Firm Heterogeneity in Skill Returns

Are specific skills rewarded differently across firms? If so, is this reflected in worker sorting across employers? We use administrative data linking employers and employees in conjunction with measures of cognitive and non-cognitive skills to examine these questions. We find that returns are very heterogeneous across firms. Workers respond by selecting employers that better reward their bundle of skills. Worker-firm skill complementarities contribute to the thick right tail in the distribution of earnings, pushing up both average labor market returns and their dispersion.


Abstract. This paper presents new evidence on worker–firm complementarities. We combine matched employer-employee data with direct measures of workers’ cognitive and noncognitive skills, and propose an empirical approach that separately identifies the firm-level return for each attribute. We find that similar skills command different returns across employers and that workers’ sorting into firms depends on returns to both attributes. We derive theoretical restrictions that characterize many-to-one matching in employer-employee data, linking within-firm skill dispersion to between-firm differences in average skills. Estimates support these restrictions. Firm heterogeneity in skill returns raises both the average level and dispersion of earnings.


  title={Firm Heterogeneity in Skill Returns},
  author={Boehm, Michael J. and Esmkhani, Khalil and Gallipoli, Giovanni},
  institution={Discussion Paper}

Consumption and Income Inequality across Generations

How much of the cross-sectional dispersion of income and consumption can be accounted for by parental heterogeneity and family background? How strong are intergenerational linkages? We examine data on expenditures and income of parent-child pairs and document the presence of significant family persistence in earnings, consumption, saving behaviours, and marital sorting patterns. However, we also show that idiosyncratic (family independent) heterogeneity has a quantitatively bigger role than parental effects for the evolution of cross-sectional inequality.


Abstract. We characterize the joint evolution of cross-sectional inequality in earnings, other sources of income, and consumption across generations in the U.S. To account for cross-sectional dispersion, we estimate a model of intergenerational persistence and separately identify the influences of parental factors and of idiosyncratic life-cycle components. We find evidence of family persistence in earnings, consumption and saving behaviours, and marital sorting patterns. However, the quantitative contribution of idiosyncratic heterogeneity to cross-sectional inequality is significantly larger than parental effects. Our estimates imply that intergenerational persistence is not high enough to induce further large increases in inequality over time and across generations.


  title={Consumption and Income Inequality across Generations},
  author={Gallipoli, Giovanni and Low, Hamish and Mitra, Aruni},
  institution={Discussion Paper}

Match Quality and Contractual Sorting

We ask whether match-specific quality has an effect on the type of contractual arrangements that firms offer to workers. We present evidence that contractual arrangements depend on match quality and that heterogeneity in pay mechanisms has a significant effect on employment durations and wage dynamics.

LINK TO PAPER (PDF)             DATA and CODE

Abstract. This paper examines the impact of match-specific heterogeneity on compensation arrangements. In a stylized contractual choice problem, we show that employers may have an incentive to offer performance-based contracts when match-specific productivity is high. We test the empirical content of this hypothesis using the NLSY79, which contains information about individual job histories and performance pay. We find that better match quality does affect pay arrangements, employment durations, and wage cyclicality. Direct evidence on the accrual of job offers to workers lends support to the hypothesis that employers use performance-related compensation to preserve high-quality matches.


  title={Match Quality and Contractual Sorting},
  author={Galindo da Fonseca, Joao and Gallipoli, Giovanni and Yedid-Levi, Yaniv},
  journal={Labour Economics}

Permanent-Income Inequality

We estimate the distribution of lifetime wealth portfolios and permanent-income inequality using microdata for 1989-2016. We show that households at the top of the asset distribution have not increased their share of human wealth. Instead, the higher concentration of permanent income over the past decades is due to the growing importance of assets, as opposed to human capital, in lifetime wealth portfolios.
Permanent income is tightly linked to consumption expenditures. We estimate an average propensity to consume out of permanent-income of roughly 0.8; this propensity to consume varies significantly in the cross-section of US households, changing with both the value and composition of each household’s lifetime wealth portfolio.


Abstract. We characterize the distribution of permanent-income and quantify the value of assets and human capital in lifetime wealth portfolios. We estimate the distribution of human wealth using nonparametric identification results that allow for state-dependent stochastic discounting and unobserved heterogeneity. The approach imposes no restrictions on income processes or utility. Accounting for the value of human capital delivers a different view of inequality: (i) in 2016 the top 10% share of permanent-income was 1/3 lower than the corresponding share of assets; (ii) however, since 1989, the top 10% share of permanent-income has grown much faster than the corresponding share of assets. Human wealth has a mitigating influence on inequality, but this effect has waned over time due to the growing importance of assets in lifetime wealth portfolios. We find that consumption expenditures are tightly linked to permanent-income; however, liquidity constraints can lead to substantial deviations below permanent-income.


  title={Permanent-Income Inequality},
  author={Abbott, Brant and Gallipoli, Giovanni},
  institution={Discussion Paper }

Human Capital Inequality: Empirical Evidence

Heterogeneity in human capital is a key source of differences in economic well-being. This article provides a synopsis of the empirical approaches that have characterized the analysis of human capital inequality over the past few decades.


Abstract. Wealth inequality has received considerable attention, with mounting evidence of steady and economically meaningful changes in the concentration of wealth ownership. By definition, wealth inequality captures disparity in the ownership of productive capital and other non-labor factors of production. In contrast, in this article we focus on the distribution of human capital and its implications for the accrual of economic resources to individuals and households. Human capital inequality can be thought of as measuring disparity in the ownership of labor factors of production, which are usually compensated in the form of wage income.


  title={Human Capital Inequality: Empirical Evidence},
  author={Abbott, Brant and Gallipoli, Giovanni},
  journal={Oxford Research Encyclopedia of Economics and Finance},

Education Policy and Intergenerational Transfers in Equilibrium

We examine the equilibrium effects of alternative financial aid policies intended to promote college participation. We model a rich environment featuring various dimensions of heterogeneity and intergenerational linkages.


Abstract. This paper examines the equilibrium effects of alternative financial aid policies intended to promote college participation. We build an overlapping generations life cycle model with education, labor supply, and consumption/saving decisions. Cognitive and non-cognitive skills of children depend on the cognitive skills and education of parents and affect education choice and labor market outcomes. Driven by both altruism and paternalism, parents make transfers to their children which can be used to fund education, supplementing grants, loans, and the labor supply of the children themselves during college. The crowding out of parental transfers by government programs is sizable and thus cannot be ignored when designing policy. The current system of federal aid is valuable: removing either grants or loans would each reduce output by 2% and welfare by 3% in the long-run. An expansion of aid towards ability-tested grants would be markedly superior to either an expansion of student loans or a labor tax cut. This result is, in part, due to the complementarity between parental education and ability in the production of skills of future generations.


 title={Education Policy and Intergenerational Transfers in Equilibrium},
 author={Abbott, Brant and Gallipoli, Giovanni and Meghir, Costas and Violante, Giovanni Luca},
 journal={Journal of Political Economy}, 
volume = {126},
number = {6},
pages = {2569-2624},

Markov-Chain Approximations for Life-Cycle Models

Non-stationary income processes are standard in quantitative life-cycle models. Their approximation is key for numerical implementations. We develop methods to apply standard discretization algorithms within non-stationary life-cycle settings and assess their relative performance. In one extension we examine income processes in which shocks to earnings are modeled as draws from a mixture of Normal distributions and describe simple and tractable approaches to numerically describe non-Normal earnings distributions.


Abstract. Non-stationary income processes are standard in quantitative life-cycle models, prompted by the observation that within-cohort income inequality increases with age. This paper generalizes Tauchen (1986), Adda and Cooper (2003), and Rouwenhorst’s (1995) discretization methods to non-stationary AR(1) processes. We evaluate the performance of these methods in the context of a canonical life-cycle,income-fluctuation problem with a non-stationary income process. We also examine the case in which innovations to the persistent component of earnings are modeled as draws from a mixture of Normal distributions. We find that the generalized Rouwenhorst’s method performs consistently better than the others even with a relatively small number of states.


  title={Markov-Chain Approximations for Life-Cycle Models},
  author={Fella, Giulio and Gallipoli, Giovanni and Pan, Jutong},
  journal={Review of Economic Dynamics},
  volume = {34},
  pages = {183-201}

Piercing the “Payoff Function” Veil: Tracing Beliefs and Motives

We develop a non-intrusive approach to experimentally gauge the way agents make their choices. We document the presence of significant heterogeneity in the decision-making process of agents within an interactive setting. The setting is designed to elicit beliefs and motives of agents.


Abstract. This paper develops an experimental methodology that allows the identification of decision-making processes in interactive settings using tracking of non-choice data. This non-intrusive and indirect approach provides essential information for the characterization of beliefs. The analysis reveals significant heterogeneity, which is reduced to two broad types, differentiated by the importance of pecuniary rewards in agents’ payoff function. Most subjects maximize monetary rewards by best responding to beliefs shaped by recent history. Others are able to identify profit-maximizing actions but choose to systematically deviate from them. The interaction among different types is key to understanding aggregate outcomes.


  title={Piercing the ``Payoff Function'' Veil: Tracing Beliefs and Motives},
  author={Fenig, Guidon and Gallipoli, Giovanni and Halevy, Yoram},
  note={Working Paper},
  institution={University of Toronto}

Structural Transformation and the Rise of Information Technology

We use a task-based IT intensity index to: (1) document the evolution of earnings, employment, and productivity in low and high tech industries; (2) estimate the elasticity of substitution in production between IT and non-IT jobs; (3) revisit Solow’s Paradox, finding evidence of positive occupation-level effects of IT intensity on productivity growth.

LINK TO PAPER (PDF)          Online Appendix (PDF)          DATA and CODE

Abstract. Has the emergence of information technology changed the structure of employment and earnings in the US? We propose a new index of occupation-level IT intensity and document several long-term changes in the occupational landscape over the past decades. Using Census and US KLEMS micro-data, we show that: (i) the bulk of productivity growth after 1950 is concentrated in IT-intensive sectors; (ii) the share of workers in IT jobs has expanded significantly, with little or no pause and IT jobs enjoy a large and growing earnings premium, even after controlling for general task requirements (e.g., cognitive, non-routine); and (iii) the rise of the IT-intensive employment share is closely associated with declines in the manufacturing employment share. While earnings premia for college-educated and cognitive/non-routine workers have flattened in the aggregate since 2000, we show that they continued growing in IT-intensive jobs and that these jobs have played a key role in accounting for the surge of high tech service labor productivity. We also use our IT intensity index to estimate industry-specific elasticities of substitution between IT and non-IT intensive labor, finding values of 1.6 in manufacturing and 1.3 in services. Finally, we revisit a long-standing question about the relationship between technological progress and productivity and provide evidence that occupation-level IT intensity is positively associated with output growth, especially in the services sector.


 title={Structural Transformation and the Rise of Information Technology},
 author={Gallipoli, Giovanni and Makridis, Christos A.},
 journal={Journal of Monetary Economics},
 note = {Carnegie Rochester NYU Series on Public Policy}

The Costs of Occupational Mobility: An Aggregate Analysis

We define a measure of job distance based on task content and quantify its importance for occupational mobility. We estimate the size of different layers of mobility costs.


Abstract. We estimate the costs of occupational mobility and quantify the relative importance of differences in task content as a component of total mobility costs. We use a novel approach based on a model of occupational choice which delivers a gravity equation linking worker flows to occupation characteristics and transition costs. Using data from the Current Population Survey and the Dictionary of Occupational Titles we find that task-specific costs account for no more than 15% of the total transition cost across most occupation pairs. Transition costs vary widely across occupations and, while increasing with the dissimilarity in the mix of tasks performed, are mostly accounted for by task-independent occupation-specific factors. The fraction of transition costs that can be attributed to task-related variables appears fairly stable over the 1994-2013 period.


  title={The Costs of Occupational Mobility: an Aggregate Analysis},
  author={Cortes, Guido Matias and Giovanni Gallipoli},
  journal={Journal of the European Economic Association},
  pages = {275-315},
  number = {2},