Bayesian Factor Model Shrinkage for Linear IV Regression With Many Instruments

A Bayesian approach for the many instruments problem in linear instrumental variable models is presented. The new approach has two components. First, a slice sampler is developed, which leverages a decomposition of the likelihood function that is a Bayesian analogue to two-stage least squares. The new sampler permits nonconjugate shrinkage priors to be implemented easily and efficiently. The new computational approach permits a Bayesian analysis of problems that were previously infeasible due to computational demands that scaled poorly in the number of regressors. Second, a new predictor-dependent shrinkage prior is developed specifically for the many instruments setting. The prior is constructed based on a factor model decomposition of the matrix of observed instruments, allowing many instruments to be incorporated into the analysis in a robust way. Features of the new method are illustrated via a simulation study and three empirical examples.     

A note about the identifiability of causal effect estimates in randomized trials with non-compliance

We show that assumptions that are sufficient for estimating an average treatment effect in randomized trials with non-compliance restrict the subgroup means for always takers, compliers, defiers and never takers to a two-dimensional linear subspace of a four-dimensional space. Implications and special cases are exemplified.     

Identification and Inference With Many Invalid Instruments

We study estimation and inference in settings where the interest is in the effect of a potentially endogenous regressor on some outcome. To address the endogeneity, we exploit the presence of additional variables. Like conventional instrumental variables, these variables are correlated with the endogenous regressor. However, unlike conventional instrumental variables, they also have direct effects on the outcome, and thus are “invalid” instruments. Our novel identifying assumption is that the direct effects of these invalid instruments are uncorrelated with the effects of the instruments on the endogenous regressor. We show that in this case the limited-information-maximum-likelihood (liml) estimator is no longer consistent, but that a modification of the bias-corrected two-stage-least-square (tsls) estimator is consistent. We also show that conventional tests for over-identifying restrictions, adapted to the many instruments setting, can be used to test for the presence of these direct effects. We recommend that empirical researchers carry out such tests and compare estimates based on liml and the modified version of bias-corrected tsls. We illustrate in the context of two applications that such practice can be illuminating, and that our novel identifying assumption has substantive empirical content.     

A Consistent Method for the Selection of Relevant Instruments

Abstract

In many applications, a researcher must select an instrument vector from a candidate set of instruments. If the ultimate objective is to perform inference about the unknown parameters using conventional asymptotic theory, then we argue that it is desirable for the chosen instrument vector to satisfy four conditions which we refer to as orthogonality, identification, efficiency, and non‐redundancy. It is impossible to verify a priori which elements of the candidate set satisfy these conditions; this can only be done using the data. However, once the data are used in this fashion it is important that the selection process does not contaminate the limiting distribution of the parameter estimator. We refer to this requirement as the inference condition. In a recent paper, Andrews [[Andrews, D. W. K. (1999)] Andrews, D. W.K. 1999. Consistent moment selection procedures for generalized method of moments estimation. Econometrica, 67: 543–564. [Crossref], [Web of Science ®] , [Google Scholar]. Consistent moment selection procedures for generalized method of moments estimation. Econometrica67:543–564] has proposed a method of moment selection based on an information criterion involving the overidentifying restrictions test. This method can be shown to select an instrument vector which satisfies the orthogonality condition with probability one in the limit. In this paper, we consider the problem of instrument selection based on a combination of the efficiency and non‐redundancy conditions which we refer to as the relevance condition. It is shown that, within a particular class of models, certain canonical correlations form the natural metric for relevancy, and this leads us to propose a canonical correlations information criterion (CCIC) for instrument selection. We establish conditions under which our method satisfies the inference condition. We also consider the properties of an instrument selection method based on the sequential application of [Andrews, D. W. K. (1999)] Andrews, D. W.K. 1999. Consistent moment selection procedures for generalized method of moments estimation. Econometrica, 67: 543–564. [Crossref], [Web of Science ®] , [Google Scholar]. Consistent moment selection procedures for generalized method of moments estimation. Econometrica67:543–564 method and CCIC.     

Estimation of Causal Effects in Latent Strata with an Encouragement for Response

We consider a new approach to deal with non ignorable non response on an outcome variable, in a causal inference framework. Assuming that a binary instrumental variable for non response is available, we provide a likelihood-based approach to identify and estimate heterogeneous causal effects of a binary treatment on specific latent subgroups of units, named principal strata, defined by the non response behavior under each level of the treatment and of the instrument. We show that, within each stratum, non response is ignorable and respondents can be properly compared by treatment status. In order to assess our method and its robustness when the usually invoked assumptions are relaxed or misspecified, we simulate data to resemble a real experiment conducted on a panel survey which compares different methods of reducing panel attrition.     

Estimation of Conditional Ranks and Tests of Exogeneity in Nonparametric Nonseparable Models

Consider a nonparametric nonseparable regression model Y = ?(Z, U), where ?(Z, U) is strictly increasing in U and U ～ U[0, 1]. We suppose that there exists an instrument W that is independent of U. The observable random variables are Y, Z, and W, all one-dimensional. We construct test statistics for the hypothesis that Z is exogenous, that is, that U is independent of Z. The test statistics are based on the observation that Z is exogenous if and only if V = F_Y|Z(Y|Z) is independent of W, and hence they do not require the estimation of the function ?. The asymptotic properties of the proposed tests are proved, and a bootstrap approximation of the critical values of the tests is shown to be consistent and to work for finite samples via simulations. An empirical example using the U.K. Family Expenditure Survey is also given. As a byproduct of our results we obtain the asymptotic properties of a kernel estimator of the distribution of V, which equals U when Z is exogenous. We show that this estimator converges to the uniform distribution at faster rate than the parametric n^{? 1/2}-rate.     

Nonparametric Additive Instrumental Variable Estimator: A Group Shrinkage Estimation Perspective

In this article, we study a nonparametric approach regarding a general nonlinear reduced form equation to achieve a better approximation of the optimal instrument. Accordingly, we propose the nonparametric additive instrumental variable estimator (NAIVE) with the adaptive group Lasso. We theoretically demonstrate that the proposed estimator is root-n consistent and asymptotically normal. The adaptive group Lasso helps us select the valid instruments while the dimensionality of potential instrumental variables is allowed to be greater than the sample size. In practice, the degree and knots of B-spline series are selected by minimizing the BIC or EBIC criteria for each nonparametric additive component in the reduced form equation. In Monte Carlo simulations, we show that the NAIVE has the same performance as the linear instrumental variable (IV) estimator for the truly linear reduced form equation. On the other hand, the NAIVE performs much better in terms of bias and mean squared errors compared to other alternative estimators under the high-dimensional nonlinear reduced form equation. We further illustrate our method in an empirical study of international trade and growth. Our findings provide a stronger evidence that international trade has a significant positive effect on economic growth.     

Random Effects Estimators with many Instrumental Variables

In this paper we propose a new estimator for a model with one endogenous regressor and many instrumental variables. Our motivation comes from the recent literature on the poor properties of standard instrumental variables estimators when the instrumental variables are weakly correlated with the endogenous regressor. Our proposed estimator puts a random coefficients structure on the relation between the endogenous regressor and the instruments. The variance of the random coefficients is modelled as an unknown parameter. In addition to proposing a new estimator, our analysis yields new insights into the properties of the standard two‐stage least squares (TSLS) and limited‐information maximum likelihood (LIML) estimators in the case with many weak instruments. We show that in some interesting cases, TSLS and LIML can be approximated by maximizing the random effects likelihood subject to particular constraints. We show that statistics based on comparisons of the unconstrained estimates of these parameters to the implicit TSLS and LIML restrictions can be used to identify settings when standard large sample approximations to the distributions of TSLS and LIML are likely to perform poorly. We also show that with many weak instruments, LIML confidence intervals are likely to have under‐coverage, even though its finite sample distribution is approximately centered at the true value of the parameter. In an application with real data and simulations around this data set, the proposed estimator performs markedly better than TSLS and LIML, both in terms of coverage rate and in terms of risk.     

Instrumental variables estimation: Assumptions,pitfalls, and guidelines

Researchers striving to ensure rigor in their scientific findings face a common pitfall: Endogeneity. To tackle this problem, scholars have increasingly adopted instrumental variables estimation (IVE). Although there are many published works showing how IVE should be used, many applied researchers still have trouble understanding how to use the method correctly. In this article, we provide a methodological overview of IVE by discussing the underlying conditions valid instruments must satisfy as well as common mistakes made in using IVE. Using simulated data, we further demonstrate the sensitivity of IVE to violations of its conditions. We then take stock of the literature in a social science discipline (i.e., leadership research) and provide insights regarding trends and shortcomings in the application of IVE. Based on our review, we categorize the different types of instruments used and discuss the potential appropriateness of each type. We conclude by providing non–technical guidelines targeted at the study design, analysis, and reporting phases, which will help applied social science researchers to ensure they use IVE correctly.     

Evaluating theory-based evaluation: information, norms, and adherence

Programmatic social interventions attempt to produce appropriate social-norm-guided behavior in an open environment. A marriage of applicable psychological theory, appropriate program evaluation theory, and outcome of evaluations of specific social interventions assures the acquisition of cumulative theory and the production of successful social interventions - the marriage permits us to advance knowledge by making use of both success and failures. We briefly review well-established principles within the field of program evaluation, well-established processes involved in changing social norms and social-norm adherence, the outcome of several program evaluations focusing on smoking prevention, pro-environmental behavior, and rape prevention and, using the principle of learning from our failures, examine why these programs often do not perform as expected. Finally, we discuss the promise of learning from our collective experiences to develop a cumulative science of program evaluation and to improve the performance of extant and future interventions.