David Green

The first chapter studies how low-skilled immigrant entry can explain the falling labour unionization rate in the U.S. economy. This paper argues that the entry of immigrants has significantly altered the incentives of native-born workers to join labour unions and for firms to hire unionized workers, prompting a fall in unionization. The chapter uses spatial variation in immigrant entry to show that a higher entry of immigrants leads to a higher fall in unionization rates across regions in the U.S. It develops a search-theoretic framework to bear out the mechanism and test some over-identifying predictions. The model is further calibrated and finds that low-skilled immigrant entry can explain 48-55% of the total fall in union density.The second chapter exploits plausibly exogenous changes in exchange rates across source countries for immigrants in Canada to evaluate how these changes impact their earnings. It presents evidence that Canadian immigrants, in response to a 10 per cent depreciation of the home currency relative to the Canadian dollar, reduce their annual earnings by 0.36 per cent, mainly by reducing hours worked. The effect is greater for recent male immigrants, who are less educated and their spouses abroad. They also tend to be from lower-income countries and located in immigrant enclaves. Crucially, remittance senders are more affected, but these exchange rate fluctuations do not affect the amount of remittance sent. Thus, suggesting that immigrants tend to be target earners and react accordingly to exchange rate fluctuations.The third chapter examines how immigrants' labour market conditions at the point of entry affect their earnings, labour market outcomes, and reverse migration decisions both in the short and long run. Using administrative tax data, this chapter finds that it takes 12-15 years for an initial adverse effect of entering the labour market when unemployment is high to dissipate completely. It further documents the heterogeneity existing in this impact based on age, gender, marital status, country of origin, and education. The chapter provides novel insights into the outmigration behaviour of immigrants and how it depends on the initial conditions they face post-arrival.

View record

This dissertation examines various aspects of the geographic segmentation of economic activity across local labour markets. In Chapter 1, I present a methodological improvement to traditional methods of selection bias correction. Selection bias plagues many empirical studies that exploit variation across regions or cities, due to the residential sorting of individuals. Using machine learning tools I develop a semi-parametric method which relaxes the strong assumptions and restrictions imposed by more traditional methods. Through numerical experiments I show that this method tends to outperform more traditional methods and more flexibly corrects for selection bias. Using this improved methodology in Chapter 2, I correct for selection bias in estimating the returns to education at the State level using the 1990 US Census. I confirm the general pattern of an upward bias in the returns to a college education as found by previous studies, but my results depart significantly from those obtained using more traditional means of selection correction. My findings indicate that traditional methods overstate the upward bias in the returns to a college education, and conversely, understate the upward bias in the returns to an advanced degree. These results confirm the importance of using the improved methodology outlined in Chapter 1 of this dissertation.In Chapter 3, I analyse the cross-city wage effects resulting from workers having the opportunity to find relatively high paying unionised jobs. I examine, and separately identify, two distinct channels of effects; the wage bargaining channel whereby workers with relatively high alternative employment opportunities can negotiate higher wages, and the wage emulation channel whereby firms pay higher wages to stave off unionisation. I build a novel model of union formation and wage setting which informs the empirical methodology employed. Specifically, I use Bartik style shift-share instruments to deal with issues related to endogeneity. I find evidence in favour of substantial spillover effects: average hourly earnings would be 2 percentage points higher had the composition of union work remained at its 1980 level. Furthermore, I find substantial heterogeneity across sub-populations in the role played by union spillovers in wage fluctuations.

View record

In the first chapter, I develop and estimate a model of life-cycle labour supply that incorporates the role of divorce. To do so, I set a collective model of household decision-making in an intertemporal context. The reduced-form literature has produced contradictory results on the effect of divorce and divorce risk on women's labour decisions. My model provides a unifying framework within which to view these findings. It also contributes to the structural literature, which has mostly studied how divorce alters bargaining power in marriage, and ignored women's insurance response to the risk of dissolution. I find that divorce risk decreases the lifetime expected income of married women, increasing their labour in all periods. However, this effect is mitigated over time, as women stay in and learn about their marriage. Among women who experience divorce, it exacerbates pre-existing differences across marriages.In the second chapter, we investigate the role played by education in the intergenerational mobility of Canadian children. We use the Longitudinal and International Study of Adults (LISA), a panel of integrated survey and administrative data covering 1982 to 2013. We estimate that the education level of children accounts for one third to one half of the correlation between their income and their parents'. Furthermore, our estimates show comparable experiences of intergenerational mobility across individuals with differing levels of education. Both results suggest untapped opportunities in the education system to improve people's mobility.In the third chapter, we exploit the Longitudinal Administrative Databank (LAD) to study the evolution of wealth inequality in Canada, from 1982 to 2011. Until now, research has relied largely on the Survey of Financial Security, which has sporadic coverage, under-coverage at the top of the wealth distribution, and a small sample size. We use the income capitalization method to take advantage of the higher-frequency data and large sample size provided by the LAD. Consistent with existing work, we find that the top 10% share was fairly constant over the period considered. However, our results differ in that we observe that the top 1% share grew moderately but steadily between 1995 and 2007.

View record

Chapter 1 develops an empirical two-sided matching model with endogenous pre-investment. The model can be used to measure the impact of frictions in labour markets using a single cross-section of matched employer-employee data. The observed matching of workers to firms is the outcome of a discrete, two-sided matching process where firms with heterogeneous preferences over education sequentially choose workers according to an index correlated with worker preferences over firms. The distribution of education arises in equilibrium from a Bayesian game: workers, knowing the distribution of worker and firm types, invest in education prior to the matching process. I propose an inference procedure combining discrete choice methods with simulation. Counterfactual analysis using Canadian data shows that changes in matching frictions can lead to economically significant equilibrium changes in both inequality and the probability of investing in higher education. These effects are more pronounced when worker and firm attributes are complements in the match surplus function.In many economic settings, agents behave similarly because they share information with one another. Information-sharing relations among agents can be modeled as a network, and the strategic interactions among them as a game on a network. Chapter 2, coauthored with Kyungchul Song and Nathan Canen, develops a tractable empirical model of social interactions where each agent - without seeing the full information network - shares information with their neighbors and best responds to the other players based on simple beliefs about their strategies. We provide conditions on the information networks and beliefs of agents such that their best responses exhibit economically intuitive features and desirable external validity relative to equilibrium models of social interaction. Moreover, the setup admits asymptotic inference without requiring that the researcher observes all the players in the game, nor that the they know precisely the sampling process. Chapter 3 discusses how discrete distributions of unobserved heterogeneity can be identified using information on sample attrition. Although attrition is often seen as a source of selection problems, we argue that it can also be used to solve selection problems - even in the absence of covariates or panel data.

View record

Abstract Indigenous peoples throughout the world live in more difficult socio-economic circumstances than their non-indigenous counterparts. This dissertation investigates the long run economic and cultural consequences of one of the most infamous policies to affect Indigenous peoples: the forcible removal of children from their homes and their placement in boarding schools (also known as residential schools). These sorts of policies were instituted in numerous countries throughout the world, including the United States, Canada, and Australia, and they have been heavily criticized. The policies have often had the stated goal of cultural assimilation, are generally perceived to have been educational failures and to have harmed Indigenous peoples both directly and intergenerationally. I investigate these possibilities using several sources of variation. I find that attendance at a residential school is associated with both economic and cultural assimilation but I do not find strong evidence that this is transmitted intergenerationally. The results suggest that while residential schooling had significant impacts, it is likely not the predominate cause of the economic disparity observed today. To get a sense of plausible alternative explanations of Indigenous disparity in the Canada context, I finish by establishing some of the basic patterns in earnings differences between Indigenous groups and their non-indigenous counterparts.

View record

In this dissertation, I explore barriers to educational attainment faced by Canadianyouth from three different perspectives. The first paper, which is joint with DavidGreen and Giovanni Gallipoli, uses an extended version of an unobserved factormodel to investigate the factors that influence a youth’s decision to drop out ofhigh-school. Our results support three main conclusions. First, ability at age 15plays an important role in dropping out. Second, parental valuation of educationhas a substantial impact on medium and low ability teenagers. Third, parentaleducation has no direct effect on dropping out once we control for ability andparental valuation of education.The second paper poses the question ”Can neighbourhoods change the decisionsof youth on the margin of university participation?” I use the fraction ofadults who have at least a Bachelors degree in a small geographic area surroundinga youth’s home to proxy one source of potential peers and role models. Resultssuggest that at the mean neighbourhoods have a substantial impact on the likelihoodof attending university.I further explore this finding by estimating the marginal effect of neighbourhoodcharacteristics at different points on the socio-economic distribution, andthe distribution of reading test scores. One striking finding is that while universityparticipation among youth from families with university degrees is unaffectedby neighbourhoods, the marginal effect of neighbourhoods is largest for the highskilledyouth from lower socio-economic backgrounds.The final paper evaluates the impact of a post-secondary preparatory course onhigh school achievement. BC AVID, a pilot project implemented with a randomassignment design, encouraged students whose grade 8 grades were within the Band C range to enrol in advanced classes and provided academic support throughan elective course. I find that while the program increased the chances that studentsenrolled in an advanced math course, it also increased the likelihood that studentsfailed the provincial examination. BC AVID did, however, help students pass theircourse work and as a result more students obtained credit for the advanced mathcourse.

View record

No abstract available.

This thesis explores several issues in the adaptation process of immigrants andtheir children in Canada.Chapter 2 investigates why second-generation immigrants are better educatedthan the remaining population. Using a standard human capital framework whereindividuals choose how much to invest in both their children's and their ownhuman capital, I show that a gap in education can arise in the absence ofdifferences in unobservable characteristics between immigrants and the nativeborn. Rather, it can arise due to institutional factors such as imperfecttransferability of foreign human capital and credit constraints. The model'skey implication is a negative relationship between parental human capitalinvestments and children's educational attainment, particularly in familieswith uneducated parents. I find strong empirical evidence of such tradeoffs inhuman capital investments occurring within immigrant families.Chapter 3 re-assesses the effect of living in an ethnic enclave on labourmarket outcomes of immigrants. I find evidence of cohort effects in therelationship between mean earnings and the proportion of co-ethnics in the CMAwhich vary by education level. Next, using information on the proportion ofone's friends who share one's ethnicity, I test a common assumption that theenclave effect is a network effect. I find that traditional, geography-basedmeasures of the ethnic enclave effect capture the impact of factor(s) otherthan social networks. In fact, the two effects generally offset each other tosome degree in determining immigrant employment outcomes. Neither measure has astatistically significant effect on average immigrant earnings, at least incross-sectional data.Chapter 4, co-authored with David Green and Craig Riddell, tests twoalternative theories about why immigrants earn less than native-born workerswith similar educational attainment and experience - discrimination versuslower skills (measured by literacy test scores). We find that immigrant workerseducated abroad have lower cognitive skill levels (assessed in English orFrench) than similar native-born workers. This skills gap can explain much ofthe earnings gap. At the same time, foreign-educated immigrants receive nolower returns to skills than the native born. These results offer strongevidence against the discrimination hypothesis.

View record