In this paper, we review the evidence for job polarization in the U.S. and provide a description of the occupational employment changes that characterized the U.S. labor market during the 1970s, 1980s, and 1990s. We begin by replicating the existing job polarization trends, which are produced using a modified occupational coding scheme in- tended to make occupational categories comparable over time. Using two alternative procedures to obtain consistent occupational codes across decades, we show that the finding that jobs polarized in the 1990s rela- tive to the 1980s no longer holds. Instead, we find that occupational em- ployment shifts were very similar during the two decades. In addition, we demonstrate that the method used to rank occupations according to their skill content has a substantial impact on the employment growth in low-skill job categories. Finally, using an additional occupational crosswalk that allows us to obtain consistent occupational categories from 1970 to 2002, we provide evidence in favor of a long-term trend to- wards employment growth in high-skill jobs and employment decline in some middle-skill jobs, but no sharp contrast between the 1980s and the 1990s. Our findings suggest that the evolution of the occupational em- ployment structure and the divergent wage growth patterns observed during the 1980s and 1990s do not easily fit within the routinization story as usually told.

This study calculates the impact of federal income taxes on the progressiveness of the Old Age and Survivors Insurance (OASI) program. It uses the Health and Retirement Study (HRS) data linked with the Social Security Earnings Records to estimate OASI contributions and benefits for individuals and households, before and after income taxes, for three birth cohorts. It uses two measures of progressivity: redistribution by decile (the difference between the share of total benefits received and the share of total taxes paid) and “effective progression” (the change in the Gini coefficient). Under both measures, the results without the income tax confirm previous findings: Social Security is progressive on an individual basis, but that progressivity is dramatically cut when one calculates it on a household basis. Adding income taxes could make the program either more or less progressive. On the one hand, the tax treatment of contributions makes the system even less progressive than generally reported. On the other hand, the taxation of benefits makes it more progressive. The net result is that adding the personal income tax to the analysis makes Social Security more progressive than without taxes, on both the individual and household bases. Importantly, however, the impact of taxation on redistribution increases significantly among younger cohorts. Under current law, the Social Security system becomes more progressive over time.

In this paper, we investigate the relationship of male and female economic opportunity to marriage formation in 2000, examining how relationships vary with race and ethnicity. To address these issues, we carry out multi-level analysis of the effects of local economic conditions on the marriage decisions of young men and women. We will compare results obtained using the 5% sample of Census 2000 from the Integrated Public Use Microdata Series (IPUMS) with those from an analysis of restricted access census files available only through Census Bureau Research Data Centers (RDCs). The internal files allow the identification of small geographic regions to measure local economic conditions. This analysis will help gauge the value of the restricted access data by presenting a benchmark using public data.

In this paper I consider the possible mechanisms by which occupation sex segregation might be associated with the odds of marriage. I use data from the National Longitudinal Survey of Youth 1979 (NLSY79) to estimate the odds of marriage over a period of fourteen years as a function of occupational and personal characteristics. I find that both women and men benefit from making gender-typical occupational choices. Additionally, men are penalized from working in female-dominated occupations. In a supplementary analysis I use data from the National Longitudinal Study of Adolescent Health 1994-2008 (Add Health) to investigate whether these findings might be attributed to stereotypic differences in sexual orientation or in personal attractiveness.

This thesis consists of three chapters pertaining to issues of long-term relationships in labour markets. In Chapter 1, I analyze a model of a two-period advice game. The decision maker chooses to retain or replace the advisor after the first period depending on the first period events. Even though the decision maker and the advisor have identical preferences, this potential replacement creates incentive for the advisor to avoid telling the truth. I show the condition under which the decision maker can find a random retention rule that induces a truthful report from the advisor, and I characterize an optimal retention rule that maximizes the decision maker’s expected payoff. In Chapter 2, I propose a search theoretic model of optimal employment contract under repeated moral hazard. The model integrates two important attributes of the labour market: workers’ work incentive on the job and their mobility in the labour market. Even though all workers and firms are ex ante homogeneous, these two factors jointly generate (1) wages and productivity that increase with worker’s tenure and (2) endogenous dynamic heterogeneity of the labour productivity of the match. The interaction of these factors provides novel implica- tions for wage dispersion, labour mobility, and the business cycle behaviour of macroeconomic variables. Lastly, in Chapter 3, I quantitatively assess wage dispersion and business cycle implications of the model developed in Chapter 2. In terms of wage dispersion, the model with on-the-job search with wage-tenure contracts seems to accommodate sizable frictional wage dispersion. The model, however, generates very small productivity difference among workers, and shows weak evidence that the productivity difference generated by the endogenous variations in incentives is responsible for frictional wage dispersion. In terms of business cycle implications, workers’ endogenous effort choice first amplifies the effect of productivity shock on unemployment rate. Second, responses of workers to productivity shocks generate marked difference between the effects of temporary productivity shock and that of permanent shock. Third, the analysis shows the importance of the distributional effect on macroeconomic variables during the transitory periods after a shock.

k-Anonymisation is an approach to preventing sensitive information about individuals being identified or inferred from a dataset. Existing work achieves this by ensuring that each individual is linked to multiple sensitive values, but they have not adequately considered how the range formed by these sensitive values may affect privacy protection. When such a range is small, sensitive information about individuals may still be inferred quite accurately, thereby breaching privacy. In this paper, we study the problem of range disclosure (i.e. estimating sensitive information through ranges) in k-anonymisation, and propose Range Diversity for quantifying the effect of range disclosure on privacy protection. Our measure considers several possible attacks and allows anonymisers to specify the level of protection required in a flexible manner. Extensive experiments show that range diversity provides better protection for range disclosure and higher level of data utility than the existing methods.

Scholars have long been interested in exchange and matching (assortative mating) in romantic partner selection. But many analyses of exchange, particularly those that examinebeauty, fail to control for matching and/or consider only gendered patterns of exchange. Because traits that are desirable in mates are positively correlated between partners and within individuals, ignoring matching produces spurious evidence of exchange. I use data from the National Longitudinal Study of Adolescent Health Romantic Pair Sample, a large (N = 1,502), nationally representative probability sample of dating, cohabiting and married couples, to investigate whether desirable characteristics are traded for different desirable traits, net of matching. I revisit findings from a variety of previously published studies and find that controlling for matching eliminates nearly all evidence of exchange. The discussion focuses onthe implications that these results have for socioeconomic stratification, gender equality, and sociobiological theories of partner selection.

Workers with more education typically earn more than those with less education, and the difference has been growing in recent decades. Not surprisingly, the percentage of the population going after and getting a college degree has been rising as well. Since the late 1970s, though, the increase in college attainment has stalled for men and gathered steam for women. Among college-age individuals, more women now graduate than men. Changes in labor market incentives appear to explain the increased investment in education made by women. But men’s investments in education have been much less responsive to the same incentives.

Immigration is transforming the U.S. labor force with important consequences for Social Security’s adequacy and finances. Using longitudinal data from the Survey of Income and Program Participation matched to rich administrative data on lifetime earnings and benefit receipt, we measure the extent to which non-natives’ lifetime earning patterns, payroll taxes paid, benefits received, and total incomes differ from those for the U.S.- born population. We consider other outcomes important to retirement security, like health status, marital status, and financial wealth. We also compare various immigrant groups with one another. Our findings stress heterogeneity in labor force and Social Security experiences among immigrants.

The exploration of the evolution of systems of cities has a long history illustrated by Berrys (1964) classic paper. In this chapter we take a simple model that is usually used for modelling the evolution of retail centres within a city and re-interpret it as a model of a system of cities. The retail model is outlined initially and its system interpretation thereafter. We then explain the idea of urban DNA and its evolution. This is followed by a description of the current system of interest the evolution of Chicago from 1790 to 1870 in the context of the development of the United States in that period with particular reference to railways. Some results are presented that explore the evolution of the populations of cities in this system and some of the many possible avenues for further research are discussed in the concluding section.

Local efforts at economic development have increased in the United States over the last three decades. In 1977, 2 of the 55 largest municipalities in the Milwaukee-Racine combined metropolitan statistical area (CMSA) used tax increment financing (TIF) to help pay for development projects. Three had an official Community Development or Redevelopment Authority (CDA/RA) to help coordinate and promote local development. Thirty years later, 43 were using TIF and 24 had a CDA or an RA. Municipal economic development officials and city councils, with the support and/or approval of state governments, often engage in local development and provide incentives to businesses to encourage growth. However, research regarding the substantive efficacy of local incentives in promoting economic development within a given metropolitan area is quite limited.This analysis seeks to partially fill that gap by replicating a regression study first offered by Robert Wassmer in a 1994 paper modeling the effect of local incentives in the Detroit metropolitan area during the 1954 1987 period. The original model has been adapted to the Milwaukee-Racine area, updated with newer data, and refined with the use of more robust statistical method. This research puts TIF, the primary local economic development incentive (EDI) in the Milwaukee-Racine area, into historical context before proceeding to answer the question are levels of economic development in the Milwaukee-Racine CMSA affected by municipal economic development efforts? Economic development efforts are operationalized here as per-capita cumulative municipal spending on TIF projects (categorized as relating to manufacturing, non-retail service, or retail industries) and by the cumulative time a CDA or RA has been active in a given municipality, both derived from original research. The measures are regressed with corresponding levels of local economic development in the manufacturing, non-retail service, and retail industries over the period 1977-2007. The model incorporates controls for important exogenous and endogenous factors including macroeconomic trends and unique local characteristics affecting economic development in order to isolate the relationship of municipal development effort to local economic results. The results indicate that greater spending on improvements financed by tax increments has not been associated with increased manufacturing employment, non-retail service employment, or non-retail service receipts, though TIF was associated with increased manufacturing value added. The effects of CDAs and RAs are more ambiguous. Their presence is found to be significantly positively associated with three development measures, but negatively associated with two others. The patterns observed in the Milwaukee-Racine area from 1977 to 2007 differ in their details from those found in prior research, but the conclusion they suggest is consistent with that from prior work: while local efforts at economic development may have been associated with positive outcomes for particular localities, when the metropolitan region is considered as a whole, local development appears to have been an ineffective anodyne for urban decay. This research suggests that it may in the interest of the State of Wisconsin to consider closer monitoring of the use of local economic development incentives.

This study explores the mechanisms of labor force participation of immigrants age 50 and above with two particular interests: (1) the effect of aging and immigration status by time of arrival, (2) the effect of aging, immigration status, and country of origin. Data comes from three year pooled sample of American Community Survey, 2006-2008. I will run sets of weighted logistic regression which include interaction terms. Preliminary results show that the immigrants who arrived between 1965 and 2000 are more likely to have worked than the U.S. born counterparts. As population ages, the magnitude of difference by nativity as well as timing of immigration decrease although immigrants and Post IRCA immigrants in particular, continue to show higher presence in the labor market well into 70s. When compared to Asian counterparts, among Latin American immigrants, 1965-1985 cohort is less likely, but Post 2001 cohort is more likely to have worked.

The goal of federal food and nutrition programs in the United States is to improve the nutritional well-being and health of low income families. A large body of literature evaluates the extent to which the Supplemental Program for Women Infants and Children (WIC) has accomplished this goal, but most studies have been based on research designs that compare program participants to non-participants. If selection into these programs is non-random then such comparisons will lead to biased estimates of the program's true effects. In this study we use the rollout of the WIC program across counties to estimate the impact of the program on infant health. We find that the implementation of WIC led to an increase in average birth weight and a decrease in the fraction of births that are classified as low birth weight. We find no evidence that these estimates are driven by changes in fertility or selection into live births. Our preferred estimates suggest that WIC initiation raised average birth weight by 2 g, or by 7 g among infants born to mothers with low education levels. These translate into estimated birth weight increases among participating mothers of approximately 18 to 29 g. Estimated treatments on the treated impacts among infants born to participating mothers with low education are of similar magnitude.

Using data from the National Hospital Discharge Survey, this paper analyzes the effect of Medicaid eligibility expansions from 1985 to 1996 on the health insurance coverage of women giving birth. We find that the eligibility expansions reduced the proportion of pregnant women who were uninsured by approximately 10%, although the magnitude of this decrease is sensitive to specification. The decrease in the proportion of uninsured pregnant women came at the expense of a substantial reduction in private insurance coverage (crowd-out) of at least 55%. Substantial crowd-out and the relatively small change in the proportion uninsured suggest that Medicaid eligibility expansions may have had small effects on infant and maternal health.