Publications

This paper studies the welfare effects of encouraging rural–urban migration in the developing world. To do so, we build and analyze a dynamic general-equilibrium model of migration that features a rich set of migration motives. We estimate the model to replicate the results of a field experiment that subsidized seasonal migration in rural Bangladesh, leading to significant increases in migration and consumption. We show that the welfare gains from migration subsidies come from providing better insurance for vulnerable rural households rather than from correcting spatial misallocation by relaxing credit constraints for those with high productivity in urban areas that are stuck in rural areas.

This paper uses data from the 2019 Annual Business Survey to document that firms adopting advanced technologies are larger in terms of employment than other firms in their same industry and cohort. Using data from the Longitudinal Business Survey, we show that adopters were already large and growing faster before artificial intelligence, robotics, cloud computing, and specialized software systems became broadly available. These findings support the view that adopters are large because of selection and not because adopting advanced technologies for automation causally expands their employment.

We develop an axiomatic theory of information acquisition that captures the idea of constant marginal costs in information production: the cost of generating two independent signals is the sum of their costs, and generating a signal with probability half costs half its original cost. Together with Blackwell monotonicity and a continuity condition, these axioms determine the cost of a signal up to a vector of parameters. These parameters have a clear economic interpretation and determine the difficulty of distinguishing states.

We develop a dynamic model of input-output networks that incorporates adjustment costs of changing inputs. Our closed-form solution for the dynamics of the economy shows that temporary shocks to upstream sectors, whose output travels through long supply chains, have disproportionately significant welfare impact compared to affected sectors’ Domar weights. We conduct a spectral analysis of the U.S. production network and reveal that the welfare impact of temporary sectoral shocks can be represented by a low-dimensional, 4-factor structure.

What do recent advances in economic geography teach us about the spatial distribution of economic activity? We show that the equilibrium distribution of economic activity can be determined simply by the intersection of labor supply and demand curves. We discuss how to estimate these curves and highlight the importance of global geography—the connections between locations through the trading network—in determining how various policy relevant changes to geography shape the spatial economy.

Do foreign students affect the likelihood that domestic students obtain a STEM degree and occupation? Using administrative student records from a US university, we exploit idiosyncratic variation in the share of foreign classmates in introductory math classes and find that foreign classmates displace domestic students from STEM majors and occupations. However, displaced students gravitate toward high-earning social science majors, so their expected earnings are not penalized. We explore several mechanisms. Results indicate that displacement is concentrated in classes where foreign classmates possess weak English language ability, suggesting that diminished in-class communication and social interactions might play an important role.

This paper studies control function (CF) approaches in endogenous threshold regression where the threshold variable is allowed to be endogenous. We first use a simple example to show that the structural threshold regression (STR) estimator of the threshold point in Kourtellos, Stengos and Tan (2016, Econometric Theory 32, 827–860) is inconsistent unless the endogeneity level of the threshold variable is low compared to the threshold effect. We correct the CF in the STR estimator to generate our first CF estimator using a method that extends the two-stage least squares procedure in Caner and Hansen (2004, Econometric Theory 20, 813–843). We develop our second CF estimator which can be treated as an extension of the classical CF approach in endogenous linear regression. Both these approaches embody threshold effect information in the conditional variance beyond that in the conditional mean. Given the threshold point estimates, we propose new estimates for the slope parameters. The first is a by-product of the CF approach, and the second type employs generalized method of moment (GMM) procedures based on two new sets of moment conditions. Simulation studies, in conjunction with the limit theory, show that our second CF estimator and confidence interval for the threshold point together with the associated second GMM estimator and confidence interval for the slope parameter dominate the other methods. We further apply the new estimation methodology to an empirical application from international trade to illustrate its usefulness in practice.

We offer an approach to cooperation in repeated games of private monitoring in which players construct models of their opponents' behavior by observing the frequencies of play in a record of past plays of the game in which actions but not signals are recorded. Players construct models of their opponent's behavior by grouping the histories in the record into a relatively small number of analogy classes for which they estimate probabilities of cooperation. The incomplete record and the limited number of analogy classes lead to misspecified models that provide the incentives to cooperate. We provide conditions for the existence of equilibria supporting cooperation and equilibria supporting high payoffs for some nontrivial analogy partitions.

This paper investigates a two-agent mechanism design problem without transfers, where the principal must decide one action for each agent. In our framework, agents only care about their own adaptation, and any deterministic dominant incentive compatible decision rule is equivalent to contingent delegation: the delegation set offered to one agent depends on the other's report. By contrast, the principal cares about both adaptation and coordination. We provide sufficient conditions under which contingent interval delegation is optimal and solve the optimal contingent interval delegation under fairly general conditions. Remarkably, the optimal interval delegation is completely determined by combining and modifying the solutions to a class of simple single-agent problems, where the other agent is assumed to report truthfully and choose his most preferred action.

We argue that noisy aggregation of dispersed information provides a unified explanation for several prominent cross-sectional return anomalies such as returns to skewness, returns to disagreement and corporate credit spreads. We characterize asset returns with noisy information aggregation by means of a risk-neutral probability measure that features excess weight on tail risks, and link the latter to observable moments of earnings forecasts, in particular forecast dispersion and accuracy. We calibrate our model to match these moments and show that it accounts for a large fraction of the empirical return premia. We further develop asset pricing tools for noisy information aggregation models that do not impose strong parametric restrictions on economic primitives such as preferences, information, or return distributions.

We present an approach to analyse learning outcomes in a broad class of misspecified environments, spanning both single-agent and social learning. We introduce a novel “prediction accuracy” order over subjective models and observe that this makes it possible to partially restore standard martingale convergence arguments that apply under correctly specified learning. Based on this, we derive general conditions to determine when beliefs in a given environment converge to some long-run belief either locally or globally (i.e. from some or all initial beliefs). We show that these conditions can be applied, first, to unify and generalize various convergence results in previously studied settings. Second, they enable us to analyse environments where learning is “slow”, such as costly information acquisition and sequential social learning. In such environments, we illustrate that even if agents learn the truth when they are correctly specified, vanishingly small amounts of misspecification can generate extreme failures of learning.

This paper studies the relation between volatility and informativeness in financial markets. We identify two channels (noise-reduction and equilibrium-learning) that determine the volatility-informativeness relation. When informativeness is sufficiently high (low), volatility and informativeness positively (negatively) comove in equilibrium. We identify conditions on primitives that guarantee that volatility and informativeness comove positively or negatively. We introduce the comovement score, a statistic that measures the distance of a given asset to the positive/negative comovement regions. Empirically, comovement scores (i) have trended downwards over the last decades, (ii) are positively related to value and idiosyncratic volatility and negatively to size and institutional ownership.

This paper studies high-dimensional vector autoregressions (VARs) augmented with common factors that allow for strong cross-sectional dependence. Models of this type provide a convenient mechanism for accommodating the interconnectedness and temporal co-variability that are often present in large dimensional systems. We propose an ℓ1-nuclear-norm regularized estimator and derive the non-asymptotic upper bounds for the estimation errors as well as large sample asymptotics for the estimates. A singular value thresholding procedure is used to determine the correct number of factors with probability approaching one. Both the LASSO estimator and the conservative LASSO estimator are employed to improve estimation precision. The conservative LASSO estimates of the non-zero coefficients are shown to be asymptotically equivalent to the oracle least squares estimates. Simulations demonstrate that our estimators perform reasonably well in finite samples given the complex high-dimensional nature of the model. In an empirical illustration we apply the methodology to explore dynamic connectedness in the volatilities of financial asset prices and the transmission of ‘investor fear’. The findings reveal that a large proportion of connectedness is due to the common factors. Conditional on the presence of these common factors, the results still document remarkable connectedness due to the interactions between the individual variables, thereby supporting a common factor augmented VAR specification.

We compare contrarian to conformist advice, a contrarian expert being one whose preference bias is against the decision-maker’s prior optimal decision. Optimality of an expert depends on characteristics of prior information and learning. If either the expert is fully informed or fine information can be acquired cheaply, then for symmetric distributions F (of the state), a conformist (contrarian) is superior if F is single peaked bimodal. If only coarse information can be acquired, then a contrarian acquires more on average and hence is superior. If information is verifiable, a contrarian has less incentive to hide unfavorable evidence and again is superior.

Spatial units typically vary over many of their characteristics, introducing potential unobserved heterogeneity which invalidates commonly used homoskedasticity conditions. In the presence of unobserved heteroskedasticity, methods based on the quasi-likelihood function generally produce inconsistent estimates of both the spatial parameter and the coefficients of the exogenous regressors. A robust generalized method of moments estimator as well as a modified likelihood method have been proposed in the literature to address this issue. The present paper constructs an alternative indirect inference (II) approach which relies on a simple ordinary least squares procedure as its starting point. Heteroskedasticity is accommodated by utilizing a new version of continuous updating that is applied within the II procedure to take account of the parameterization of the variance–covariance matrix of the disturbances. Finite-sample performance of the new estimator is assessed in a Monte Carlo study. The approach is implemented in an empirical application to house price data in the Boston area, where it is found that spatial effects in house price determination are much more significant under robustification to heterogeneity in the equation errors.