ROBIN Jean Marc

This doctoral dissertation focuses on wealth inequality through the prism of two markets: the marriage market (chapter 1) and the labor market (chapters 2 and 3). In the first chapter, Normann Rion and I analyze the role of marriage, and more specifically of financial endogamy, in the formation of the high wealth inequalities observed in the United States. We use a dynamic life cycle model where each agent rationally chooses his or her spouse. This model predicts that a decrease in the marriage rate reduces inequality while an increase in endogamy increases the wealth gap between Americans. In the second chapter, we analyze not the sources but the consequences of wealth inequality, again in the United States where inequality is at its highest. In particular, we propose a link between the low wealth of young people (aged 21 to 30) and their wage trajectory following episodes of involuntary job loss. Indeed, the after-effects of youth unemployment are significant, which seems a priori inconsistent with the predictions of existing labor market models. Our contribution is to show that an explanation based on a career choice under a financial constraint is theoretically and empirically plausible. The third chapter, co-authored with Cem Ozguzel, is purely empirical. We revisit the very fact that young people suffer significantly from job losses, using German administrative data. The age profile of the cost of unemployment that we obtain for Germany is comparable to that documented in the literature on the United States.

My thesis is composed of two independent parts. The first part deals with crypto-economics and the second with theoretical econometrics. In the first chapter, I present a model that predicts the total computing power deployed by miners using the bitcoin/dollar exchange rate. The second chapter uses a simplified version of the previous model to note the inefficiency of the current Bitcoin protocol and proposes a simple way to reduce the electricity consumption generated by this cryptocurrency. The third chapter explains how to identify and estimate the exact bounds of the region of identification of the average marginal effect in a logit model with fixed effects on panel data.

No summary available.

This thesis examines the themes of matching, inequality and the impact of technological change on the labor market. In particular, it addresses questions about matching between employees and firms and how this influences inequality in the labor market, both within the entire population, as well as between demographic and skill groups. It also examines how technological change affects the labor market conditions faced by workers and firms. These issues are addressed over three chapters. The first chapter, entitled "Multidimensional Heterogeneity and Matching in a Frictional Labor Market - An Application to Polarization" discusses the matching of workers to firms along multidimensional characteristics and quantifies the impact of technological change on the matching, wage, and employment patterns of different demographic groups. I construct a directed search model with multidimensional heterogeneity and estimate the model on U.S. data. I find that production complementarities between cognitive and interpersonal skills and tasks have increased, compared to those between manual skills and tasks. This shift in production technology explains much of the wage and job polarization in the United States. Moreover, although it does not control for gender differences, the model can explain a substantial fraction of the narrowing of gender gaps in wages and jobs. Co-authored with Nicolo Dalvit and Aseem Patel, the second chapter, "Intra-firm Hierarchies and Gender Disparity," examines the ranking of women in hierarchies within firms. It uses French administrative data and examines the impact of wage and employment gaps across hierarchies over time. In addition, by exploiting a corporate board quota policy in France, it assesses the impact of increased female leadership on wages and employment outcomes within firms. We find that hierarchies matter in gender wage and employment gaps. Gender wage and employment gaps increase with each level of corporate hierarchy, although these gaps narrow more over time at higher levels. Moreover, improving female leadership has different impacts across hierarchies. While a higher proportion of female board members reduces the gender pay gap at the top of the hierarchy, it does not have such an impact at the bottom. Instead, it increases the proportion of women in the lower levels working part-time, at the expense of full-time employment. The opposite is true for women in the upper levels. The third chapter, "Labor Shortages and Labor Market Adjustments: An Island Theory," co-authored with Riccardo Zago, discusses the impact of labor shortages and whether they lead to wage and salary adjustments. Using unique data on vacancies reported by firms to be difficult to fill, we document the impact of shortages across regions, industries, and occupational groups. We find that shortages lead only to wage and employment adjustments in non-routine occupations, but not in routine occupations. We show how the secular decline in routine occupations, caused by technological change, can explain the persistence of shortages in this sector and its inability to adjust.

No summary available.

No summary available.

No summary available.

No summary available.

No summary available.

No summary available.

No summary available.

No summary available.

We quantify the contribution of labor market reforms to unemployment dynamics in nine OECD countries (Australia, France, Germany, Japan, Portugal, Spain, Sweden, UK, US). We estimate a dynamic stochastic search-matching model with heterogeneous workers and aggregate productivity shocks. The heterogeneous-worker mechanism proposed by Robin (2011) explains unemployment volatility by productivity shocks well in all countries. Placement and employment services, UI benefit reduction and product market deregulation are found to be the most prominent policy levers for unemployment reduction. Business cycle shocks and LMPs explain about the same share of unemployment volatility (except for Japan, Portugal and the US).

No summary available.

This thesis deals with the formation of couples in the marriage market, and proposes as a guideline to focus on the issue of inequalities, both inter- and intra-household. The first chapter examines the role of marital preferences in the rise of income inequality between households. Edoardo Ciscato and I use US data to measure the impact of changing marital preferences on inter-household income inequality. Using structural methods, we show that if marital preferences had not changed since 1971, the Gini coefficient today would be 6% lower. In Chapter 2, I introduce the idea of bringing together the literature on matching models and that on collective models. For this purpose, Alfred Galichon, Scott Kominers and I have worked on a matching model with imperfectly transferable utility. We prove the existence and uniqueness of equilibrium in this framework. In addition, we construct two algorithms for determining the equilibrium. We show how the model can be estimated by maximum likelihood and propose an illustration. In the last chapter, I focus on resource sharing within couples. The idea is that collective models are inseparable from the marriage market, in the sense that the sharing of bargaining power is endogenous to the determination of an equilibrium in the marriage market. I discuss in depth the connection between collective models and matching models. In particular, I characterize the classes of collective models that can be integrated with the imperfectly transferable utility (ITU) matching model developed earlier. I propose a general method to estimate these models. Finally, I propose to illustrate my results on data extracted from the US PSID, and to estimate a model with private consumption, leisure and domestic work.

We develop a search model of marriage where men and women draw utility from private consumption and leisure, and from a non-market good that is produced in the home using time resources. We condition individual decisions on wages, education, and an index of family attitudes. A match-specific, stochastic bliss shock induces variation in matching given wages, education, and family values, and triggers renegotiation and divorce. Using BHPS (1991–2008) data, we take as given changes in wages, education, and family values by gender, and study their impact on marriage decisions and intra-household resource allocation. The model allows to evaluate how much of the observed gender differences in labor supply results from wages, education, and family attitudes. We find that family attitudes are a strong determinant of comparative advantages in home production of men and women, whereas education complementarities induce assortative mating through preferences.

We document changes in labour supply, wage and education by gender and marital status using the British Household Panel Survey, 1991-2008, and seek to disentangle the main channels behind these changes. To this end, we use a version of Goussé, Jacquemet, and Robin (2016)'s search-matching model of the marriage market with labour supply, which does not use information on home production time inputs. We derive conditions under which the model is identified. We estimate different parameters for each year. This allows us to quantify how much of the changes in labour supply, wage and education by gender and marital status depends on changes in the preferences for leisure of men and women and how much depends on changes in homophily.

We propose a two-step method to nonparametrically estimate multivariate models in which the observed outcomes are independent conditional on a discrete latent variable. Applications include microeconometric models with unobserved types of agents, regime-switching models, and models with misclassification error. In the first step, we estimate weights that transform moments of the marginal distribution of the data into moments of the conditional distribution of the data for given values of the latent variable. In the second step, these conditional moments are estimated as weighted sample averages. We illustrate the method by estimating a model of wages with unobserved heterogeneity on PSID data.

In this thesis, I examine the effect of uncertainty on consumption behavior in life cycle models. Although it has been recognized since the 1980s that uncertainty can substantially alter the predictions of life cycle models, some of the mechanisms involved are still poorly understood. In Chapter 1, I study the consequences of the presence of uncertainty on consumption growth, and I show that they challenge the existing belief that consumption follows a random walk in standard life cycle models. Indeed, uncertainty leads prudent households (i.e. those with convex marginal utility) to reallocate part of their present consumption to the future, which is uncertain, and thus to choose a level of present consumption lower than their expected future consumption. This implies that variables other than present consumption improve the prediction of future consumption, as they predict the precautionary difference between present and future consumption. In chapter 2, I consider the impact of uncertainty on the level of consumption of households, how it varies with their income and wealth, and thus how their consumption responds to income shocks. Since the presence of uncertainty induces households to allocate a larger share of their resources to future periods, they save more in the present period. I highlight the fact that this additional saving, called precautionary saving, varies in a decreasing and concave manner with the transitory share of income and with wealth, but in an increasing and convex manner with the permanent share of income. In chapter 3, I draw the consequences of these results for the empirical measurement of the response of consumption to income shocks. I build on the results of Chapters 1 and 2 to show that, in standard life-cycle models, consumption growth is negatively correlated with the realizations of past transitory shocks, due to precautionary behavior. Such a correlation would induce a bias in a frequently used method for estimating the response of consumption to transitory income shocks, developed by Blundell, Pistaferri and Preston (2008). Indeed, their method would attribute to present transitory shocks the changes in consumption explained by past shocks, thus predicting a too small response to transitory shocks. I generalize this method to take into account the possible influence of past shocks on consumption growth. With this more flexible estimator, I obtain that the response of consumption to transitory shocks is statistically significant and that its magnitude is consistent with the literature on the response of consumption to transitory tax cuts.

In this thesis, I analyze the role of institutions in two areas of study: the labor market and the housing market. In Chapter 1 (with Elisa Guglielminetti), I present a randomized search model to explain the selection of new hires between short and long term contracts. Mining an Italian database, we find that the probability of obtaining a permanent contract increases with a stronger match between the worker's education and occupation. In Chapter 2, I explore the effect of liberalizing fixed-term contracts and find a negative effect on new permanent jobs. The results validate the simulations of the first chapter and confirm the role of match quality in firms' hiring choices. Chapter 3 (with Robert Gary-Bobo), studies the distribution of housing and real estate wealth between generations and, by finding an imbalance in favor of the oldest, identifies taxation instruments to correct it.

We study the treatment effect of grade retention using a panel of French junior high-school students, taking unobserved heterogeneity and the endogeneity of grade repetitions into account. We specify a multistage model of human-capital accumulation with a finite number of types representing unobserved individual characteristics. Class-size and latent student-performance indices are assumed to follow finite mixtures of normal distributions. Grade retention may increase or decrease the student’s knowledge capital in a type-dependent way. Our estimation results show that the average treatment effect on the treated (ATT) of grade retention on test scores is positive but small at the end of grade 9. Treatment effects are heterogeneous: we find that the ATT of grade retention is higher for the weakest students. We also show that class size is endogenous and tends to increase with unobserved student ability. The average treatment effect of grade retention is negative, again with the exception of the weakest group of students. Grade repetitions reduce the probability of access to grade 9 of all student types.

A constructive proof of identification of multilinear decompositions of multiway arrays is presented. It can be applied to show identification in a variety of multivariate latent structures. Examples are finite-mixture models and hidden Markov models. The key step to show identification is the joint diagonalization of a set of matrices in the same non-orthogonal basis. An estimator of the latent-structure model may then be based on a sample version of this joint-diagonalization problem. Algorithms are available for computation and we derive distribution theory. We further develop asymptotic theory for orthogonal-series estimators of component densities in mixture models and emission densities in hidden Markov models.

No summary available.

We extend the search-matching model of the marriage market of Shimer and Smith (2000) to allow for labor supply, home production, match-specific shocks and endogenous divorce. We study nonparametric identification using panel data on marital status, education, family values, wages, and market and non market hours, and we develop a semiparametric estimator. We estimate how much sorting results from time use specialization or homophilic preferences. We estimate how equilibrium marriage formation affects the wage elasticities of market and non market hours. We estimate individuals’ willingness to pay for marriage and quantify the redistributive effect of intra- household resource sharing.

This paper provides methods to estimate finite mixtures from data with repeated measurements non-parametrically. We present a constructive identification argument and use it to develop simple two-step estimators of the component distributions and all their functionals. We discuss a computationally efficient method for estimation and derive asymptotic theory. Simulation experiments suggest that our theory provides confidence intervals with good coverage in small samples.

I use search models as a starting point for my analysis, examining the impact of frictions from either a theoretical or an empirical perspective. In Chapter 1 I analyze the effects of uncertainty on the macroeconomy. Empirical estimates show that uncertainty has a negative impact on the economy and that the labor market is an important transmission channel. A general equilibrium model with DMP frictions is able to reproduce the observed facts. In Chapter 2 I use a Time Varying Parameter SVAR with stochastic volatility to investigate the properties of job creation in the United States and their variation over time. The estimates indicate that the volatility depends largely on demand and price shocks. Jobs responded negatively to technology shocks until the early 1990s. Chapter 3 incorporates the spatial dimension into a search model. This allows to explain some regularities observed in Austrian data: i) the existence of a reservation frontier between wage and distance. ii) the change in job search strategy. iii) the discouraging effect of unemployment benefits. In Chapter 4 I present a model that explains the selection of new hires between short and long term contracts. Using an Italian database, we find that the probability of obtaining a permanent contract depends negatively on the degree of mismatch between the worker's education and the occupation. Moreover, reforms that liberalize fixed-term contracts encourage their use, but they have non-linear effects on the unemployment rate.

This thesis studies family formation, family organization, and parenting systems. The first two chapters deal with spousal choice and resource sharing within the couple. I first model how individuals meet and decide whether or not to pair up based on their education, income, and physical appearance. I use U.S. data where I observe who is married to whom and for how long, allowing me to identify individuals' preferences. The second chapter analyzes the effects of marriage on certain economic trends such as income inequality or labor supply. In this chapter, people in a couple share their income and choose together the best organization for working and performing domestic tasks and raising children. Using British data, I identify the income transfers that exist between men and women and show that these transfers increase the work of married men and decrease the work of married women. The last two chapters study the efficiency of the French college and in particular the practice of repeating grades. The third chapter uses decomposition methods to assess the extent to which the decline in French students' PISA scores can be attributed to changes in student characteristics or to changes in the returns to education. Finally, the last chapter focuses specifically on repetition and uses panel data on French middle school students to assess the impact of repetition on their scores.

No summary available.

No summary available.

The aim of this paper is to provide simple nonparametric methods to estimate finitemixture models from data with repeated measurements. Three measurements suffice for the mixture to be fully identified and so our approach can be used even with very short panel data. We provide distribution theory for estimators of the mixing proportions and the mixture distributions, and various functionals thereof. We also discuss inference on the number of components. These estimators are found to perform well in a series of Monte Carlo exercises. We apply our techniques to document heterogeneity in log annual earnings using PSID data spanning the period 1969 - 1998.

A constructive proof of identification of multilinear decompositions of multiway arrays is presented. It can be applied to show identification in a variety of multivariate latent structures. Examples are finite-mixture models and hidden Markov models. The key step to show identification is the joint diagonalization of a set of matrices in the same non-orthogonal basis. An estimator of the latent-structure model may then be based on a sample version of this simultaneous-diagonalization problem. Simple algorithms are available for computation. Asymptotic theory is derived for this joint approximate-diagonalization estimator.

This thesis is composed of four chapters that can be read independently. Each of the chapters focuses on a particular moment in the trajectories of French youth through the education system to their entry into the labor market and their transition to employment. Throughout these trajectories, individuals orient themselves in the school system and make occupational choices. This work focuses on understanding these choices and the impact they may have on the future of individuals. Although each of these works is autonomous, they all focus on explaining inequalities in the school and labor market of young people in France and on characterizing how these are linked. The thesis contributes to economic research by asking original questions about the individual decisions that young people make during their life course and by answering these questions with empirical methods adapted to the problem and the available data.

The aim of this paper is to provide simple nonparametric methods to estimate finitemixture models from data with repeated measurements. Three measurements suffice for the mixture to be fully identified and so our approach can be used even with very short panel data. We provide distribution theory for estimators of the mixing proportions and the mixture distributions, and various functionals thereof. We also discuss inference on the number of components. These estimators are found to perform well in a series of Monte Carlo exercises. We apply our techniques to document heterogeneity in log annual earnings using PSID data spanning the period 1969-1998.

The essays in this thesis use empirical evidence to answer two questions that are of central importance given our growing understanding of the relationship between social preferences and economic growth and welfare at the country level: the basis for prosocial behavior and the impact of policies aimed at increasing it. Levels of prosocial behavior have often been taken as fixed, but these tests provide evidence that they are susceptible to change from policy interventions. Given that there are few interventions specifically focused on trust and cooperation, there may be considerable scope for improving well-being by increasing policy focus on this issue. This is what is demonstrated in these essays. Chapter 1 discusses the foundations of pro-social behavior using different frameworks in demands for a contribution to the public good, and shows that information about the social norm is the most powerful motivating factor. Chapter 2 provides empirical and theoretical results that pro-social behavior at the community level (by contributing to local public services) depends on the expected effectiveness of that behavior. Chapter 3 provides new results on the impact of trust at the individual level, and shows that a childhood training program that increased trust (as well as improved attention and reduced delinquency), triggered a chain of events to improve long-term outcomes in terms of education, crime, and economic performance.

A first work considers the identification and estimation of the binary choice model with random coefficients and its asymptotic study. A second work gives minimax lower bounds and presents an adaptive estimator. The third paper considers a binary treatment effect model when the selection model is a binary choice random coefficient model that relaxes the monotonicity assumption. We discuss the identification and estimation of marginals, a generalization of the marginal treatment effect and the joint law of counterfactuals, conditional on the vector of unobservables in the selection equation, as well as the treatment effect parameters. The fourth work deals with the identification of the joint distribution of ex-ante and ex-post earnings in a generalized Roy model with random coefficients and uncertainty. The fifth chapter obtains confidence regions for instrumental variable estimation of a high-dimensional linear model with endogenous regressors. The method is robust to identification, to very many weak instruments, to heteroskedasticity. We study the properties of the method when the structural model is not parsimonious and obtain model selection results. We present several extensions. The last work considers the estimation of confidence intervals for wealth inequality indicators from the 2004 wealth survey where the wealth components are intervals, and from tax data. They take into account the uncertainty on the data, the parameters, and the uncertainty related to sampling and total non-response.

What role does job quality play in a frictional labor market? This is the question that this thesis, consisting of three chapters, has tried to answer, with each contribution focusing on particular frictions and different quality dimensions: on the frictional side, the impact of competition between firms for the best employees in a context of information shortage, and on the other hand the possibility that institutions such as the civil service do not allow for equal access to certain jobs, or that financial markets do not allow for efficient money investments. Regarding the quality dimensions, the contributions focus on wages and a non-monetary dimension for which data from different subcategories of job satisfaction are used. All three chapters take seriously the possibility that workers may be in the "wrong" job while they are rational and maximizing their utility in the labor market. In a market where information about available jobs arrives stochastically, one rarely finds the job that exactly matches one's preferences. If access to credit depends on initial wealth, the best entrepreneurial ideas may not be realized. Chapter (1) presents a job search model with hedonic and productive characteristics. Chapter (2) identifies the impact of a change in wealth on the demand for hedonic characteristics. Chapter (3) traces the evolution of income and its determinants in the self-employed and employed sectors in Ghana.

This thesis presents three independent research topics, linked nevertheless by the question of the identification of economic models. The first chapter is devoted to non-parametric instrumental models. I first study the completeness condition, which is recently used to identify instrumental non-parametric regressions for instance. This essay considers an additively separable non-parametric model with a broad support condition. In this framework, different versions of the completeness condition are obtained. I then consider a new method to deal with endogenous selection, based on the independence between instruments and selection variable, and the completeness condition. In addition to the identification, an estimation method and an application are proposed. The second chapter focuses on two industrial economy models. The first test considers the non-parametric identification of the common value auction model. The identifying assumption is that the support of the distribution of signals conditional on the value of the good varies with this value. The interest of this approach is that apart from this condition, it does not rely on functional restrictions. The second essay studies the adverse selection model. It shows that in the absence of exogenous contract changes, the identification of the model requires the knowledge of at least one of the model parameters. However, in the presence of such changes, the model is partially or completely identified. An estimation method and an application are also proposed. The third chapter, finally, focuses on peer effect models. While these are considered as unidentified, a slight modification of the standard linear model allows us to find the structural parameters thanks to the variations in group size. These results are extended to a binary model of interactions.

This thesis first proposes a structural evaluation of the microeconomic effects of two key policies of the Plan d'Aide au Retour à l'Emploi (PARE) implemented in France in 2001: the accompaniment and the training of job seekers. As business creation is often cited as an alternative to these policies, we then establish a model of the life cycle of self-employed workers. However, as the social cost of the measures studied is not available, a complete cost-benefit analysis could not be carried out in the framework of this work. From a methodological point of view, the structural approach is approached here as an extension of reduced form evaluations and not as their antithesis. Given the algorithmic difficulties posed by this method, we pay particular attention to the optimization techniques used and we endeavor - following Judd's recommendations - to simplify the implementation of the estimates as much as possible through the use of the AMPL software.

The transition to a market economy and accession to the European Union have profoundly changed the structure and functioning of the economies of Central and Eastern Europe. This thesis proposes an analysis of the evolutions observed in the regional and national labor markets of the new EU member states as well as an evaluation of the public policies implemented in this context of economic transition. The analysis of the matching process between workers and employers reveals the importance of labor demand in the creation of new hires in Latvia, Slovenia and Estonia and underlines the need to integrate flows (unemployed and vacant jobs) and spatial effects in the modeling. The effectiveness of public policies is documented at the macro and micro levels and demonstrates the positive influence of training programs on unemployment exit rates and the employability of participants.

This thesis introduces two new classes of nonlinear state-space models and establishes algorithms for their estimation in a Bayesian framework. The first class is a multivariate generalization of Markov regime-switching models in the sense that it allows the incorporation of several latent variables. This methodology is illustrated by an application in finance where a multivariate stochastic volatility model with several discrete latent variables is estimated on portfolio returns. The second class, called quartic state-space models, allows the study of the dynamics of bimodal laws in a very general framework. An illustration on a longitudinal basis of gross domestic products shows the evolution of the two modes of world income distribution and characterizes the convergence between different groups of countries.

This thesis highlights the impact of the massification of education in the 1980s on the rate of return to education in France from 1983 to 2002. In addition, it provides an update on the progress of econometric techniques that have been developed over the last fifty years. We explain why it now seems essential to take into account, in the various formalizations, the implicit and explicit costs of choosing to study and the probability of unemployment which differs according to the years of education. Repetition and the specificity of the French education system are elements that lead us to believe that "years of study" are not representative of the level of individuals and that the explanatory variable "diploma" is more appropriate. Taking these different factors into account, the rate of return on the baccalauréat has fallen significantly over the period and is now negative at around -1%, whereas the rate of return on the post-bac years has converged towards 10%, with the licence still being more profitable than the master's degree. On the other hand, the problem of the downgrading of the general baccalaureate is increasing. The use of surveys that are richer in individual information and information on family background shows that they also condition academic success. The question of whether or not to continue studying was raised. Many students have continued their studies when they should not have done so, and vice versa. Sensitivities and taste for education have a strong impact on this choice but are typically unmeasurable qualitative variables that escape econometrics.

This work brings together four essays devoted to the study of heterogeneity and individual dynamics in the labor market. The first chapter highlights the link between wage mobility (or inertia) and the degree of persistence of inequality. We use a simple and original statistical method to study individual wage trajectories, and apply it to French data covering the period 1990-2002. We find that the recession of the early 1990s was associated with a significant increase in longitudinal inequality. In the second chapter, we study the effect of job mobility on the correlations between wages and non-wage characteristics. In our model, strong preferences for these characteristics do not necessarily translate into negative correlations if mobility frictions are important. On European data, we estimate strong preferences for certain characteristics such as the type of work or job security, as well as very small wage differentials between levels of amenity. Chapters 3 and 4 introduce a method for modeling unobserved heterogeneity: independent component analysis. This differs from principal component analysis in that the factors are not simply assumed to be uncorrelated, but statistically independent. This assumption allows the factors to be identified in an unambiguous way. We apply our method to education wage data in France for the year 1995. Our results suggest a complex and multidimensional relationship between educational attainment and labor market performance.

This thesis is a contribution to the study of the possible links between wage dispersion and firm heterogeneity from a theoretical and empirical perspective. It is composed of five chapters. The first three chapters follow the framework of the employment research models developed by Burdett and Mortensen (1999), in which wage dispersion is endogenous, without the addition of other heterogeneity. The first chapter proposes two extensions to this model: on the one hand, it considers the possibility for firms to differentiate on their choice of capital, and on the other hand, it endogenizes the arrival rate of job offers to employees. The second chapter estimates the model developed in the first chapter for certain sectors and concludes that the heterogeneity generated by the theoretical model is insufficient, compared with the heterogeneity observed in the data. The third chapter extends the model constructed in chapter 1 to the possibility for firms to recruit part-time workers. Chapters 4 and 5 characterize the empirical differences in wage policies between firms using matched firm-employee data over 20 years. It highlights and characterizes the extreme heterogeneity of their practices. Chapter 5 examines in particular the link between wage dispersion and productivity, interpreted in relation to the organization of firms.

The right to housing is an essential value of our society. Its purpose is to help the most disadvantaged people to benefit from a home offering a minimum of comfort. Economic policies have been put in place to implement this right, such as low-income housing and subsidized loans to encourage access to property. This thesis aims to contribute to the evaluation of their effectiveness. It first determines to what extent economic policies have influenced the housing choices of disadvantaged households in order to guarantee them decent living conditions. It then examines the macroeconomic consequences of such policies. Indeed, the public rental sector and homeownership are sources of residential inertia and may therefore limit the ability of migration to address local labor market imbalances. In particular, the lack of worker mobility contributes to local wage disparities, which are discussed in more detail.

After a presentation of the evolution of cold drink consumption from 1950 to 1997 as shown by national accounts figures and food surveys of the National Institute of Statistics and Economic Studies, the author focuses his study on the year 1997. The data used for the empirical estimates come from a panel of households residing in France and monitored daily. In the first chapter, the author presents a descriptive analysis of 35 groups of beverages covering all cold drinks and characterizes the buyers and non-buyers. In the second chapter, he estimates an almost ideal quadratic demand system to analyze the demand for beverages divided into nine distinct groups. Price elasticities are calculated using unit values instead of prices, as is usually done. As a result, these elasticities reflect both a quantity effect and a quality effect. In the third chapter, a recent method that allows the quality effect to be separated from the quantity effect and the corresponding elasticities to be calculated is implemented. The observed substitutions between drinks highlight situations of use: drinks drunk with meals, as aperitifs, for conviviality or for quenching thirst. Finally, the fourth chapter proposes a supply model to explain household storage behavior. The model assumes that the current price and the price of the product at the last purchase are sufficient.

This thesis is a contribution to the study of equilibrium job search models. It is composed of three chapters. The first one is devoted to a review of the literature on the subject, the next two present the models we have developed which incorporate features already present in the first models of this kind. In particular, they allow an employee to make job-to-job transitions. Two forms of heterogeneity are considered: firm and worker heterogeneity. Chapter II considers only the case of heterogeneous firms while the last chapter considers both forms of heterogeneity in the same market. The equilibrium of the economy is analyzed in detail. It is shown that, when the distribution of firm productivity is continuous, there is a bijective relationship between the wage offered by a firm and labor productivity. Particular attention is paid to the properties of the equilibrium wage distributions induced by these models. Indeed, one of the drawbacks of the previously developed models is to generate a cross-sectional wage distribution that is incompatible with the observed distributions. We show that firm heterogeneity is necessary to obtain a good fit. Other qualitative results concern the influence of the shape of the tails of the productivity distribution on those of the wages offered. Finally, we take into account the introduction of a legal minimum wage and derive necessary and sufficient conditions for the wage distribution to peak at this minimum wage. We develop multi-stage semi-parametric structural estimation methods of these models, based on an inversion of the wage-productivity function. We use these methods to estimate these models on French data from the INSEE employment survey.