This DPhil thesis is a collection of three empirical papers that study the role of firms in the UK labour market. Each chapter focuses on firms at different points in their lifecycle. Young firms are an engine of job creation but little is known about the quality of the jobs that they offer. In Chapter 1, I use a matched employer-employee dataset to study how starting wages and lifecycle earnings of employees differ between young and mature firms. I find that young firms pay a small premium to new hires, but subsequent wage growth is better at mature firms, both within continuing job matches and when individuals change jobs. Crucially, highly-paid and stable jobs at young firms have become increasingly rare over time, as young firms themselves have become less likely to survive and attain high productivity levels - both in absolute terms and relative to mature firms over the same period. Policies that aim to stimulate job growth by encouraging the formation of new firms should therefore pay close attention to the types of firms that form. Chapter 2 asks what determines the proportion of a firm's income that workers receive as compensation. I use longitudinal firm data from a period of substantial labour share variation to understand the firm-level determinants of the labor share of income - a question that has typically only been addressed with country- and sector-level data. Estimating a dynamic model using GMM, I find that firms with greater market power and a higher ratio of capital to labour allocate a smaller proportion of their value added to workers. Testing the impact of tangible and intangible capital on low- and high-wage firms leads to conclusions consistent with the hypothesis of capital-skill complementarity. Overall, the results suggest that firm-level drivers play a key role in the evolution of the aggregate labour share, which has declined significantly since the 1970s. Chapter 3 co-authored with Brian Bell, focuses on mature firms and asks how wages at such firms respond to idiosyncratic firm-level cost shocks. We create a unique dataset that links longitudinal data on workers' compensation to the unexpected costs related to firms' legacy defined benefit pension plans. We show that firms are able to share the burden of such costs when a significant share of their workers are current or former members of the plan. We also find that firms that respond to deficits by closing down the pension plans effectively reduce the total compensation of plan members. These results point to significant frictions in the labour market, which we show are a direct result of the pension arrangement that workers have. Yet closing schemes has an implicit cost for firms, since it reduces the frictions that workers face, and increases mobility.

This thesis consists of four separate chapters which address different aspects of inequality and income determination. The first three chapters are country-level studies which examine (1) how incomes are shaped by spatial price differences, (2) the factor income composition, and (3) enterprise size. The final chapter analyses how income inequality changed at the global level. The first chapter investigates the implications of regional price differences for earnings differentials and inequality in Germany. I combine a district-level price index with administrative earnings data from social security records. Prices have a strong equalising effect on district average wages in West Germany, but a weaker effect in East Germany and at the national level. The change in overall inequality as a result of regional price differences is small (although significant in many cases), because inequality is mostly explained by differences within rather than between districts. The second chapter is motivated by the rapid increase in top income shares in the United States since the 1980s. Using data derived from tax filings, I show that this pattern is very similar after controlling for changes in tax unit size. Over the same period as top income shares increased, the composition of these incomes changed dramatically, with the labour share rising. Using a non-parametric copula framework, I show that incomes from labour and capital have become more closely associated at the top. This association is asymmetric such that top wage earners are more likely to also receive high capital incomes, compared with top capital income recipients receiving high wages. In the third chapter, I investigate the positive cross-sectional relationship between enterprise size and earnings using panel data from Ghana. I find evidence for a significant firm size effect in matched firm-worker data and a labour force panel, even after controlling for individual fixed effects. The size effect in self-employment is stronger in the cross-section, but it is driven by individual time-invariant characteristics. The final chapter studies the global interpersonal income distribution using a newly constructed and improved database of national household surveys between 1988 and 2008. The chapter finds that the global Gini remains high and approximately unchanged at around 0.7. However, this hides a substantial change in the global distribution from a twin-peaked distribution in 1988 into a single-peaked one now. Furthermore, the regional composition of the global distribution changed, as China graduated from the bottom ranks. As a result of the growth in Asia, the poorest quantiles of the global distribution are now largely from Sub-Saharan Africa. By exploiting the panel dimension of the dataset, the analysis shows which decile-groups within countries have benefitted most over this 20-year period. In addition, the chapter presents a preliminary assessment of how estimates of global inequality are affected by the likely underreporting of top incomes in surveys.

This thesis consists of a short introduction and three self-contained analytical chapters. Chapter 1 focuses on the question of learning gaps and divergence in achievement across countries. I use unique child-level panel data from Ethiopia, India, Peru and Vietnam to ask at what ages do gaps between different populations emerge, how they increase or decline over time, and what the proximate determinants of this divergence are. I document that learning gaps between the four countries are already evident at the age of 5 years and grow throughout the age trajectory of children, preserving country ranks from 5 to 15 years of age. At primary school age, the divergence between Vietnam and the other countries is largely accounted for by substantially greater learning gains per year of schooling. Chapter 2 focuses on learning differences between private and government school students in India. I present the first value-added models of learning production in private and government schools in this context, using panel data from Andhra Pradesh. I examine the heterogeneity in private school value-added across different subjects, urban and rural areas, medium of instruction, and across age groups. Further, I also estimate private school effects on children's self-efficacy and agency. I find modest or insignificant causal effects of attending private schools in most test domains other than English and on children's academic self-concept and agency. Results on comparable test domains and age groups correspond closely with, and further extend, estimates from a parallel experimental evaluation. Chapter 3 uses panel data from the state of Andhra Pradesh in India to estimate the impact of the introduction of a national midday meal program on anthropometric z-scores of primary school students, and investigates whether the program ameliorated the deterioration of health in young children caused by a severe drought. Correcting for self-selection into the program using a non-linearity in how age affects the probability of enrollment, we find that the program acted as a safety net for children, providing large and significant health gains for children whose families suffered from drought.

This thesis is an investigation into the microeconomic mechanisms that govern some of the occupational choices faced by workers in Sub-Saharan Africa, and into the monetary and non-monetary returns to their decisions. Chapter 1 begins by exploring the decision process that leads workers to allocate themselves to different occupations within the economy. In particular, I investigate the role of risk-aversion in the allocation of workers between formal and informal jobs in Ghana, hence attempting to explain a fundamental dimension of duality through an investigation into workers' preferences. In my model of sectoral allocation risk-averse workers can opt between entering the free-entry informal sector and queuing for formal occupations. Conditional on identifying the riskier option, the model yields testable implications on the relationship between risk-aversion and workers' allocation. My testing strategy proceeds in two steps. First, using the first three waves of the Ghana Household Urban Panel Survey (GHUPS) dataset, I estimate expected income uncertainty and find it considerably higher in the informal sector than in formal employment. Second, using experimental data to elicit risk-attitudes I estimate the effect of risk-aversion on occupational choices and I find that, in line with the first result, more risk-averse workers are more likely to queue for formal jobs and less likely to be in the informal sector. The conclusion of the first chapter is that attitudes to risk should feature more prominently in models of sector allocation and in the design of labour market policies, in particular when those policies aim to impact workers' vulnerability to risk and uncertainty. Chapter 2 focuses on the largest occupational category in the Developing world, self-employed workers with small productive activities, and it tries to estimate the returns to different productive assets, namely physical capital, labour and human capital. These are the workers that form most of the informal sector analysed in chapter 1, which allows me to draw a direct link with the analysis so far. The chapter begins by specifying a model for the income-generating process grounded in the literature on firms' production and hence abridging the gap between the analysis of individual earnings and the study of firms' value added. Identification in the empirics is achieved by means of panel estimators that are suitable to address the endogeneity of input choices, which derives from both time-varying and time-invariant unobservable heterogeneity. The use of these estimators is made feasible by the length of the Ghanaian Household Urban Panel Survey dataset at CSAE. I also explore issues of endogeneity in the selection of different technologies, defined by their relative capital and labour intensity. Finally, I analyse the shape of returns to capital, with the aim to detect potential non-convexities in technology. The results show that capital and work-experience play the strongest role in income-generation, while the shares of value added attributed to labour and to formal schooling are low. Marginal returns to investment are high at low capital levels and they decrease very rapidly, pointing against the existence of non-convexities due to minimum scale requirements, but implying that real income gains resulting form micro-investment are modest. Chapter 3 returns to the issue of earnings uncertainty and risk-aversion explored in Chapter 1, but it now takes the allocation choice as given and explores the direct welfare implications of income uncertainty for worker's well-being. Namely, the chapter explores the relationship between income and welfare, with a particular attention on the link between income vulnerability and happiness. Using unique longitudinal data on life-satisfaction and labour market outcomes, I estimate an individual measure of vulnerability (defined as the probability of falling below a low-income threshold) and investigate its effect on well-being. After controlling for unobservable individual fixed effects, work-satisfaction, relative income and other relevant worker characteristics, I find a sizable impact of vulnerability, over and above the income effect. When I explore the mechanisms behind my results, I find that aspiration adaptation to current income may result in a transitory income effect. Moreover, using my direct measure of attitudes to risk from field-experiments (already used in chapter 1), I can test directly the hypothesis that more risk-averse agents suffer more heavily from a given increase in income vulnerability. Overall, my findings support policy interventions that aim to reduce vulnerability, as I expect such policies to have a 'direct' impact on agents' happiness given the prevailing attitudes to risk and uncertainty in the population. Finally, from the point of view of overall social welfare, my results suggest that non-Rawlsian growth models, whereby 'someone may be left behind', may fail to enhance general welfare, for high enough levels of risk-aversion in the population, if the risk of falling behind is sufficiently widespread.

Varieties of Capitalism theory predicts that the skill specificity of workers determines their demand for social protection. In this thesis, I test this assumption using a measure of occupational mobility between pre- and post-unemployment, which I apply to European workers in different skill groups as defined by Fleckenstein et al., (2011). Using this measure as an indicator of the portability of workers' skills, I then evaluate whether the lower marketability of human capital investments is associated with greater demand for unemployment protection. The findings demonstrate that whilst this relationship is apparent in certain countries, notably Coordinated Market Economies such as Germany, the assumptions do not apply across institutional settings. Consequently, skill specificity cannot explain variation in attitudes towards unemployment protection policies between countries.