Community rating regulations equalize the insurance premiums faced by the healthy and the unhealthy. Intended reductions in the unhealthy's premiums can be undone, however, if the healthy forgo coverage. The severity of this adverse selection problem hinges largely on how health care costs are distributed across market participants. Theoretically, I show that Medicaid expansions can combat adverse selection by removing high cost individuals from the relevant risk pool. Empirically, I find that private coverage rates improved significantly in community rated markets when states expanded Medicaid's coverage of relatively unhealthy adults. The effects of Medicaid expansions and community rating regulations are fundamentally linked.

The US experienced two dramatic changes in the structure of education in a 50 year period. The first was a large expansion of educational attainment; the second, an increase in test score gaps between college-bound and non-college-bound students. This paper documents the impact of these two trends on the composition of school groups by ability and the importance of these composition effects for wages. The main finding is that there is a growing gap between the abilities of high school and college-educated workers that accounts for one-half of the college wage premium for recent cohorts and for the entire rise of the college wage premium between the 1910 and 1960 birth cohorts.

Given today's extensive religious pluralism it is hardly surprising to find many buildings with ties outside the Norwegian Church, especially in the urban religious landscape. Many of these originate from religious non-conformism during the period when 96% or more of the population belonged to the State Church and after the ordinance ban on competing faiths was abolished in the 1840s. Our population censuses provide one of the worlds longest source series about the distribution of alternative beliefs from 1865 to 1980. The slow growth is attributable to Norway as an anti-pluralistic society, where the largely voluntary activity among dissenters had difficulty competing with the professional organization of the State Church. Locally, however, groups of dissenters could still be strong, with up to one-third of the population as followers, for instance in Vegrdshei parish, half way between Oslo and Kristiansand. Dissenters had solid bastions, especially in parts of the northern Troms and Nordland provinces, an area in stfold province and along the coast of Rogaland province, including a surprisingly strong position for Lutheran congregations in parts of Vestfold, a province associated more with capitalist than with religious values. Relatively speaking, dissenters were stronger in cities than in the countryside, which among other things was related to the predominance of women among the many migrants to urban areas, and that in-migrants were easier to influence. A link between ruralurban migration and not belonging to any religious society was even easier to prove for the many men in this group analogous to theories of radicalization of the labour movement. Key words: Dissenters, religiosity, non-conformism, censuses.

Every school district has authority to determine how it will assign public school students to its schools. The principal assignment mechanism used by most school districts is to draw lines that comprise attendance zones. These zones designate the public school in which children should enroll given where they live. Some districts create sprawling, truly irregular attendance zones that pull children from far-flung neighbourhoods; others create compact zones that resemble regular-looking squares. Still, no one knows if and how many irregularly-shaped zones exist and whether they integrate or segregate students by race. This research addresses educational gerrymandering in three ways. First, it describes measures of four attendance zone properties: (1) shape; (2) internal racial diversity; (3) the distribution of racial groups across small areas within them; (4) the degree to which they exclude members of nearby racial groups in order to isolate the members of other racial groups. Second, it documents the correlations among these measures. The third and last contribution is to classify attendance zones into categories with the aim of identifying zones that represent the typical cases of shape irregularity and racial balance. To meet these goals, over 13,169 attendance zones embedded in 307 of the largest school districts are analyzed. Results indicate that educational gerrymandering is infrequent and when zones are shaped irregularly, they are almost always racially balanced.

Outlier detection is a fundamental task that is used in numerous data analytic applications. It tackles the problem of identifying rare or atypical points that widely diverge from the general behavior or model of the data. The process of detecting outliers and subsequently using them for data analysis relies on the underlying application. For example, outlier detection can be employed as a preprocessing step to clean the data set from erroneous measurements and noisy data points. On the other hand, it can also be used to isolate suspicious or interesting patterns in the data. Examples include fraud detection, customer relationship management, network intrusion, clinical diagnosis, and biological data analysis. Although many successful algorithms have been developed for outlier detection, several challenges have haunted researchers and practitioners for decades. The first one is limited algorithm scalability. Due to the fast evolution of World Wide Web, the collected data can easily reach terabyte- or even petabyte- scale. Most existing approaches, ranging from statistical methods to geometric methods, and from density-based approaches to information theory based approaches, suffer from limited scalability and do not work well on large scale data. The second one is to detect outliers in the irregular, dynamic semi-structured data such as trees and graphs. There have been some research on finding outliers from the graphs. What are the definitions for meaningful outliers in the graph context? How can we detect them accurately and efficiently? The third challenge is to build a unified and modular detection system which provides researchers a complete toolbox for outlier detection tasks. Our research aims at designing the next-generation outlier detection algorithms that tackle the above three challenges. To achieve better scalability, we have done an extensive empirical study on different optimization techniques for distance-based outlier detection. Also, we proposed an ranking scheme driven by the Locality Sensitive Hashing (LSH), which finds all outliers by only visiting a small portion of the data (10%). Find similar points of each point, or all pair similarity search, is the key operation for many distance-based, density-based and cluster-based outliers. We optimized this fundamental kernel in metric space on MapReduce platform, and scaled the algorithm to hundreds of machines and solved the inadequate memory issue. For semi-structured outlier detection, we first designed a clustering-based algorithm, and a generic clustering algorithm for sets/multisets, tree and graphs. We also studied a concrete detection application on the semi-structured knowledge base, and found more than one million anomalies. Finally, we integrated our work seamlessly into a detection framework, which accepts different types of data. Users also enjoy the freedom of choosing and comparing different algorithms.

This dissertation analyzes how government regulations can affect entrepreneurship and small business performance. The first essay focuses on the effects of occupational licensing regulation, which has increased dramatically in importance over the last several decades and currently affects more than one thousand occupations in the United States. I use confidential U.S. Census Bureau micro-data to study the relationships between occupational licensing and key business outcomes. Among findings that shed light on the effect of occupational licensing on entrepreneurship are that occupational licensing regulation does not affect the equilibrium number of practitioners, but substantially reduces their entry and exit rates and that providers of occupational licensing training, namely, schools, are larger and seem to be more profitable in states with more stringent occupational licensing regulation. In the second essay, I explore (with a coauthor) whether businesses started as franchises survive longer than those launched as independent businesses, and whether there is a relationship between state franchise relationship regulation aimed at preventing franchisor opportunism and the survival of franchised businesses. We find the difference in one-year survival rate between franchised and independent businesses to be about five percentage points, and this gap to persist across two and three-year survival rates. State franchise relationship laws, however, do not seem to affect the survival of franchised businesses. In the final essay, I analyze (with a coauthor) how personal bankruptcy laws affect entrepreneurship. Lenient bankruptcy laws may encourage entrepreneurship by limiting the possible negative consequences of business failure. We examine this relationship using variation in state bankruptcy homestead exemptions, and analyze the impact of the Bankruptcy Abuse Prevention and Consumer Protection Act of 2005 that affected these exemptions. We argue that sole proprietorships are expected to be affected by the differences in homestead exemptions, corporations, because they have limited liability, not to be affected. Consistent with these predictions, entrepreneurs choice of legal form of organization does not seem to be affected by homestead exemptions, and we find no evidence of any significant effect of the Bankruptcy Abuse Prevention and Consumer Protection Act on entry rates.

The aim of this paper is twofold: a) to explore the evolution of occupational segregation of women and men of different racial/ethnic groups in the U. S. during the period 1940- 2010 and b) to assess the consequences of segregation for each of them. For that purpose, this paper proposes a simple index that measures the monetary loss or gain of a group derived from its overrepresentation in some occupations and underrepresentation in others. This index has a clear economic interpretation. It represents the per capita advantage (if the index is positive) or disadvantage (if it is negative) of the group, derived from its segregation, as a proportion of the average wage of the economy. Our index is a helpful tool not only for academics but also for institutions concerned with inequalities among demographic groups because it makes it possible to rank them according to their segregation nature.

In this paper, we compare the relative roles of multi-factor productivity (MFP) growth and factor accumulation in goods and services for states in the US from 1980 to 2007 using the dual growth accounting framework. We find that while MFP growth was relatively high, and converged in the goods sector, it was low and diverged in services. Though the low growth in MFP in services was due to declining real user cost, the divergence itself was due to variation in wage growth. We also document that while the gap between productivity and wage growth was higher in goods, the two series were more strongly correlated in services. Finally, states with higher initial human capital experienced higher growth in both sectors.

What drives governments with similar revenues to provide very different amounts of goods with private sector substitutes? Education is a prime example. I use exogenous shocks to Brazilian municipalities revenue during 19952008 generated by non-linearities in federal transfer laws to demonstrate two things. First, municipalities with higher income inequality or higher median income allocate less of a revenue shock to education and are less likely to expand public school enrollment. They aremore likely to invest in public infrastructure that is broadly enjoyed, like parks and roads, or to save the shock. Second, I find no evidence that the quality of public education suffers as a result. If anything, unequal and high-income areas are more likely to improve public school inputs and test scores following a revenue shock, given their heavy use of private education. I further provide evidence that an increase in public sector revenue lowers private school enrollment.

This work examines the effect of cultural differences on self-employment. All the individuals considered in the analysis are second-generation immigrants who were born and live under the same laws and institutions in the US. Following an epidemiological approach, the variation in self-employment rates by ancestors' national origin can be considered as supporting evidence of the effect of culture on self-employment. Our results show that culture has quantitatively significant effects on self-employment. This finding is robust to alternative specifications and to the introduction of several controls. Additional analysis shows that there are differences in the impact of culture on self-employment by gender, in that men are more sensitive than women to culture; and by economic activity, in that those individuals involved in professional, scientific, and technical activities, and those in accommodation and food service activities, are more affected by the impact of cultural differences. We also examine the transmission of culture, observing an important role of the inter-generational transfer of culture, although the impact of culture on self-employment diminishes from generation to generation.

Americans are convinced that employment stability has declined in recent decades, but previous research on this question has led to mixed conclusions. A key challenge is that trends for men and women are in opposite directions and appear to cancel each other out. We clarify this situation by examining trends in employer tenure by sex, marital status, and parental status. We find that married mothers are behind the increase in women’s job tenure, but men and never-married women have seen declines in tenure. Furthermore, we show that the timing of tenure trends for women parallels periods of increased labor force attachment. Finally, we find that shifts in industry and occupation composition can account for the decline in tenure among men and never-married women before 1996 but not afterward. We situate these diverging trends in two broad shifts in expectations, norms, and behaviors in the labor market: the end-of-work discourse and the revolution in women’s identification with paid work. Our findings support the view that job tenure is declining for all groups, but women’s greater labor force attachment, especially their more continuous employment around childbirth, countered and masked this trend.

Organized activity participation is associated with a wide array of positive developmental outcomes. Latinos are one of the largest and fastest growing ethnic groups in the U.S., yet are less likely to participate in organized activities than their peers. Theoretically, the alignment or fit between adolescents’ and their activities’ characteristics is critical to support youths’ use and engagement in organized activities. Using qualitative data in Study 1, I examined parents’ and adolescents’ perspectives and experiences related to several indicators of ethnicity and culture in their activities. Results suggested that alignment on Spanish-language use was critical for participation. However, some Latino families did not prefer aspects of ethnicity and culture in their activities because adolescents learned about their culture with family or because adolescents wanted to fit in with their majority White peers. Study 2 tested quantitatively whether features of ethnicity and culture in the activity mattered for Latino adolescents’ experiences during activities. Ethnic and cultural features in activities, particularly respect for one’s ethnicity and culture, fostered positive experiences during activities. Unexpectedly, some ethnic and cultural features were detrimental, such that overt teaching about ethnicity and culture was related to negative feelings during the activity. There was little evidence that the relation between ethnic and cultural features in activities and concurrent experiences varied by Latino cultural orientation. Integrating the findings across these two studies, there was mixed evidence for the traditional theoretical notions that optimal development occurs in environments that fit with individual’s characteristics. Complementary fit was optimal when adolescents’ needs were considered across the many contexts in which their lives are embedded, including their families and neighborhoods. I recommend that practitioners should take care in learning about the specific families and youth that their activity serves to best understand how to meet their needs. Some aspects of culture, such as Spanish-language use may be critical for participation; other aspects may require special attention from activity leaders, such as teaching about ethnicity and culture. This dissertation is an important step in understanding how to best design activities that promote the recruitment and retention of Latino youth in organized activities.

Septic systems were identified at 241,733 locations in a 2,539-square-mile (mi2) study area that includes all or parts of 12 counties in the Metropolitan Atlanta, Georgia, area. Septic system percolation may locally be an important component of streamflow in small drainage basins where it augments natural groundwater recharge, especially during extreme low-flow conditions. The amount of groundwater reaching streams depends on how much is intercepted by plants or infiltrates to deeper parts of the groundwater system that flows beyond a basin divide and does not discharge into streams within a basin. The potential maximum percolation from septic systems in the study area is 62 cubic feet per second (ft3/s), of which 52 ft3/s is in the Chattahoochee River Basin and 10 ft3/s is in the Flint River Basin. These maximum percolation rates represent 0.4 to 5.7 percent of daily mean streamflow during the 201112 period at the farthest downstream gaging site (station 02338000) on the Chattahoochee River, and 0.5 to 179 percent of daily mean streamflow at the farthest downstream gaging site on the Flint River (02344350). To determine the difference in base flow between basins having different septic system densities, hydrograph separation analysis was completed using daily mean streamflow data at streamgaging stations at Level Creek (site 02334578), with a drainage basin having relatively high septic system density of 101 systems per square mile, and Woodall Creek (site 02336313), with a drainage basin having relatively low septic system density of 18 systems per square mile. Results indicated that base-flow yield during 201112 was higher at the Level Creek site, with a median of 0.47 cubic feet per second per square mile ([ft3/s]/mi2), compared to a median of 0.16 (ft3/s)/mi2, at the Woodall Creek site. At the less urbanized Level Creek site, there are 515 septic systems with a daily maximum percolation rate of 0.14 ft3/s, accounting for 11 percent of the base flow in September 2012. At the more urban Woodall Creek site, there are 50 septic systems with an average daily maximum percolation rate of 0.0097 ft3/s, accounting for 5 percent of base flow in September 2012. Streamflow measurements at 133 small drainage basins (less than 5 mi2 in area) during September 2012 indicated no statistically significant difference in streamflow or specific conductance between basins having high and low density of septic systems (HDS and LDS, respectively). The median base-flow yield was 0.04 (f3/s)/mi2 for HDS sites, ranging from 0 to 0.52 (ft3/s)/mi2, and 0.10 (ft3/s)/mi2 for LDS sites, ranging from 0 to 0.49 (ft3/s)/mi2. A Wilcoxon rank-sum test indicated the median base-flow yields for HDS and LDS sites were not statistically different, with a p-value of 0.345. Because of the large size of the study area and associated variations in basin characteristics, data collected in September 2012 were also evaluated on the basis of the basins physical characteristics in an attempt to reduce or eliminate other basin characteristics that might affect base flow. Basins were evaluated based on geologic area, four geographic subareas, and 45-meter (147.6 ft) buffer zone; there were no statistically significant differences between median base-flow yield for HDS and LDS basins. It is probable that detection of the contribution from septic system percolation in base flow at many of the sites visited in September 2012 was obscured by a combination of the limitations of measurement accuracy and evapotranspiration. Detection of septic system percolation may also have been complicated by leaky water and sewer mains, which may have resulted in higher streamflows in LDS basins relative to HDS basins.

We use data on immigrants who live in the United States to study the effects of exposure to hyperinflation on occupational choice. To do so, we calculate the number of years an individual had lived under hyperinflation before arriving to the US. We find that its marginal effect on the probability of being self-employed instead of wage-earner is 0.87 percentage point. This finding suggests that the macroeconomic environment one lives in permanently affects his economic behavior. The estimated effect depends on the age individuals had when exposed to hyperinflation. In particular, it vanishes for those over the age of 40.

Academic and government researchers from most industrialized countries and an increasing number of developing countries are collecting and analyzing full 24-hour time diary data. These diary data provide a comprehensive and accurate basis for generating national accounts of time spent working, sleeping and spending free time, and how daily life varies across demographic groups, across decades and across countries. Diaries can also provide data on multi-tasking, interacting with others and the location and timing of daily activities in the American Time Use Survey (ATUS) and in more than 25 other (mainly European) counties. These diary data are being standardized and archived via the Multinational Time Use Survey (MTUS) at the University of Oxford. Considerable progress has been made in producing tables of comparative time use in and across more than 20 countries and in identifying and visualizing patterns of multinational time use using Multidimensional Scaling (MDS). Major attention has been focused on uses of time in families, with special attention to the implications for our understanding of the division of paid and unpaid work within households. Given the richness of diary data available, less attention is now given to trends from time estimate questions, although brief mention is made here to useful estimates from the US General Social Survey (GSS) on important free-time activities and subjective time measures. In order to provide greater insight into the sociological meaning of these activities for national accounts, supplemental data have also been collected on subjective and qualitative characteristics of diary activities, such as enjoyment, stress and outcomes. Both in the US and other countries, diarists generally rate activities related to social life, religion and interactive childcare higher in enjoyment, and activities related to household upkeep and health care at the bottom. In more recent surveys, paid work activities and TV have received less positive ratings. Furthermore, recent US surveys indicate no increase in time pressure, a subjective factor that has been consistently related to lower life satisfaction. Fuller coverage of these ratings is covered in a separate entry in this volume titled Subjective Time.

How much additional tax revenue can the government generate by increasing labor income taxes? In this paper we provide a quantitative answer to this question, and study the importance of the progressivity of the tax schedule for the ability of the government to generate tax revenues. We develop a rich overlapping generations model featuring an explicit family structure, extensive and intensive margins of labor supply, endogenous accumulation of labor market experience as well as standard intertemporal consumption-savings choices in the presence of uninsurable idiosyncratic labor productivity risk. We calibrate the model to US macro, micro and tax data and characterize the labor income tax Laffer curve under the current choice of the progressivity of the labor income tax code as well as when varying progressivity. We find that more progressive labor income taxes significantly reduce tax revenues. For the US, converting to a flat tax code raises the peak of the Laffer curve by 6%, whereas converting to a tax system with progressivity similar to Denmark would lower the peak by 7%. We also show that, relative to a representative agent economy tax revenues are less sensitive to the progressivity of the tax code in our economy. This finding is due to the fact that labor supply of two earner households is less elastic (along the intensive margin) and the endogenous accumulation of labor market experience makes labor supply of females less elastic (around the extensive margin) to changes in tax progressivity.

This dissertation addresses two research questions: 1. To what degree do states differ in measuring the achievement gap? 2. Are there predictors to suggest why this differential occurs? The first research question requires that the degree of difference, the differential quotient, is determined. For the purposes of this dissertation, I calculated the reported achievement gaps between white and black fourth graders for the years 2005, 2007, and 2009 on the individual state reading and math assessments and the reading and math portions of the National Assessment of Educational Progress, NAEP, by percentage. For example, if whites scored 90% proficient on the state-developed measure and blacks scored 70%, the state-based achievement gap was recorded as 20. If the white proficiency percentage on the NAEP was reported at 40% and the black percentage was reported at 10%, the NAEP-based achievement gap was recorded as 30. The state-established achievement gap percentage was then subtracted from the NAEP- based achievement gap percentage to create an assessment differential quotient. In the previously explained example, the differential quotient would be 10. These calculations were also completed for the differences between fourth grade white achievement and Hispanic achievement on these assessments. The larger the differential quotient, the further the state assessment achievement gap was from the NAEP achievement gap. The differential quotient was calculated for the reading and math assessments at the fourth grade level creating six individual differential quotients for each state or 300 observations for analysis. The findings of this dissertation evidence that states vary in measuring the achievement gap locally from the national exam (NAEP) at differing degrees. These disparities fluctuate by content, by year, and from state to state. For example, the number of states reporting a positive differential quotient on the white/black achievement gap for the fourth grade reading assessment over the years studied remained relatively stable at an average of 49%. This means that 49% of states reported a smaller achievement gap between whites and blacks in fourth grade reading than was reported by the NAEP. In comparison, the percentage of states reporting a positive differential quotient for the white/black achievement gap on fourth grade math reached 78% in 2009. More state reports differed from the national reports in math than in reading with more states showing growth in math as compared to the NAEP. This dissertation also found that differential quotients increased over the three years studied in all areas except white/black reading. As reporting for No Child Left Behind was not required until 2006, it logically follows that states became more interested in reporting narrowed achievement gaps in 2007 and 2009. States struggled to reach 100% proficiency by the year 2014 and adjustments to test content, format, and procedures were made to construct an illusion of better results at the local level thus causing larger differential quotients. Finally, these findings show that some states such as Nebraska, Texas, and Virginia are consistently found to have the highest disparity between state reports and the national reports. This output leads to the second research question: are there predictors to suggest why this differential occurs more frequently in some states versus others? The answer to this question is yes. The level of black and the level of Hispanic within a state were statistically significant in both math models. The coefficients of these predictors indicate that states with higher black and Hispanic populations show a greater disparity between state-based achievement reports and national-based reports.

This thesis examines the operation of urban labour and housing markets. I bring new insights to old questions about migration, unemployment and homeownership. The first essay studies the impact of immigration on the wages of native-born workers. In standard competitive models, the effect comes entirely through changes in marginal products of different labour types. But, I argue that firms with monopolistic power can exploit the lower reservation wages of recent migrants by cutting wages for natives and migrants alike. I present evidence from cross-city variation in local skill distributions, wages, and employment rates. The second essay looks at why higher skilled workers are more likely to migrate long distances within a country. It is commonly argued that they face comparatively low migration costs. But, US survey evidence on reported reasons for moving suggests this explanation is at best incomplete. I argue that high skilled workers are relatively mobile, more fundamentally, because of larger potential gains from a successful job match. The third essay documents descriptive facts on regional unemployment differentials. In the UK, unemployment has remained persistently high in less productive cities since the 1980s. But, there is no such relationship in the US: local populations adjust quickly to meet local demand. I speculate that relatively generous out-of-work benefits in the UK may allow unemployed workers to remain in poor-performing cities, while low local housing costs discourage them from searching elsewhere. The final (co-authored) essay focuses on the determinants of homeownership. It is commonly argued that households bring forward their home purchase because of uncertainty over future house price fluctuations. But, using a life cycle model, we argue that households are more likely to respond to price risk by increasing their liquid savings. We present supporting evidence from cross-city variation in ownership rates and loan-to-value ratios.