Working papers

In our laboratory experiment, subjects, in sequence, have to predict the value of a good. We elicit the second subject’s belief twice: first (“first belief”), after he observes his predecessor’s action; second (“posterior belief”), after he observes his private signal. Our main result is that the second subjects weigh the private signal as a Bayesian agent would do when the signal confirms their first belief; they overweight the signal when it contradicts their first belief.

This paper provides estimation and inference methods for a structural function, such as Conditional Average Treatment Effect (CATE), based on modern machine learning (ML) tools.

We develop a theory for estimation of a high-dimensional sparse parameter 𝛳 defined as a minimizer of a population loss function LD(𝛳,g0) which, in addition to 𝛳, depends on a, potentially infinite dimensional, nuisance parameter g0.

Factor structures or interactive effects are convenient devices to incorporate latent variables in panel data models. We consider fixed effect estimation of nonlinear panel single-index models with factor structures in the unobservables, which include logit, probit, ordered probit and Poisson specifications.

An oil lease auction is the classic example motivating a common values model. However, formal testing for common values has been hindered by unobserved auction-level heterogeneity, which is likely to affect both participation in an auction and bidders' willingness to pay. We develop and apply an empirical approach for first-price sealed bid auctions with affiliated values, unobserved heterogeneity, and endogenous bidder entry.

We consider the estimation and inference in a system of high-dimensional regression equations allowing for temporal and cross-sectional dependency in covariates and error processes, covering rather general forms of weak dependence.

Recent years have seen substantial reductions in public spending on social care for older people in England. This has not only led to large falls in the number of people over the age of 65 receiving publicly funded social care, but also to growing concern about the potential knock-on effects on other public services, and in particular the National Health Service (NHS).

Multivalued treatment models have typically been studied under restrictive assumptions: ordered choice, and more recently unordered monotonicity. We show how treatment eects can be identied in a more general class of models that allows for multidimensional unobserved heterogeneity.

This chapter presents key concepts and theoretical results for analyzing estimation and inference in high-dimensional models. High-dimensional models are characterized by having a number of unknown parameters that is not vanishingly small relative to the sample size.

This paper studies the differential effect of targeting cash transfers to men or women on the structure of household expenditures on non-durables. We study a policy intervention in the Republic of Macedonia, offering cash transfers to poor households, conditional on having their children attending secondary school.

We show that the identification results of finite mixture and misclassification models are equivalent in a widely-used scenario except an extra ordering assumption. In the misclassification model, an ordering condition is imposed to pin down the precise values of the latent variable, which are also of researchers' interests and need to be identified.

Many questions in econometrics can be cast as inference on a parameter selected through optimization. For example, researchers may be interested in the effectiveness of the best policy found in a randomized trial, or the best-performing investment strategy based on historical data.

We give a general construction of debiased/locally robust/orthogonal (LR) moment functions for GMM, where the derivative with respect to first step nonparametric estimation is zero and equivalently first step estimation has no effect on the influence function.

The multinomial logit model with random coefficients is widely used in applied research. This paper is concerned with estimating a random coefficients logit model in which the distribution of each coefficient is characterized by finitely many parameters.

This paper investigates individuals' expectations about their own survival to older ages and compares patterns in average responses about survival chances with actual and projected survival rates.

This paper characterizes and proposes a method to correct for errors-in-variables biases in the estimation of rank correlation coeffcients (Spearman's ρ and Kendall's τ).

This paper studies the properties of the wild bootstrap-based test proposed in Cameron et al. (2008) in settings with clustered data.

This paper explores the effects of simulated moments on the performance of inference methods based on moment inequalities.

Consider a simple general equilibrium economy with one representative consumer, a single competitive firm and the government. Suppose that the government has to finance public expenditures using linear consumption taxes and/or a lump-sum tax on profits redistributed to the consumer. We show that, if the tax rate on profits cannot exceed 100 percent, one cannot improve upon the second-best optimum of an economy with constant returns to scale by using a less efficient profit-generating decreasing returns to scale technology.

We propose simultaneous mean-variance regression for the linear estimation and approximation of conditional mean functions.