This thesis studies the effect of repeated and long term relationships between actors engaged in economic markets. Firms hire workers for long periods and offer contracts that evolve over time, and where the history shared with the worker might affect future payments. This thesis shows that understanding the nature and implications of such relationships is central to correctly measure the realized allocation in the market and predict the effects of changes in labour policies. The opening chapter is a theoretical contribution to the repeated games literature. It demonstrates how differences in time preferences between players can be used to sustain equilibrium payoffs that are unattainable under identical discount parameters. This reveals how rich inter-temporal strategies can be utilized to sustain improbable transfers between individuals. The second chapter embeds such a relationship inside an equilibrium where actors randomly meet with each other. It contributes to the literature on labour markets with friction by demonstrating how widely available matched employer-employee data can be used to recover the production function in the economy as well as the assignment of workers to firms. This has important implications for the effectiveness of policies aiming at reallocating workers to more productive jobs. In the final chapter, workers are risk averse and productivity is uncertain. I show that in this context firms choose to offer partial insurance contracts to their workers. The repeated interactions between the firm and the worker are fundamental to understanding how employers choose to transmit part of the uncertainty to the workers. I estimate the model on Swedish data and evaluate the effects of a hypothetical progressive tax aimed at reducing income inequality and uncertainty. The exercise reveals that firms will respond to the policy by transferring more risk to the employees negating around 30% of the direct effect of the policy.

In this thesis, I study the influence of country-specific institutions on the labour supply decision of prime-aged women. In chapter 1, I use reduced-form methods, to evaluate the impact of wide changes to the benefit system and childcare subsidies targeted at households with pre-school age children in France. I estimate strong responses to maternity leave type benefits, and a positive impact of childcare subsidies on mothers employment rates in the long-run. In chapter 2, I develop and estimate a static labour supply model with part-time wage equations as well as demand-side constraints. I can compare the elasticities estimates and predictions from tax reform simulations to those of models assuming a unique hourly productivity and the absence of any employment constraint. The structural model also enables me to clearly simulate the impact of each component of the wide policy reform studied in the first chapter. In chapter 3, I try to understand why the number of hours worked by British married women is lower than that of French married women. I find that in the presence of children, British mothers are far more responsive to financial incentives. Husbands’ earnings and their interaction with childcare prices seem to play an important role. Nevertheless, the fall in hours worked in the British households with children, despite facing lower taxes than in France, remains somewhat puzzling. It could be mainly attributed (in the framework used) to different preferences. In the final chapter, I study the impact of joint-taxation on the labour supply choices of married women with working husbands in France. I simulate a revenue-neutral reform that would cancel the tax penalty or gain associated with being married. I find that the overall labour supply of these women would increase by 1.2%.

This thesis uses microeconometric methods to study the armed conflict in Colombia. Chapter 2 uses geographical and temporal variation in potential drug trafficking networks to instrument for the prevalence of violent crime in different regions of Colombia. Using changes in the prices of different international cocaine markets I identify exogenous changes in violence at municipality level. I exploit the comparative advantage that different regions in Colombia have when they are serving either the cocaine market in the US or Europe. My results suggest that homicide rates increase according to the comparative advantage each municipality has in the drug trade. However, I do not find strong evidence of the influence of drug trade over the placement and actions of left wing guerrillas or right wing paramilitary groups. Thereafter, chapter 3 uses the exogenous variation from drug trafficking to analyse the effect of homicides on the prevalence of early motherhood. My results suggest the one standard deviation increase in the homicide rate induces 2.65 p.p. increase in the probability of early motherhood. Chapter 4 proposes a microeconomic framework to describe the features of an optimal reintegration contract. In my model, a Principal (government) collects taxes from the community in order to fund a reintegration contract with the Agents (illegal soldiers). This contract involves a set of threats and benefits. The shape of such threats and benefits depends on the relative productivity of the security technology and the level of absorption of the labour markets. Last chapter concludes.

The thesis consists of four chapters utilising applied micro-econometric techniques to develop a deeper understanding of the education sector. I apply traditional economic concepts such as productivity, immigration, insurance and technological innovation to the field of education economics. Chapter one considers the consequences of academic rank in primary school on later test scores. Using administrative data tracking the student population in England, I estimate the impact of rank on later attainment through the variation in the test-score distributions across schools. The positive impact of rank on attainment helps to explain some puzzles in the education literature, such as the lack of impact of selective schools. The second chapter involves immigration and investigates how the influx of overseas students has affected enrolment of domestic students at UK universities. Using administrative data, I employ methods used in the labour literature to model crowd-out. I find no evidence of crowd-out of domestic students, and some evidence of crowd in amongst postgraduate students. Chapter three establishes the threat of accusations as new source of demand for trade union membership amongst teachers. I model union membership as legal insurance, where demand is determined by the threat of accusations. I measure threat primarily through the incidence of media stories concerning teachers in the local area. Combining these data with union membership data from Labour Force Surveys, I find that unionisation rates increase with media coverage of allegations. The final chapter is an estimation of the impact of restricting technology in the workplace on productivity. This is applied to the education setting using the autonomous decisions by schools to ban mobile phones. Obtaining histories of phone policies through surveys and combining this with administrative data on individual pupil level attainment, I use a difference in difference analysis to estimate the impact on student performance.

This thesis presents an analysis of the determinants of public sector productivity in the Federal Government of Nigeria. The first chapter introduces the thesis, with an emphasis on the creation of measurement tools and methods for, and the collection of, the data that acts as the foundation of the contribution of my thesis work. The second chapter studies how management practices that bureaucrats in the Nigerian Civil Service operate under, correlate to the quantity and quality of public services delivered. For each of 4700 projects, I have hand coded independent engineering assessments of each project's completion rate and delivered quality. I supplement this information with a survey to elicit management practices for bureaucrats in the 63 civil service organizations responsible for these projects, following the approach of Bloom and Van Reenen [2007]. I find that management practices matter: a one standard deviation increase in autonomy for bureaucrats corresponds to significantly higher project completion rates of 18%; a one standard deviation increase in practices related to incentives and monitoring corresponds to significantly lower project completion rates of 14%. I provide evidence that the negative impacts of practices related to incentive provision/monitoring arise because bureaucrats multi-task and incentives are poorly targeted, and because these management practices capture elements of subjective performance evaluation that further leave scope for dysfunctional responses from bureaucrats. The backdrop to these results, where 38% of projects are never started, implies there are potentially large gains to marginally changing management practices for bureaucrats. The third chapter studies the causes and consequences of interactions between politicians and bureaucrats in the Nigerian Civil Service along two key margins: which bureaucrats a politician delegates the delivery of public projects to, and the incentives that politicians provide to those bureaucrats. To investigate these issues, I combine data on the political careers of politicians and measures of their interactions with bureaucrats with the data presented in my first chapter. I find that politicians facing high levels of political competition are more likely to (1) delegate the implementation of public projects in their constituency to more autonomous organizations; and (2) provide informal incentives to bureaucrats in those organizations. Guided by a moral hazard model, I assess the separate impacts of the delegation and incentive margins using an instrumental variables strategy. I find that delegation to more productive bureaucrats is the key channel through which politicians improve the bureaucracy's output when faced with high levels of political competition. The results have implications for the design of organizations that regulate politicians' interactions with the bureaucracy. The final chapter presents some concluding comments.

My dissertation addresses questions in two topic areas, intergenerational mobility and migration. I first study the dynamic response of income mobility to structural changes in a model of intergenerational transmission. I illustrate that mobility today depends on past policies and institutions, such that a decline in mobility may reflect past gains rather than a recent deterioration of “equality of opportunity”. How to measure income mobility is addressed in the next chapter, in which I document that heterogeneity in the shape of income profiles generates large life-cycle biases, which cannot be eliminated with standard methods used in the current literature. Finally, in the fourth chapter I study how elements of the transmission process affect the relation between mobility over two generations and the long-run persistence of economic status within families. I provide various arguments why long-run mobility is likely lower than predictions from intergenerational evidence suggest. In the final chapter I analyse the effect of migration on labor markets. Triggered by the fall of the iron curtain, Germany experienced a sudden inflow of Czech workers that reached a local employment share of up to ten percent. I exploit this natural experiment to assess how immigration affects native workers, and to examine the mechanisms by which labor markets adjust. I find a strong and rapid response in both native wages and employment, and document substantial heterogeneity across age groups: native employment decreases most strongly among older workers, even though their wages are less affected than for other age groups. This finding suggests that the elasticity of labor supply differs across demographic groups, with important implications for the analysis of responses to labor supply shocks. When distinguishing between the different types of adjustment, I find that native employment decreases predominantly through diminished inflows into work, less so through outflows into non-employment.

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 monopsonistic 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.

In this dissertation I exploit observational and experimental data to study individual decision making when agents face social interactions or are described by non-standard neoclassical preferences. In the first chapter I study how social interactions, could explain occupational choice in an incomplete information setting. In a discrete choice framework I allow for group unobservables affecting decisions. I show that asymmetries in the peer influence enables to separately identify the effects from group members’ expected behaviour and the effects from their characteristics. I provide an empirical application to nineteenth century London. The results show that social networks were important in determining occupations but are somewhat lower than estimates which do not impose consistent beliefs nor allow for unobservables. Secondly, I implement an artefactual field experiment with small entrepreneurs. Subjects were given an initial amount of money to be invested across alternatives. Some of the subjects were informed about the possibility of getting either a high or a low income level. The income level was either predetermined or allocated after a fair lottery. Agents who started with a low income after the lottery were more risk loving. A model of reference–dependent preferences with multiple reference points, formed through recently held expectations on foregone and actual outcomes, fits most of the experimental results. In the last chapter I study game interactions in interdependent value auctions fol- lowing Kim (2003). Agents are asymmetrically informed in terms of how precisely they know the different aspects of the object’s value. Due to the mismatch of bidding strategies between these bidders, the second-price auction is inefficient. The English auction has an equilibrium in which bidders can infer information and attain efficiency. The increase in perfectly informed bidders increases the seller’s revenue. A laboratory experiment confirms key predictions about efficiency and revenues and reveals naive bidding.

This thesis studies the relationship between pension incentives and formal labour market participation in a multi-tier de ned contribu- tion pension system. During 2008 a mayor pension reform was imple- mented in Chile, changing simultaneously the redistributive welfare and contributory tier of the system, introducing several elements to boost formal labour market participation and reduce inequalities. The expected pension wealth at retirement and the accrual rate have di er- ently changed for di erent group of the population due to the reform. I estimate the e ects of the reform on formal labour market partici- pation using two di erent empirical strategies: First, I use a di erence in di erence estimator to address the e ect of the expected pension wealth on formal labour market participation. I exploit the di eren- tial e ects of the reform on individuals belonging to di erent groups to gain identi cation. The endogenous pension wealth is instrumen- talized using time and group dummies. Second, I solve and estimate a dynamic consumption, labour supply and pension savings accumula- tion life cycle structural model. It complements the existing literature by incorporating the choice of two sectors in the labour market, the formal and informal labour sectors and by allowing for intrahouse- hold bargaining power. Households choose individuals' sector labour supply and consumption in an environment with uncertainty given by sectoral wage shocks, future marital status and future fertility choices. The main results of the thesis are threefold. Firstly, the changes in the nal pension wealth at retirement and the accrual rate have reduced formal labour market participation. Secondly, the reform has increased not only the self- nanced pension wealth but also has importantly improved the nal pension due to the rst tier reform. Finally, even though the nal pension changes have been positive for both gender, the female pension improvement has been much higher than the rise for men reducing the gender inequalities.

This thesis has two major themes. The first theme (Chapters 1 – 3) focuses on search theory and industrial organisation. I build a novel model of costly search where firms choose price distributions, draws are then taken from these distributions and the collection of draws is presented to consumers. Consumers are aware of the set of prices that firms charge but do not know which price is associated with which firm. As they search, they learn and update their beliefs about the prices offered by the other firms. I extend the model by introducing a price comparison website as well as informed and uninformed consumers. Equilibrium in the model has certain desirable properties. The second theme (Chapters 4 – 6) focuses on game theory, and in particular, games with forgetful players. One of the 'conclusions' of the Games and Economics Behaviour collection of papers on imperfect recall is that there is no one way to model imperfect recall in games; in particular, modelling issues which are of little or no consequence in games of perfect recall suddenly become substantive in games of imperfect recall. Furthermore, there is little consensus on how to proceed. I introduce a class of decision problems where, if we think of forgetting in a novel but intuitive way, we can transform the game into a game of perfect recall – thus resolving the modelling ambiguities. I extend the model to show that an agent with self-control problems may in certain cases be better off having a bad memory. I consider whether a firm can offer different contracts to discriminate between different types of consumer with varying degrees of memory and sophistication. Finally, I consider the consumption behaviour of a forgetful consumer.