Search

The instrumental variable method is widely used in causal inference research to improve the accuracy of estimating causal effects. However, the weak correlation between instruments and exposure, as well as the direct impact of instruments on the outcome, can lead to biased estimates. To mitigate the bias introduced by such instruments in nonlinear causal inference, we propose a two‐stage nonlinear causal effect estimation based on model averaging. The model uses different subsets of instruments in the first stage to predict exposure after a nonlinear transformation with the help of sliced inverse regression. In the second stage, adaptive Lasso penalty is applied to instruments to obtain the estimation of causal effect. We prove that the proposed estimator exhibits favorable asymptotic properties and evaluate its performance through a series of numerical studies, demonstrating its effectiveness in identifying nonlinear causal effects and its capability to handle scenarios with weak and invalid instruments. We apply the proposed method to the Atherosclerosis Risk in Communities dataset to investigate the relationship between BMI and hypertension. [ABSTRACT FROM AUTHOR]

Instrumental variable (IV) strategies are widely used in political science to establish causal relationships, but the identifying assumptions required by an IV design are demanding, and assessing their validity remains challenging. In this paper, we replicate 67 articles published in three top political science journals from 2010 to 2022 and identify several concerning patterns. First, researchers often overestimate the strength of their instruments due to non-i.i.d. error structures such as clustering. Second, IV estimates are often highly uncertain, and the commonly used t -test for two-stage-least-squares (2SLS) estimates frequently underestimate the uncertainties. Third, in most replicated studies, 2SLS estimates are significantly larger in magnitude than ordinary-least-squares estimates, and their absolute ratio is inversely related to the strength of the instrument in observational studies—a pattern not observed in experimental ones—suggesting potential violations of unconfoundedness or the exclusion restriction in the former. We provide a checklist and software to help researchers avoid these pitfalls and improve their practice. [ABSTRACT FROM AUTHOR]

We discuss the fundamental issue of identification in linear instrumental variable (IV) models with unknown IV validity. With the assumption of the 'sparsest rule', which is equivalent to the plurality rule but becomes operational in computation algorithms, we investigate and prove the advantages of non-convex penalized approaches over other IV estimators based on two-step selections, in terms of selection consistency and accommodation for individually weak IVs. Furthermore, we propose a surrogate sparsest penalty that aligns with the identification condition and provides oracle sparse structure simultaneously. Desirable theoretical properties are derived for the proposed estimator with weaker IV strength conditions compared to the previous literature. Finite sample properties are demonstrated using simulations and the selection and estimation method is applied to an empirical study concerning the effect of body mass index on diastolic blood pressure. [ABSTRACT FROM AUTHOR]

With mean instrumental variables regression, k-class estimators have the potential to reduce bias, which is larger with weak instruments. With instrumental variables quantile regression, weak instrument-robust estimation is even more important because there is less guidance for assessing instrument strength. Motivated by this, we introduce an analogous k-class of estimators for instrumental variables quantile regression. We show the first-order asymptotic distribution under strong instruments is equivalent for all conventional choices of k. We evaluate finite-sample median bias in simulations for a variety of k, including the k for the conventional k-class estimator corresponding to limited information maximum likelihood (LIML). Computation is fast for all k, and compared to the k = 1 benchmark estimator (analogous to 2SLS), using the LIML k reliably reduces median bias in a variety of data-generating processes, especially when the degree of overidentification is larger. We also revisit some empirical estimates of consumption Euler equations derived from quantile utility maximization. All code is provided online (https://kaplandm.github.io). [ABSTRACT FROM AUTHOR]

Evaluating policy in imperfectly competitive markets requires understanding firm behavior. While researchers test conduct via model selection and assessment, we present the advantages of Rivers and Vuong (2002) (RV) model selection under misspecification. However, degeneracy of RV invalidates inference. With a novel definition of weak instruments for testing, we connect degeneracy to instrument strength, derive weak instrument properties of RV, and provide a diagnostic for weak instruments by extending the framework of Stock and Yogo (2005) to model selection. We test vertical conduct (Villas‐Boas (2007)) using common instrument sets. Some are weak, providing no power. Strong instruments support manufacturers setting retail prices. [ABSTRACT FROM AUTHOR]

Linear instrumental variable regressions are widely used to estimate causal effects. Many instruments arise from the use of 'technical' instruments and more recently from the empirical strategy of 'judge design'. This paper surveys and summarises ideas from recent literature on estimation and statistical inferences with many instruments for a single endogenous regressor. We discuss how to assess the strength of the instruments and how to conduct weak identification robust inference under heteroskedasticity. We establish new results for a jack-knifed version of the Lagrange Multiplier test statistic. Furthermore, we extend the weak identification robust tests to settings with both many exogenous regressors and many instruments. We propose a test that properly partials out many exogenous regressors while preserving the re-centring property of the jack-knife. The proposed tests have correct size and good power properties. [ABSTRACT FROM AUTHOR]

Mendelian randomization (MR) is a widely used method to estimate the causal effect of an exposure on an outcome by using genetic variants as instrumental variables. MR analyses that use variants from only a single genetic region (cis‐MR) encoding the protein target of a drug are able to provide supporting evidence for drug target validation. This paper proposes methods for cis‐MR inference that use many correlated variants to make robust inferences even in situations, where those variants have only weak effects on the exposure. In particular, we exploit the highly structured nature of genetic correlations in single gene regions to reduce the dimension of genetic variants using factor analysis. These genetic factors are then used as instrumental variables to construct tests for the causal effect of interest. Since these factors may often be weakly associated with the exposure, size distortions of standard t‐tests can be severe. Therefore, we consider two approaches based on conditional testing. First, we extend results of commonly‐used identification‐robust tests for the setting where estimated factors are used as instruments. Second, we propose a test which appropriately adjusts for first‐stage screening of genetic factors based on their relevance. Our empirical results provide genetic evidence to validate cholesterol‐lowering drug targets aimed at preventing coronary heart disease. [ABSTRACT FROM AUTHOR]

Instrumental variables (IVs) are extensively used to handle unmeasured confounding. However, weak IVs may cause problems. Many matched studies have considered strengthening an IV through discarding some of the sample. It is widely accepted that strengthening an IV tends to increase the power of non-parametric tests and sensitivity analyses. We re-evaluate this conventional wisdom and offer new insights. First, we evaluate the trade-off between IV strength and sample size assuming a valid IV and exhibit conditions under which strengthening an IV increases power. Second, we derive a criterion for checking the validity of a sensitivity analysis model with a continuous dose and show that the widely used Γ sensitivity analysis model, which was used to argue that strengthening an IV increases the power of sensitivity analyses in large samples, does not work for continuous IVs. Third, we quantify the bias of the Wald estimator with a possibly invalid IV and leverage it to develop a valid sensitivity analysis framework and show that strengthening an IV may or may not increase the power of sensitivity analyses. We use our framework to study the effect on premature babies of being delivered in a high technology/high volume neonatal intensive care unit. [ABSTRACT FROM AUTHOR]

Mendelian randomization utilizes genetic variants as instrumental variables (IVs) to estimate the causal effect of an exposure variable on an outcome of interest even in the presence of unmeasured confounders. However, the popular inverse‐variance weighted (IVW) estimator could be biased in the presence of weak IVs, a common challenge in MR studies. In this article, we develop a novel penalized inverse‐variance weighted (pIVW) estimator, which adjusts the original IVW estimator to account for the weak IV issue by using a penalization approach to prevent the denominator of the pIVW estimator from being close to zero. Moreover, we adjust the variance estimation of the pIVW estimator to account for the presence of balanced horizontal pleiotropy. We show that the recently proposed debiased IVW (dIVW) estimator is a special case of our proposed pIVW estimator. We further prove that the pIVW estimator has smaller bias and variance than the dIVW estimator under some regularity conditions. We also conduct extensive simulation studies to demonstrate the performance of the proposed pIVW estimator. Furthermore, we apply the pIVW estimator to estimate the causal effects of five obesity‐related exposures on three coronavirus disease 2019 (COVID‐19) outcomes. Notably, we find that hypertensive disease is associated with an increased risk of hospitalized COVID‐19; and peripheral vascular disease and higher body mass index are associated with increased risks of COVID‐19 infection, hospitalized COVID‐19, and critically ill COVID‐19. [ABSTRACT FROM AUTHOR]

We propose exact exogeneity tests and weak-instruments-robust tests on factor loadings for a system of regressions with possibly non-Gaussian disturbances. Our methodology is valid in finite samples and accounts for common cross-sectional factors. Analytical invariance results are derived, with companion simulation studies. Finally, a total-effect parameter is introduced that embeds the unobservable endogeneity factor. Proposed tests are applied to assess whether Catastrophe bond mutual funds co-move with financial markets. Significant risk premiums are detected globally and over time, although they are less pervasive from a domestic currency perspective. Findings underscore the importance of instrumenting and assessing direct and total effects. [ABSTRACT FROM AUTHOR]

We compare two approaches to using information about the signs of structural shocks at specific dates within a structural vector autoregression (SVAR): imposing "narrative restrictions" (NR) on the shock signs in an otherwise set-identified SVAR; and casting the information about the shock signs as a discrete-valued "narrative proxy" (NP) to point-identify the impulse responses. The NP is likely to be "weak" given that the sign of the shock is typically known in a small number of periods, in which case the weak-proxy robust confidence intervals in Montiel Olea, Stock, and Watson are the natural approach to conducting inference. However, we show both theoretically and via Monte Carlo simulations that these confidence intervals have distorted coverage—which may be higher or lower than the nominal level—unless the sign of the shock is known in a large number of periods. Regarding the NR approach, we show that the prior-robust Bayesian credible intervals from Giacomini, Kitagawa, and Read deliver coverage exceeding the nominal level, but which converges toward the nominal level as the number of NR increases. [ABSTRACT FROM AUTHOR]

For a binary outcome Y, generated by a simple threshold crossing model with a single exogenous normally distributed explanatory variable X, the OLS estimator of the coefficient on X in a linear probability model is a consistent estimator of the average partial effect of X. Even in this very simple setting, we show that when allowing for X to be endogenously determined, the 2SLS estimator, using a normally distributed instrumental variable Z, does not identify the same causal parameter. It instead estimates the average partial effect of Z, scaled by the coefficient on Z in the linear first-stage model for X, denoted γ1, or equivalently, it estimates the average partial effect of the population predicted value of X, Z γ 1. These causal parameters can differ substantially as we show for the normal Probit model, which implies that care has to be taken when interpreting 2SLS estimation results in a linear probability model. Under joint normality of the error terms, IV Probit maximum likelihood estimation does identify the average partial effect of X. The two-step control function procedure of Rivers and Vuong can also estimate this causal parameter consistently, but a double averaging is needed, one over the distribution of the first-stage error V and one over the distribution of X. If instead a single averaging is performed over the joint distribution of X and V, then the same causal parameter is estimated as the one estimated by the 2SLS estimator in the linear probability model. The 2SLS estimator is a consistent estimator when the average partial effect is equal to 0, and the standard Wald test for this hypothesis has correct size under strong instrument asymptotics. We show that, in general, the standard weak instrument first-stage F-test interpretations do not apply in this setting. [ABSTRACT FROM AUTHOR]

Summary: This paper proposes a new method for estimating linear dynamic structural models. The proposed generalized band spectrum estimator (GBSE) generalizes band spectrum regression to simultaneously account for endogeneity and nonlinearities of unknown form in the first stage, while having a computationally convenient closed‐form expression in the time domain. We apply the GBSE to the hybrid New Keynesian Phillips curve (NKPC) with U.S. postwar data. We find a stable marginal cost coefficient in the short run, with relatively large and statistically significant values, and small and insignificant values when both short‐run and long‐run frequencies are included. The forward‐looking component and the inflation inertia are relatively stable and equally quantitatively important. Overall, our estimates present much less sampling uncertainty than the corresponding generalized method of moment (GMM) estimates, both in extensive Monte Carlo simulations and the empirical application, and they provide formal empirical support for misspecification of the NKPC as a model for all frequencies. [ABSTRACT FROM AUTHOR]

ObjectiveTo evaluate the association between participation in the Supplemental Nutrition Assistance Program (SNAP) and body mass index (BMI) in the presence of unmeasured confounding.MethodsWe applied new matching methods to determine whether previous reports of associations between SNAP participation and BMI were robust to unmeasured confounders. We applied near-far matching, which strengthens standard matching by combining it with instrumental variables analysis, to the nationally-representative National Household Food Acquisition and Purchasing Survey (FoodAPS, N = 10,360, years 2012-13).ResultsIn ordinary least squares regressions controlling for individual demographic and socioeconomic characteristics, SNAP was associated with increased BMI (+1.23 kg/m2, 95% CI: 0.84, 1.63). While propensity-score-based analysis replicated this finding, using instrumental variables analysis and particularly near-far matching to strengthen the instruments' discriminatory power revealed the association between SNAP and BMI was likely confounded by unmeasured covariates (+0.21 kg/m2, 95% CI: -3.88, 4.29).ConclusionsPrevious reports of an association between SNAP and obesity should be viewed with caution, and use of near-far matching may assist similar assessments of health effects of social programs.

Two‐sample summary data Mendelian randomization is a popular method for assessing causality in epidemiology, by using genetic variants as instrumental variables. If genes exert pleiotropic effects on the outcome not entirely through the exposure of interest, this can lead to heterogeneous and (potentially) biased estimates of causal effect. We investigate the use of Bayesian model averaging to preferentially search the space of models with the highest posterior likelihood. We develop a Metropolis‐Hasting algorithm to perform the search using the recently developed MR‐RAPS as the basis for defining a posterior distribution that efficiently accounts for pleiotropic and weak instrument bias. We demonstrate how our general modeling approach can be extended from a standard one‐component causal model to a two‐component model, which allows a large proportion of SNPs to violate the InSIDE assumption. We use Monte Carlo simulations to illustrate our methods and compare it to several related approaches. We finish by applying our approach to investigate the causal role of cholesterol on the development age‐related macular degeneration. [ABSTRACT FROM AUTHOR]

Many models in econometrics involve endogeneity and lagged dependent variables. We start by observing that usual identification-robust (IR) tests are unreliable when model variables are nonstationary or nearly nonstationary. We propose IR methods which are also robust to nonstationarity: one Anderson-Rubin type procedure and two split-sample procedures. Our procedures are also robust to missing instruments. For distributional theory, three different sets of assumptions are considered. First, on assuming Gaussian structural errors, we show that two of the proposed statistics follow the standard F distribution. Second, for more general cases, we assume that the distribution of errors is completely specified up to an unknown scale factor, allowing the Monte Carlo test method to be applied. This assumption enables one to deal with non-Gaussian error distributions. For example, even when errors follow heavy-tailed distribution, such as the Cauchy distribution or more generally the family of stable distributions—which may not have moments and thus make inference difficult—our procedures provide simple and exact solutions. Third, we establish the asymptotic validity of our procedures under quite general distributional assumptions. We present simulation results showing that our procedures control their level correctly and have good power properties. The methods are applied to an empirical example, the New Keynesian Phillips curve, in which both weak identification and nonstationarity present challenges. The results of this empirical study suggest forward-looking behavior of U.S. inflation. [ABSTRACT FROM AUTHOR]

In this article, we propose methods to construct confidence intervals for the bias of the two-stage least squares estimator, and the size distortion of the associated Wald test in instrumental variables models with heteroscedasticity and serial correlation. Importantly our framework covers the local projections—instrumental variable model as well. Unlike tests for weak instruments, whose distributions are nonstandard and depend on nuisance parameters that cannot be consistently estimated, the confidence intervals for the strength of identification are straightforward and computationally easy to calculate, as they are obtained from inverting a chi-squared distribution. Furthermore, they provide more information to researchers on instrument strength than the binary decision offered by tests. Monte Carlo simulations show that the confidence intervals have good, albeit conservative, in some cases, small sample coverage. We illustrate the usefulness of the proposed methods in two empirical situations: the estimation of the intertemporal elasticity of substitution in a linearized Euler equation, and government spending multipliers. for this article are available online. [ABSTRACT FROM AUTHOR]

We study the asymptotic behavior of instrumental variable estimators in the static panel model under many-instruments asymptotics. We provide new estimators and standard errors based on concentrated instruments as alternatives to an estimator based on maximum likelihood. We prove that the latter estimator is consistent under many-instruments asymptotics only if the starting value in an iterative procedure is root-N consistent. A similar approach for continuous updating GMM shows the derivation is nontrivial. For the standard cross-sectional case (T = 1), the simple formulation of standard errors offer an alternative to earlier formulations. [ABSTRACT FROM AUTHOR]

Limited-information methods are commonly used to estimate forward-looking models with rational expectations, such as the "New Keynesian Phillips Curve" of Galí and Gertler (1999). In this paper, we address issues of identification that have been overlooked due to the incompleteness of the single-equation formulation. We show that problems of weak instruments may arise, depending on the properties of the 'exogenous' variables, and that they are empirically relevant. We also uncover a link between identification and dynamic mis-specification, and examine the (lack of) power of Hansen's (1982) J test to detect invalid over-identifying restrictions. With regards to the New Phillips curve, we find that problems of identification cannot be ruled out, and they deserve further attention. [ABSTRACT FROM AUTHOR]

This paper analyzes the problem of weak instruments on identification, estimation, and inference in a simple nonparametric model of a triangular system. The paper derives a necessary and sufficient rank condition for identification, based on which weak identification is established. Then nonparametric weak instruments are defined as a sequence of reduced‐form functions where the associated rank shrinks to zero. The problem of weak instruments is characterized as concurvity, which motivates the introduction of a regularization scheme. The paper proposes a penalized series estimation method to alleviate the effects of weak instruments and shows that it achieves desirable asymptotic properties. A data‐driven procedure is proposed for the choice of the penalization parameter. The findings of this paper provide useful implications for empirical work. To illustrate them, Monte Carlo results are presented and an empirical example is given in which the effect of class size on test scores is estimated nonparametrically. [ABSTRACT FROM AUTHOR]

This paper introduces a new identification‐ and singularity‐robust conditional quasi‐likelihood ratio (SR‐CQLR) test and a new identification‐ and singularity‐robust Anderson and Rubin (1949) (SR‐AR) test for linear and nonlinear moment condition models. Both tests are very fast to compute. The paper shows that the tests have correct asymptotic size and are asymptotically similar (in a uniform sense) under very weak conditions. For example, in i.i.d. scenarios, all that is required is that the moment functions and their derivatives have 2 + γ bounded moments for some γ > 0. No conditions are placed on the expected Jacobian of the moment functions, on the eigenvalues of the variance matrix of the moment functions, or on the eigenvalues of the expected outer product of the (vectorized) orthogonalized sample Jacobian of the moment functions. The SR‐CQLR test is shown to be asymptotically efficient in a GMM sense under strong and semi‐strong identification (for all k ≥ p, where k and p are the numbers of moment conditions and parameters, respectively). The SR‐CQLR test reduces asymptotically to Moreira's CLR test when p = 1 in the homoskedastic linear IV model. The same is true for p ≥ 2 in most, but not all, identification scenarios. We also introduce versions of the SR‐CQLR and SR‐AR tests for subvector hypotheses and show that they have correct asymptotic size under the assumption that the parameters not under test are strongly identified. The subvector SR‐CQLR test is shown to be asymptotically efficient in a GMM sense under strong and semi‐strong identification. [ABSTRACT FROM AUTHOR]

The paper studies the asymptotic size property of various specification tests in linear structural models where instrumental variables may locally violate the exclusion restrictions. Our results provide some new insights and extensions of earlier studies. In particular, we derive an explicit formula of the asymptotic size of the tests which shows clearly the factors that influence their size under instrument endogeneity. We show that all tests have correct asymptotic size when the usual orthogonality condition holds, but their asymptotic size can be arbitrary large even if only one instrument is slightly correlated with the error term. We present a Monte Carlo experiment that confirms our theoretical findings. [ABSTRACT FROM AUTHOR]

We study subvector inference in the linear instrumental variables model assuming homoskedasticity but allowing for weak instruments. The subvector Anderson and Rubin (1949) test that uses chi square critical values with degrees of freedom reduced by the number of parameters not under test, proposed by Guggenberger, Kleibergen, Mavroeidis, and Chen (2012), controls size but is generally conservative. We propose a conditional subvector Anderson and Rubin test that uses data‐dependent critical values that adapt to the strength of identification of the parameters not under test. This test has correct size and strictly higher power than the subvector Anderson and Rubin test by Guggenberger et al. (2012). We provide tables with conditional critical values so that the new test is quick and easy to use. Application of our method to a model of risk preferences in development economics shows that it can strengthen empirical conclusions in practice. [ABSTRACT FROM AUTHOR]

This paper surveys the state of the art in the econometrics of regression models with many instruments or many regressors based on alternative – namely, dimension – asymptotics. We list critical results of dimension asymptotics that lead to better approximations of properties of familiar and alternative estimators and tests when the instruments and/or regressors are numerous. Then, we consider the problem of estimation and inference in the basic linear instrumental variables regression setup with many strong instruments. We describe the failures of conventional estimation and inference, as well as alternative tools that restore consistency and validity. We then add various other features to the basic model such as heteroskedasticity, instrument weakness, etc., in each case providing a review of the existing tools for proper estimation and inference. Subsequently, we consider a related but different problem of estimation and testing in a linear mean regression with many regressors. We also describe various extensions and connections to other settings, such as panel data models, spatial models, time series models, and so on. Finally, we provide practical guidance regarding which tools are most suitable to use in various situations when many instruments and/or regressors turn out to be an issue. [ABSTRACT FROM AUTHOR]

We study the finite-sample properties of tests for overidentifying restrictions in linear regression models with a single endogenous regressor and weak instruments. Under the assumption of Gaussian disturbances, we derive expressions for a variety of test statistics as functions of eight mutually independent random variables and two nuisance parameters. The distributions of the statistics are shown to have an ill-defined limit as the parameter that determines the strength of the instruments tends to zero and as the correlation between the disturbances of the structural and reduced-form equations tends to plus or minus one. This makes it impossible to perform reliable inference near the point at which the limit is ill-defined. Several bootstrap procedures are proposed. They alleviate the problem and allow reliable inference when the instruments are not too weak. We also study their power properties.

Copyright of EconoQuantum is the property of Universidad de Guadalajara, Centro Universitario de Ciencias Economico Administrativas and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this article, we consider inference in the linear instrumentalvariables models with one or more endogenous variables and potentially weak instruments. I developed a command, twostepweakiv, to implement the twostep identification-robust confidence sets proposed by Andrews (2018, Review of Economics and Statistics 100: 337-348) based on Wald tests and linear combination tests (Andrews, 2016, Econometrica 84: 2155-2182). Unlike popular procedures based on first-stage F statistics (Stock and Yogo, 2005, Testing for weak instruments in linear IV regression, in Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg), the two-step identificationrobust confidence sets control coverage distortion without assuming the data are homoskedastic. I demonstrate the use of twostepweakiv with an example of analyzing the effect of wages on married female labor supply. For inference on subsets of parameters, twostepweakiv also implements the refined projection method (Chaudhuri and Zivot, 2011, Journal of Econometrics 164: 239-251). I illustrate that this method is more powerful than the conventional projection method using Monte Carlo simulations. [ABSTRACT FROM AUTHOR]

Abstract: We analyze whether popular measures of narrative tax shocks can be treated as relevant instruments for observable endogenous tax series of interest. We find that narrative tax measures are only weakly correlated with cyclically adjusted tax revenues for the US and the UK. Using weak‐instrument robust inference, narrative tax measures often yield insignificant estimates of tax multipliers. We conclude that the literature currently understates the uncertainty associated with estimating the tax multiplier using the narrative approach. [ABSTRACT FROM AUTHOR]

Summary: In this paper, I investigate the estimation of linear regression models with strictly exogenous instruments under minimal identifying assumptions. I introduce a uniformly (in the data‐generating process) consistent estimator under nearly minimal identifying assumptions. The proposed estimator, called the integrated instrumental variables (IIV) estimator, is a simple weighted least‐squares estimator. It does not require the choice of a bandwidth or tuning parameter, or the selection of a finite set of instruments. Thus, the estimator is extremely simple to implement. Monte Carlo evidence supports the theoretical claims and suggests that the IIV estimator is a robust complement to optimal instrumental variables in finite samples. In an application with quarterly UK data, the IIV estimator estimates a positive and significant elasticity of intertemporal substitution and an equally sensible estimate for its reciprocal, in sharp contrast to instrumental variables methods that fail to identify these parameters. [ABSTRACT FROM AUTHOR]

We compare the finite sample performance of a number of Bayesian and classical procedures for limited information simultaneous equations models with weak instruments by a Monte Carlo study. We consider Bayesian approaches developed by Chao and Phillips, Geweke, Kleibergen and van Dijk, and Zellner. Amongst the sampling theory methods, OLS, 2SLS, LIML, Fuller’s modified LIML, and the jackknife instrumental variable estimator (JIVE) due to Angrist et al. and Blomquist and Dahlberg are also considered. Since the posterior densities and their conditionals in Chao and Phillips and Kleibergen and van Dijk are nonstandard, we use a novel “Gibbs within Metropolis−Hastings” algorithm, which only requires the availability of the conditional densities from the candidate generating density. Our results show that with very weak instruments, there is no single estimator that is superior to others in all cases. When endogeneity is weak, Zellner’s MELO does the best. When the endogeneity is not weak and ρ ω 12 > 0 , where ρ is the correlation coefficient between the structural and reduced form errors, and ω 12 is the covariance between the unrestricted reduced form errors, the Bayesian method of moments (BMOM) outperforms all other estimators by a wide margin. When the endogeneity is not weak and β ρ < 0 ( β being the structural parameter), the Kleibergen and van Dijk approach seems to work very well. Surprisingly, the performance of JIVE was disappointing in all our experiments.

Instrumental variable (IV) methods are widely used to deal with the issue of unmeasured confounding and are becoming popular in health and medical research. IV models are able to obtain consistent estimates in the presence of unmeasured confounding, but rely on assumptions that are hard to verify and often criticized. An instrument is a variable that influences or encourages individuals toward a particular treatment without directly affecting the outcome. Estimates obtained using instruments with a weak influence over the treatment are known to have larger small-sample bias and to be less robust to the critical IV assumption that the instrument is randomly assigned. In this work, we propose a weighting procedure for strengthening the instrument while matching. Through simulations, weighting is shown to strengthen the instrument and improve robustness of resulting estimates. Unlike existing methods, weighting is shown to increase instrument strength without compromising match quality. We illustrate the method in a study comparing mortality between kidney dialysis patients receiving hemodialysis or peritoneal dialysis as treatment for end-stage renal disease. [ABSTRACT FROM AUTHOR]

Instrumental variable analysis is an approach for obtaining causal inferences on the effect of an exposure (risk factor) on an outcome from observational data. It has gained in popularity over the past decade with the use of genetic variants as instrumental variables, known as Mendelian randomization. An instrumental variable is associated with the exposure, but not associated with any confounder of the exposure-outcome association, nor is there any causal pathway from the instrumental variable to the outcome other than via the exposure. Under the assumption that a single instrumental variable or a set of instrumental variables for the exposure is available, the causal effect of the exposure on the outcome can be estimated. There are several methods available for instrumental variable estimation; we consider the ratio method, two-stage methods, likelihood-based methods, and semi-parametric methods. Techniques for obtaining statistical inferences and confidence intervals are presented. The statistical properties of estimates from these methods are compared, and practical advice is given about choosing a suitable analysis method. In particular, bias and coverage properties of estimators are considered, especially with weak instruments. Settings particularly relevant to Mendelian randomization are prioritized in the paper, notably the scenario of a continuous exposure and a continuous or binary outcome. [ABSTRACT FROM AUTHOR]

We provide Monte Carlo evidence on the finite-sample behavior of the conditional empirical likelihood (CEL) estimator of Kitamura, Tripathi, and Ahn and the conditional Euclidean empirical likelihood (CEEL) estimator of Antoine, Bonnal, and Renault in the context of a heteroscedastic linear model with an endogenous regressor. We compare these estimators with three heteroscedasticity-consistent instrument-based estimators and the Donald, Imbens, and Newey estimator in terms of various performance measures. Our results suggest that the CEL and CEEL with fixed bandwidths may suffer from the no-moment problem, similarly to the unconditional generalized empirical likelihood estimators studied by Guggenberger. We also study the CEL and CEEL estimators with automatic bandwidths selected through cross-validation. We do not find evidence that these suffer from the no-moment problem. When the instruments are weak, we find CEL and CEEL to have finite-sample properties—in terms of mean squared error and coverage probability of confidence intervals—poorer than the heteroscedasticity-consistent Fuller (HFUL) estimator. In the strong instruments case, the CEL and CEEL estimators with automatic bandwidths tend to outperform HFUL in terms of mean squared error, while the reverse holds in terms of the coverage probability, although the differences in numerical performance are rather small. [ABSTRACT FROM PUBLISHER]

A standard test for weak instruments compares the first-stage F-statistic to a table of critical values obtained by Stock and Yogo (2005) using simulations. We derive a closed-form solution for the expectation from which these critical values are derived, as well as present some second-order asymptotic approximations that may be of value in the presence of multiple endogenous regressors. Inspection of this new result provides insights not available from simulation, and will allow software implementations to be generalised and improved. Finally, we explore the calculation of p-values for the first-stage F-statistic weak instruments test.

We consider the static linear panel data model with a single regressor. For this model, we derive the LIML estimator. We study the asymptotic behavior of this estimator under many-instruments asymptotics, by showing its consistency, deriving its asymptotic variance, and by presenting an estimator of the asymptotic variance that is consistent under many-instruments asymptotics. We briefly indicate the extension to the static panel data model with multiple regressors. [ABSTRACT FROM AUTHOR]

Economists rely frequently on instrumental variables estimation to overcome biases that endogenous explanatory variables cause in ordinary least squares estimation. However, traditional instrumental variables estimators, such as two-stage least squares and limited information maximum likelihood estimation, can suffer persistent estimator biases and size-of-test biases in even very large samples if the instruments used are large in number or are only weakly correlated with an endogenous explanatory variable. This paper reviews strategies for grappling with weak instruments and with large numbers of instruments in linear regression models. [ABSTRACT FROM AUTHOR]

Confidence intervals are commonly used to describe parameter uncertainty. In nonstandard problems, however, their frequentist coverage property does not guarantee that they do so in a reasonable fashion. For instance, confidence intervals may be empty or extremely short with positive probability, even if they are based on inverting powerful tests. We apply a betting framework and a notion of bet-proofness to formalize the 'reasonableness' of confidence intervals as descriptions of parameter uncertainty, and use it for two purposes. First, we quantify the violations of bet-proofness for previously suggested confidence intervals in nonstandard problems. Second, we derive alternative confidence sets that are bet-proof by construction. We apply our framework to several nonstandard problems involving weak instruments, near unit roots, and moment inequalities. We find that previously suggested confidence intervals are not bet-proof, and numerically determine alternative bet-proof confidence sets. [ABSTRACT FROM AUTHOR]

IO economists often estimate demand for differentiated products using data sets with a small number of large markets. This paper addresses the question of consistency and asymptotic distributions of instrumental variables estimates as the number of products increases in some commonly used models of demand under conditions on economic primitives. I show that, in a Bertrand-Nash equilibrium, product characteristics lose their identifying power as price instruments in the limit in certain cases, leading to inconsistent estimates. The reason is that product characteristic instruments achieve identification through correlation with markups, and, depending on the model of demand, the supply side can constrain markups to converge to a constant quickly relative to sampling error. I find that product characteristic instruments can yield consistent estimates in many of the cases I consider, but care must be taken in modeling demand and choosing instruments. A Monte Carlo study confirms that the asymptotic results are relevant in market sizes of practical importance. [ABSTRACT FROM AUTHOR]

Many articles which have estimated models with forward looking expectations have reported that the magnitude of the coefficients of the expectations term is very large when compared with the effects coming from past dynamics. This has sometimes been regarded as implausible and led to the feeling that the expectations coefficient is biased upwards. A relatively general argument that has been advanced is that the bias could be due to structural changes in the means of the variables entering the structural equation. An alternative explanation is that the bias comes from weak instruments. In this article, we investigate the issue of upward bias in the estimated coefficients of the expectations variable based on a model where we can see what causes the breaks and how to control for them. We conclude that weak instruments are the most likely cause of any bias and note that structural change can affect the quality of instruments. We also look at some empirical work in Castle et al. (2014) on the new Kaynesian Phillips curve (NYPC) in the Euro Area and U.S. assessing whether the smaller coefficient on expectations that Castle et al. (2014) highlight is due to structural change. Our conclusion is that it is not. Instead it comes from their addition of variables to the NKPC. After allowing for the fact that there are weak instruments in the estimated re-specified model, it would seem that the forward coefficient estimate is actually quite high rather than low. [ABSTRACT FROM AUTHOR]

Mendelian randomization is the use of genetic instrumental variables to obtain causal inferences from observational data. Two recent developments for combining information on multiple uncorrelated instrumental variables (IVs) into a single causal estimate are as follows: (i) allele scores, in which individual-level data on the IVs are aggregated into a univariate score, which is used as a single IV, and (ii) a summary statistic method, in which causal estimates calculated from each IV using summarized data are combined in an inverse-variance weighted meta-analysis. To avoid bias from weak instruments, unweighted and externally weighted allele scores have been recommended. Here, we propose equivalent approaches using summarized data and also provide extensions of the methods for use with correlated IVs. We investigate the impact of different choices of weights on the bias and precision of estimates in simulation studies. We show that allele score estimates can be reproduced using summarized data on genetic associations with the risk factor and the outcome. Estimates from the summary statistic method using external weights are biased towards the null when the weights are imprecisely estimated; in contrast, allele score estimates are unbiased. With equal or external weights, both methods provide appropriate tests of the null hypothesis of no causal effect even with large numbers of potentially weak instruments. We illustrate these methods using summarized data on the causal effect of low-density lipoprotein cholesterol on coronary heart disease risk. It is shown that a more precise causal estimate can be obtained using multiple genetic variants from a single gene region, even if the variants are correlated. [ABSTRACT FROM AUTHOR]

We examine statistical models, including the workhorse linear instrumental variables model, in which the mapping from the reduced form distribution to the structural parameters of interest is singular. The singularity of this mapping implies certain fundamental restrictions on the finite sample properties of point estimators: they cannot be unbiased, quantile-unbiased, or translation equivariant. The nonexistence of unbiased estimators does not rule out bias reduction of standard estimators, but implies that the bias-variance tradeoff cannot be avoided and needs to be considered carefully. The results can also be extended to weak instrument asymptotics by using the limits of experiments framework. [ABSTRACT FROM AUTHOR]

We study the asymptotic validity of the bootstrap for Durbin-Wu-Hausman tests of exogeneity, with or without identification. We provide an analysis of the limiting distributions of the proposed bootstrap statistics under both the null hypothesis of exogeneity (size) and the alternative hypothesis of endogeneity (power). We show that when identification is strong, the bootstrap provides a high-order approximation of the null limiting distributions of the statistics and is consistent under the alternative hypothesis if the endogeneity parameter is fixed. However, the bootstrap only provides a first-order approximation when instruments are weak. Moreover, we provide the necessary and sufficient condition under which the proposed bootstrap tests exhibit power under (fixed) endogeneity and weak instruments. The latter condition may still hold over a wide range of cases as long as at least one instrument is relevant. Nevertheless, all bootstrap tests have low power when all instruments are irrelevant, a case of little interest in empirical work. We present a Monte Carlo experiment that confirms our theoretical findings. [ABSTRACT FROM AUTHOR]