For Faculty

Characterizing Capital Amplification 

Hannes Malmberg 

This project characterizes how capital accumulation amplifies output effects of productivity improvements in a general class of models. Using a general dynamic model with multiple investment goods, my results show that the steady state of the dynamic economy is isomorphic to the equilibrium in a distorted static economy where capital is an intermediate input subject to a tax. The implicit tax is given by the ratio of the subjective discount rate and the depreciation rate; it captures that, from a static perspective, time discounting looks like underinvestment. Using this isomorphism, I characterize capital amplification using the static misallocation framework from Baqaee and Farhi (2020). My current results show that upstream investment goods have an outsized effect on long-run output compared to their Domar weights, and that the long-run output effect of human capital improvements exceeds that predicted by Mincerian returns if skilled labor is complementary to investment.

A Structural Meta-Analysis of Welfare Reform Experiments and Their Impacts on Children 

Joseph Mullins 

The project proposes a novel quantitative method to aggregate evidence from multiple empirical studies using an economic model. Information is combined across studies to learn about economic parameters, which enables both policy and prediction in an environment that is robust to the usual critiques (Marschak, 1953; Lucas, 1976). The method is applied to a number of welfare-to-work experiments conducted in the United States in the 1990s. These experiments are aggregated to learn about labor supply elasticities and the effect of household income and childcare arrangements on child skill outcomes. 

Acquisition of Tier 3 UK Biobank data

Aldo Rustichini 

In the paper we study possible consequences of historical and future expansions in a stochastic, general equilibrium Roy model where tertiary educational attainment is determined by cognitive ability and socioeconomic disadvantage from low socioeconomic status. 

On the Nature of Self-Employment
Anmol Bhandari

Labor can be broadly thought of as being rewarded in two forms – returns to time in paid employment and returns to time in self-employment. While there is an extensive literature on factors driving returns to paid employment, little is known about self-employment returns. A major lim iting factor has been an absence of comprehensive panel data for self-employed individuals. We plan to use administrative data from tax filings of individuals and businesses to fill in this gap. After we construct the panel, we will document how returns to self-employment differ from re turns to paid employment. Our findings will shed light on several questions concerning drivers of entrepreneurship as well as income profiles of business owners.

A Structural Meta-Analysis of Welfare Reform Experiments and Their Impacts on Children
Joseph Mullins

The project proposes a novel quantitative method to aggregate evidence from multiple empirical studies using an economic model. Information is combined across studies to learn about economic parameters, which enables both policy and prediction in an environment that is robust to the usual critiques (Marschak, 1953; Lucas, 1976). The method is applied to a number of welfare-to-work experiments conducted in the United States in the 1990s. These experiments are aggregated to learn about labor supply elasticities and the effect of household income and childcare arrangements on child skill outcomes.

Individual Differences and Collective Outcomes
David Rahman

This project will undertake experiments in line with the standards of experimental economics and psychology, designed to measure the extent to which individual personality traits affect collective outcomes in various economic settings. The specific setting consists of a small number of subjects jointly deciding how to make dynamic investment decisions over time. The treatments considered include manipulation of personality traits of subjects within a team, status, and ’skin in the game,’ among others. After finishing the design of experiments and obtaining IRB approval, we will code and run the pilot experience, then analyze the data and complete a first draft of a paper summarizing our findings.

The Macroeconomics of Intensive Agriculture
Hannes Malmberg

In poor countries, the agricultural sector is large, uses few inputs per worker, and is unproductive, even relative to the local economy. We jointly study these phenomena using a neoclassical two-sector model. To inform our choice of an agricultural production function, we draw on micro and macro sources to construct a cross-country database of agricultural input use and prices. Using our framework, we separately analyze the consequences of closing the sectoral TFP gap in agriculture and non-agriculture between a poor country and the US. We find that closing the agricultural TFP gap reduces GDP per worker gap from a factor of 40 only to a factor of 14.1, whereas closing the non-agricultural TFP gap reduces the GDP per worker gap to 2.5. This finding is in stark contrast to those of simple decomposition exercises, which suggest that closing agricultural labor productivity gaps could dramatically reduce income differences. The difference emerges from general equilibrium forces. In general equilibrium, agricultural TFP improvements have diminishing marginal returns since rising incomes and falling agricultural prices reduce the importance of the agricultural sector. In contrast, non-agricultural TFP improvements have increasing marginal returns, as income effects and agricultural input intensification both work to increase the importance of the non-agricultural sector.

Our results show the importance of taking a general equilibrium perspective on agricultural labor productivity when intermediate inputs and machinery are produced in other sectors. Furthermore, they show that even when agriculture is relatively unproductive, the main impediment to development may lie outside the agricultural sector.

A General Method for Structural Estimation using Estimated Policy Functions
Joseph Mullins

I introduce a new method for estimating the parameters of economic models using observed choice behavior. My proposal builds on frontier methods in the literature by nesting those techniques within a general class. I show that this general class of estimator admits a new procedure, with equivalent asymptotic properties, that improves upon existing trade-offs between computational simplicity, finite sample performance, and asymptotic efficiency. Current results show that in finite settings, both in theory and in practice through simulation exercises, that the method offers a tradeoff between finite sample precision and computational speed. I propose to develop extensions for the case in which states and actions are continuous.

Efficient Risk Sharing with Ordinal Recursive Utilities
Jan Werner

Recursive utilities play an important role in macroeconomics and finance. Their main advantage over the standard discounted expected utility is that they discard time separability. In particular, the popular Epstein-Zin (EZ) recursive utility has been extensively used in asset pricing to explain long standing empirical puzzles such as the equity premium puzzle and the risk-free rate puzzle. The EZ recursive utility allows for separate parametrization of risk aversion and elasticity of intertemporal substitution. Another important advantage of recursive utilities is that they can be easily used in dynamic programming. Despite the importance and the prevalence of recursive utilities, their properties and implications are not yet well understood. Are they a minor - yet more flexible - variation of discounted expected utility, or a drastic departure thereof? The case in point of this proposal is efficient risk sharing. Discounted expected utility (with risk aversion and common beliefs) leads to strong implications in regard to risk sharing in dynamic as well as static settings. If there is no aggregate risk - conditional on past information, or unconditional - at any date, then there is no individual risk in any efficient allocation. More generally and with aggregate risk, efficient risk sharing is comonotone, that is, every individual plan is comonotone with the aggregate. Little is known for recursive utilities. Under what conditions do these results extend to recursive utilities, and when do they fail to hold? Can there be “betting” against each other with no aggregate risk? Are there conditions on intertemporal correlation or comonotonicity of aggregate resources that lead to simple rules of risk sharing? Good understanding of risk sharing is important for asset pricing even in models with representative agent. With this project we aim to go beyond basic insights into the nature of risk sharing with recursive utilities.

Resource Allocation with Market Incompleteness and Financial Frictions
Anmol Bhandari

We propose to study how an economy allocates resources for production and consumption in presence of market incompleteness and financial frictions. Our hypothesis is that factor misal- location is closely tied to the risk-sharing avenues available to firm owners. In contrast to the commonly studied bond-only economy with collateral constraints (for example Moll (2014)), the degree of misallocation can be increasing in the persistence of the idiosyncratic risk when firms have access to state-contingent contracts. The possibility to transfer wealth from high produc- tivity states to low productivity states allows firm owners to trade off the efficient allocation of consumption against the efficient allocation of capital.

The Missing Middle Managers: Labor Costs, Firm Structure, and Development
Hannes Malmberg

In this project, we assess whether the cost of managers, business
professionals, and administrators (MBAs) in poor countries is an important deterrent to the adoption of large, hierarchical, multi-establishment firm structures. We provide new empirical evidence drawing on a database from a compensation consulting firm that advises multinationals on pay at their local administrative support affiliates worldwide. The database includes annual total compensation for more than 170,000 country-year-firm-job observations. Our main empirical finding is that the level of compensation is in line with GDP per worker in rich countries but an order of magnitude higher in the poorest; the elasticity of compensation with respect to GDP per worker is only one-third. We aim to quantify the importance of this finding using an appropriate technology model where firms in each country and industry choose between small and large firm structures. Large firms offer productivity gains but require the use of MBAs to coordinate and implement production.

Tax Compliance of Intangible-Intensive Businesses
Ellen McGrattan

A challenge for tax policy research is to accurately assess tax noncompliance of businesses in light of the fact that their owners make substantial investments in intangible assets that are not easily
measured and generate incomes that are not easily taxed. Intangible assets include research and development (R&D), software, brands, designs, organizational capital, non-compete agreements, client lists, customer bases, and goodwill. Unlike investment in plant and equipment, investment in intangible assets affords greater opportunities to misreport incomes and valuations. In previous work with Ed Prescott and more recently with Anmol Bhandari, I have developed theory-based frameworks to measure the aggregate stock of U.S. business intangible assets for both large corporations and private businesses. Using these theories, my coauthors and I find that intangible assets are large and a crucial input to theoretical tax policy analysis. This proposal includes three next-step projects building on this research, using administrative data from the Internal Revenue Service.

1. The goal of the first project is to develop theory and statistical tools suitable to study the economic impact of changes in corporate tax policy, allowing for the fact that tax rates differ across geographical jurisdictions where income is booked, legal forms of organization, and types of income, introducing many tax arbitrage possibilities.
2. The goal of the second project is to develop theory and statistical tools suitable to study the economic impact of tax changes for private businesses that tend to be much smaller in size than multinationals, hold different investment portfolios, and present different challenges for the IRS as far as tax noncompliance is concerned.
3. In addition to theory development and evaluation, we are interested in generating new tabulations that can be replicated and produced on a regular basis. Currently, the BEA aggregates information on C and S corporations when measuring corporate entities and partners and proprietors when measuring noncorporate entities. We are finding that this aggregation is masking important economic information that is relevant for researchers interested in U.S. tax policy.

A General Method for Structural Estimation using Estimated Policy Functions
Joseph Mullins

I propose a new method for estimating the parameters of dynamic models using observed choice behavior. My proposal builds on frontier methods in the literature by nesting those techniques within a general class. I show that this general class of estimator admits a new procedure, with equivalent asymptotic properties, that improves upon existing trade-offs between computational simplicity, finite sample performance, and asymptotic efficiency.

Maturity and Multiplicity in Sovereign Debt Models
Manuel Amador
In this project, we propose and analytically solve a tractable version of the canonical Eaton and Gersovitz (1981) sovereign debt model with long duration bonds. Long-term bonds play a key role in bringing the quantitative sovereign debt models closer to the data, in large part due to the inherent incentive to dilute bondholders. We establish that the same force generates multiplicity. We are able to make progress by simplifying the underlying structure of the model. The main ingredients are the introduction of outside option shocks to the government (rather than the usual income shocks), linear utility from expenditures for the government (up to an upper bound), and continuous time. With this, we show if the government is more impatient than foreign creditors, the maturity of the bonds is intermediate, and there are deadweight losses from default; then multiple equilibria emerge. All of these are usual assumptions made in the quantitative literature.

Resource Allocation with Market Incompleteness and Financial Frictions
Anmol Bhandari
This project will study how an economy allocates resources for production and consumption in presence of market incompleteness and financial frictions. Our hypothesis is that factor misallocation is closely tied to the risk-sharing avenues available to firm owners. In contrast to the commonly studied bond-only economy with collateral constraints (for example Moll (2014)), the degree of misallocation can be increasing in the persistence of the idiosyncratic risk when firms have access to state-contingent contracts. The possibility to transfer wealth from high productivity states to low productivity states allows firm owners to trade off the efficient allocation of consumption against the efficient allocation of capital.

Financial Frictions, Startups and the Business Cycle
V.V. Chari
This project aims to better understand whether access to external finance deteriorates in recessions, and quantify its impact on the creation of fast-growing firms. We develop a quantitative general equilibrium model with heterogeneous firms, which features aggregate shocks, financial frictions and firms’ entry and exit. The model is disciplined using a microdata panel covering the universe of U.K. firms. The estimated model is used to evaluate the importance of financial constraints for the short- and medium-run behavior of the economy, and offer several policy proposals aimed at stimulating the entry of firms with high growth potential during recessions.

The Macroeconomics of Intensive Agriculture
Hannes Malmberg
In poor countries, the agricultural sector is large, uses few inputs per worker, and is unproductive, even relative to the local economy. In this project, we jointly study these phenomena using a neoclassical two-sector model. To inform our choice of an agricultural production function, we draw on micro and macro sources to construct a cross-country database of agricultural input use and prices. Using our framework, we separately analyze the consequences of closing the sectoral TFP gap in agriculture and non-agriculture between a poor country and the US. Our preliminary results suggest that closing the agricultural TFP gap is much less powerful than closing non-agricultural TFP gaps in closing aggregate income differences. This finding is very different to those found in earlier simple decomposition exercises, which have suggested that closing agricultural labor productivity gaps could dramatically reduce income differences. Our project focuses on how general equilibrium forces makes full counterfactuals very different from simple decomposition exercises.

Family Law Effects on Divorce, Fertility, and Child Investment
Joseph Mullins
We propose estimating a dynamic model of marriage, fertility and parenting, with the main goal being the investigation of how policies toward divorce influence individual welfare and child skill outcomes. We will discipline our estimates of preferences and the technology of child development with panel data on parental time inputs, child test scores, earnings, and divorce decisions. Using the estimated model, we can perform ex-ante evaluation of the impact of policies governing divorced parenting, such as child support orders, child custody and placement regulations, and marital dissolution standards. The model allows us to look beyond the effect of these policies on rates of divorce, and consider optimal policy combinations for adult welfare and for child skill outcomes.

Managing Public Debt
Anmol Bhandari
One of the central questions in macroeconomics is how the government should set its fiscal policy and manage its debts portfolio over the business cycle. The large quantities of public debts accumulated in the aftermath of the 2007 financial crises renewed debates among economists and policymakers about whether the current levels of public debt are suboptimally high, how quickly they should be re-paid and what is the desirable long run target for debt level. Our methods will incorporate fiscal hedging, liquidity, and collateral benefits of issuing debt taking into account distortionary costs of raising taxes and potential lack of commitment issues that governments face. Our main results will derive sufficient statistics that are informed with asset prices and provide practical guidance to policymakers in deciding the level and composition debt over the business cycle.

The Macrodynamics of the Wage Distribution
Jeremy Lise
The project aims to better understand how the distribution of wages in the economy evolve in response to the business cycle. A model of the labor market is developed in which workers are heterogeneous in terms of their initial skills, firms differ in productivity and which types of skills they demand, and the business cycle may have very different effects on low and high skilled workers and on firms operating in different sectors. This theory is combine with comprehensive matched worker-firm data which will allow for the measurement of differential impact on the lifetime earnings of workers across the skill distribution, as well as profits across the firm types. The estimated model is well suited to evaluate policies aimed to improve the living standards of low wages workers, for example minimum wages or earned income tax credits.

Monetary Policy, Fiscal Policy, and Crisis: Application to Latin America
Tim Kehoe
The objective of this project is to develop a conceptual framework that can be used to evaluate the role of aggregate fiscal policy on macroeconomic outcomes. This conceptual framework will be the cornerstone of a larger parallel project, whose purpose is to write the fiscal and monetary history of Latin-America. The proposed research will proceed in parallel along three complementary tracks: The systematic collection and analysis of data for a comparable set of countries, the analysis of recurrent historical episodes and, finally, the evaluation of existing models and development of new models to guide the analysis of these data.

The Impact of Consumer Credit Access on Earnings Mobility, Entrepreneurship and Income Inequality
Kyle Herkenhoff
Households have gained significant amounts of access to consumer credit, such as credit cards, student debt, and mortgages over the last 40 years. Little is known about how the rise of consumer credit access, at the family or individual level, has impacted income inequality or the way families respond to financial distress. This study aims to uncover the relationship between credit access among family members (the head-of-household, spouse, and children) and labor supply, earnings mobility (inclusive of labor and business income), and family-level responses to adverse shocks.

The Theory and Empirics of Subjective Beliefs in Macroeconomic Dynamics
Anmol Bhandari
This project lays down a general framework for incorporating subjective beliefs into models of macroeconomic dynamics. The connection of theoretical restrictions with informative data sources yields new insights into the joint determination of beliefs and equilibrium dynamics in the macroeconomy. The second contribution is to extend the existing toolkits that macroeconomists use to solve and estimate dynamic equilibrium models to allow for endogenous subjective beliefs. Finally, forward-looking policies affect these endogenous beliefs which opens the scope for explicit expectations management by the policymakers. We formalize how to pose optimal policy problems with endogenous subjective beliefs and explain the new lessons that emerge when policymakers acknowledge the feedback between policies and equilibrium belief formation.

Credit Access and Inequality
Kyle Herkenhoff
What is the impact of credit access on inequality and income mobility? Does access to credit exacerbate or mitigate inequality? We test how credit access impacts earnings through two main channels: (1) revolving credit, student debt, and human capital formation and (2) business credit and entrepreneurship. Understanding the answer to these questions is an important step toward optimal government policy for student loans, entrepreneur loans, and other government subsidized forms of credit.

Housing and Taxation over the Life Cycle
Larry Jones
This project will theoretically and quantitatively study the optimal joint taxation for consumption, housing, and labor earnings over the life cycle with unrestricted tax instruments. To connect our theory to the empirical environment, we propose to use unique administrative data for the Netherlands. This project has two contributions. First, we propose a two-step method to connect dynamic Mirrlees models to the data. In the first step, we infer preference and shock process parameters by estimating a decentralized life-cycle economy. In the second step, we use the estimated parameters to quantify the optimal allocation subject to information constraints. Second, we provide a normative prescription on how to optimally tax consumption, earnings and housing.

Deep Reinforcement Learning Analysis of Stochastic Games and Computational Problems in Economics
Aldo Rustichini
This research project will extend the techniques of Deep Reinforcement Learning (DRL) to the analysis of stochastic games, and thus to provide a foundation of the theory of rules as guidance for behavior. As follow up, we plan to extend DRL to the analysis of a large class of computational problems in Economics. The immediate motivation for the extension is to provide a psychologically plausible model of behavior in repeated and stochastic games, based on the replacement of strategies with rules. A set of rules provides a potentially incomplete guidance of behavior. The starting point will be the application of DRL techniques to any stochastic game with finite state space with action and state-dependent transition.

The Affordable Care Act and the Labor Market
Naoki Aizawa
The 2010 Affordable Care Act (ACA) dramatically changes the policy environment for both labor and health insurance markets. Claims have been made that ACA may reduce overall efficiency in the labor market by distorting firms’ optimal decisions regarding size or composition of full-time and part-time employees.  It has also been argued that under the pre-ACA health insurance system, entrepreneur decisions were distorted by poor individual insurance markets. This project will use empirical models to investigate these claims and their policy implications.

The Theory and Empirics of Sweat Equity
Anmol Bhandari
There are about 6 million firms in the U.S. with 99.6% employing fewer than 500 employees, accounting for more than half of the aggregate labor force. In the past 30 years, there has been a striking change in how small businesses organize their activity. Now, the majority of business income is earned by pass through entities, including Schedule S corporations, partnerships and sole proprietorships, with most growth coming from LLC partnerships and S corporations. The pass through entities are for the most part small businesses with one or two owners or shareholders and, according to BEA studies, are more nimble in shifting income to avoid regulations and minimize taxes. This research will examine accounting for intangible investments that pass-through business owners make that do not show up in business income reporting and ultimately GDP. These findings have implications for the debate on wealth vs. earnings inequality that is central to policy questions of taxation and redistribution. 

How Much Can Job Losers Borrow, and How Much Do They Actually Borrow? Implications for Optimal Unemployment Insurance
Kyle Herkenhoff
This project will be the first research to merge individual credit reports with census employment histories. This unique quarterly panel includes 5 million individuals across the US and allows for the observation of borrowing behavior of displaced workers and measurement of credit limits of displaced workers. This research is the first to measure credit constraints directly using measures of borrowing capacity (including credit scores and credit limits) as well as direct measures of ability-to-obtain-loans (such as loan application and denial rates).

The Monetary and Fiscal History of Latin America
Tim Kehoe
Since the mid 1970s, countries in Latin America have been plagued by economic crises. This project will systematically collect and analyze data for a comparable set of countries, analyze recurrent historical episodes, and evaluate existing models and develop new models to guide the analysis of these data. The theoretical analysis and the collected volume that will be produced as a result of this research will serve as a guide for policymakers as to how to avoid the same mistakes. Events over the past five years suggest that this guide will also be useful for researchers and policy makers outside Latin America, such as countries in the Eurozone and even the United States.

Inequality in Lifetime Earnings
Fatih Guvenen
Using a unique, confidential, and very large dataset on earnings histories from the US Social Security Administration, this project will estimate the distribution of lifetime earnings for individuals in the US. This data will then be used to address questions such as: (1) How much inequality is there in lifetime earnings among US individuals? (2) How much dispersion is there across men and women and workers with different racial backgrounds in lifetime earnings? (3) How should lifetime earnings inform tax policy? (4) Current Social Security benefits are calculated based on the maximum 35 years of earnings during the working life. Using our measurement of lifetime earnings for different US cohorts, what are some alternatives?

Medicare Advantage Private-Plan Reimbursement Rates and the Associated Savings and Costs to Taxpayers
Amil Petrin
Does adding the Medicare Advantage market reduce or increase overall health care costs and consumer welfare? The goal of this project is to develop the most advanced model of demand and supply of the Medicare Advantage market to date to estimate how its existence impacts costs and welfare. Ultimately, it will answer the question of what – if any – the optimal reimbursement rates and rules in the market should be to achieve lower health care costs.

Information Management Systems and Societal Cooperation
David Rahman
Understanding which informational institutions can help individuals cooperate better is a fundamental question of both mechanism design and public policy. Famously, Abreu, Milgrom and Pearce (1991, Econometrica) showed how simple information management institutions can dramatically impact the scope for cooperation in society. Their two main theories were (i) delaying the arrival of information to individuals in society can increase welfare, and (ii) increasing the frequency with which individuals interact can decrease welfare. In the context of the Prisoners' Dilemma with imperfect monitoring, this project will test these theories.