. Furthermore, taking into account the effect of the financial market condition, the income distribution fluctuates less pro-cyclically than previous studies indicated.

This paper presents a “Granger Causality” analysis of the relationship between the antitrust enforcement activities of the Department of Justice and economic growth in the U.S. economy from 1891 to 2002. Professor Posner posed the question 35 years ago about whether there was a relationship between the two economic variables. The empirical results show some economic impact from antitrust enforcement on economic growth and little feedback.

Responses to minimum income and minimum spending questions are used to produce economic well-being thresholds. Thresholds are estimated using a regression framework. Regression coefficients are based on U.S. Survey of Income and Program Participation (SIPP) data and then applied to U.S. Consumer Expenditure Survey (CE) data. Three different resource measures are compared to the estimated thresholds. The first resource measure is total before-tax money income, and the other two are expenditure based. The first of these two refers to expenditure outlays and the second to outlays adjusted for the value of the service flow of owner-occupied housing (rental equivalence). The income comparison is based on SIPP data while the outlays comparisons are based on CE data. Results using official poverty thresholds are shown for comparison. This is among the earliest work in the U.S. in which expenditure outlays have been used for economic well-being determinations in combination with personal assessments, and the first time rental equivalence has been used in such an exercise. Comparisons of expenditures for various bundles of commodities are compared to the CE derived thresholds to provide insight concerning what might be considered minimum or basic.

Results reveal that CE and SIPP MIQ thresholds are higher than MSQ thresholds, and resulting poverty rates are also higher with the MIQ. CE-based MSQ thresholds are not statistically different from average expenditure outlays for food, apparel, and shelter and utilities for primary residences. When reported rental equivalences for primary residences that are owner occupied are substituted for out-of-pocket shelter expenditures, single elderly are less likely to be as badly off as they would be with a strict outlays approach in defining resources.

This paper evaluates consumption-based poverty in the United States using Consumer Expenditure Survey data. The poverty measures rest upon micro-theoretic foundations of utility maximizing behavior and a complete demand system. The Translog model (Christensen et al., 1975) is used to replicate and extend Slesnick’s (1993) measures of poverty into the late 1990s. Consumption-based poverty analysis is extended by computing Sen (1976) indexes, which provide more complete measures of poverty than simple headcount ratios. The robustness of Slesnick’s results is tested under alternative assumptions concerning shares of services between housing and other durables across time.

In recent years a great deal of time and resources have been devoted to understanding poverty in general and elderly poverty in particular. This paper uses a panel data set spanning the years 1980–2001 to investigate the impact of certain economic and demographic factors on overall, non-elderly, and elderly poverty. We also investigate the robustness of income-only poverty measures when the threshold income level is altered.

Dominance techniques are used to analyze and rank inequality, welfare, and poverty across regions in Thailand in the 1990s. Inference-based dominance methods are applied to consumption expenditure microdata from the Household Socio-Economic Surveys (SES) of 1992, 1994, 1996, 1998 and 2000. Attention is focused on the period immediately before and after the economic contraction of 1996–1997. Lorenz dominance is employed to assess inequality, while first-order Engel food share dominance is applied to rank welfare across time and among regions. Poverty is evaluated by comparing truncated food-share quantile functions. The evidence reveals that the economic crisis in 1997 seems to affect inequality in Bangkok (the richest region) more than the Northeast (the poorest region), and most dramatic changes occur in the North and South. Welfare in Bangkok is unambiguously higher than in other regions before and after economic contraction. In fact, the great economic contraction changes the rankings of economic well-being and poverty only in the North, South, and Northeast.

We study Canadian national and provincial family income inequality from 1991 to 1997. We use special cases of generalized entropy measures, the Theil measures of inequality, since they are decomposable into between-provinces inequality and within provinces inequality. We draw statistical inferences from our findings by using the bootstrapping technique. We find that Canadian provinces have experienced differential trends in family income inequality over this period, a pattern that is masked when analyzing solely national trends. Changes in between-province family income inequality are found to be insignificant, indicating that the observed rise in overall inequality over this period is due to factors within provinces. Changes in within-province family income inequality are found to be significant. We further analyze two-way decompositions by province and education, and by province and age, to learn about the role of human capital and the life cycle in determining changes in family inequality among and within Canadian provinces.

In this paper we apply the Gini adjustment procedure developed by Bishop, Formby, and Smith (1997) to investigate the effects of demographic factors on earnings inequality in Taiwan. We advance their method using quantile regression to control for demographic factors between 1978 and 1999 based on the subsample of workers conducted by the DGBAS. It is found that the marginal impact effects of female on earnings inequality are larger than the effects of years of schooling and experience. Hence, gender gap has the most significant impact on earnings inequality in Taiwan. Finally, the policy implications from our study are that controlling for gender gap could reduce earnings inequality. In particular, adoption of an affirmative action policy for women may successfully reduce the overall level of earnings inequality in Taiwan.

This study examined the secular trends in socioeconomic inequality in obesity during the period 1971–1994 in the United States. We analyzed the national representative data collected from three waves of the National Health and Nutrition Examination Survey (NHANES) conducted between those years. The Concentration Index was calculated to measure the socioeconomic inequality in obesity across gender, age, and ethnic groups in each survey period. In general, socioeconomic inequality in obesity was reduced between the 1970s and 1990s in women and black men, although the trend was not statistically significant for black women and was stable in white men. Our results indicate that, first, the association between obesity and socioeconomic status (SES) weakened over time, and second, SES inequality was not an important contributor to the dramatic increase in the prevalence of obesity in the United States. Our findings suggest that other social and environmental factors, which have influenced changes in people’s lifestyle, might better explain the increasing overweight problem in the United States. Effective intervention efforts for the prevention and management of obesity should target all SES groups from a population perspective.