Simultaneous-equation estimation with no instrumental variables

For a single equation in a system of linear equations, estimation by instrumental variables is the standard approach. In practice, however, it is often difficult to find valid instruments. This letter outlines a maximum-likelihood method that does not require instrumental variables; it is applied to the estimation of a demand function and an Engel curve.     

An asymptotic variance inequality for instrumental variable estimators signaling proportional bias increases

An asymptotic variance inequality for instrumental variable (IV) estimators is proposed, which suggests a critical variance that signals proportional increases in the bias of IV estimators through the augmentation of a set of instruments.     

Iterative algorithm for non parametric estimation of the instrumental variables quantiles

This paper proposes a simple algorithm for the numerical computation of the non parametric IV quantile estimation. This algorithm is based on the Landweber iterations for solving a nonlinear integral equation. The paper is illustrated by numerical simulations.     

On adding over-identifying instrumental variables to simultaneous equations

Reducing the number of over-identifying instruments, or adding them to a structural equation, increases estimation dispersion. Added instruments should be insignificant under correct specification, with parameter estimates nearly unaffected, confirmed by Monte Carlo. Selecting instruments does not affect these results.     

General instrumental variables estimation under stochastic linear restrictions

This letter extends the Theil-Goldberger ‘mixed’ regression estimator, for models subject to stochastic linear restrictions, to the case of stochastic regressors. A general instrumental variables ‘mixed’ estimator is discussed. The asymptotic distribution of the estimator is obtained, and an asymptotic test of the compatibility of the sample and prior information is presented.     

Cumulant instrument estimators for hedge fund return models with errors in variables

We revisit the factors incorporated in asset pricing models following the recent developments in financial markets – i.e., the rise of shadow banking and the change in the transmission channel of monetary policy. We propose two versions of the Fung and Hsieh (2004) hedge fund return model, especially an augmented market model which accounts for the new dynamics of financial markets and the procyclicality of hedge fund returns. We run these models with an innovative Hausman procedure, tackling the measurement errors embedded in the models factor loadings. Our empirical method also allows for confronting the drawbacks of the instruments used to estimate hedge fund asset pricing models.     

A quantile regression approach for estimating panel data models using instrumental variables

We introduce a quantile regression approach to panel data models with endogenous variables and individual effects correlated with the independent variables. We find newly developed quantile regression methods can be easily adapted to estimate this class of models efficiently.     

Exact equivalence of instrumental variable estimators in an error component structural system

The paper establishes an equivalence result in the context of anm-equation error component structural system, whose disturbances have the usual three-component structure, and whose equations feature explanatory variables of the formz _i, z_t andz _it; the latter vary (respectively) only over individuals, only over time, and over both. Under the stochastic specification assumed, it is shown that the alternative instrumental variables (IV) estimators commonly used in the special cases of this system are all equivalent (numerically identical); the result is a generalization of the equivalences established previously for the special cases. In the single equation (m=1) context, the equivalence requires that the IV set contain variables of the formz _i and/orz _t, and further, in numbers determined by the ranks of (respectively) the individuals-mean and time-mean matrices of the instruments. If such an IV set is common to all equations, the equivalence also holds for the system under joint estimation. The result is used to recommend a couple of estimators for use in panel data, on grounds of computational simplicity.This is a revision of the December 1990 draft with the same title, and is a substantial revision of the April 1990 version entitled: Analysis of an error component structural system. This revision has benefited from comments received from a referee and a editor of this journal. I came to know from an anonymous reader that the equivalence criterion developed in my 1990 a article, used here and the two earlier versions, was infact obtained previously in an unpublished paper by Balestra (1988). Balestra's paper, which was made available to me by Badi Baltagi at the time of this revision, and subsequently by Balestra, considers the equivalence of ^b, ^c and one other estimator which differs from our ^a. Errors, if any, are my responsibility.     

Viewpoint: An extended class of instrumental variables for the estimation of causal effects

Abstract We examine how structural systems can yield observed variables instrumental in identifying and estimating causal effects. We provide an exhaustive characterization of potentially identifying conditional exogeneity relationships and demonstrate how structural relations determine exogeneity and exclusion restrictions that yield moment conditions supporting identification. This provides a comprehensive framework for constructing instruments and covariates. We introduce notions of conditioning and conditional extended instrumental variables (XIVs). These permit identification but need not be traditional instruments, as they may be endogenous. We distinguish between observed XIVs and proxies for unobserved XIVs. A main message is the importance of sufficiently specifying causal relations governing the unobservables.     

A test of the independence of subsets of instrumental variables and regressors

A Wald statistic is presented to test for the independence of subsets of regressors and instrumental variables and the disturbance in a regression equation. The test does not require that the equation be identified under maintained conditions.     

Moments of OLS estimators in an autoregressive moving average model with explanatory variables

In this paper we derive the moments of the ordinary least squares (OLS) estimators in an autoregressive moving average model by a straightforward technique compared to the one used in Carter and Ullah (1979). The model contains exogenous variables and the technique also provides simpler moment expressions and can be used to derive the moments in more general dynamic models.     

The comparative performance of the iterative instrumental variables estimator using structural and predictive criteria

The purpose of this paper is to give experimental evidence on the small-sample properties of the iterative instrumental variables estimator originally proposed byLyttkens [1970], relative to the more conventional methods including ordinary least squares, limited information single equation maximum likelihood and three stage least squares.     

An instrumental variables approach to estimating tax revenue elasticities: Evidence from Sub-Saharan Africa

This paper exploits the significant response of real GDP growth of Sub-Saharan African countries to exogenous international commodity price and rainfall shocks to construct instrumental variables estimates of the tax revenue elasticity IV estimates yield that a 1% increase in GDP increases tax revenues by up to 2.5%.     

Individual heterogeneity in the returns to schooling: instrumental variables quantile regression using twins data

Considerable effort has been exercised in estimating mean returns to education while carefully considering biases arising from unmeasured ability and measurement error. Recent work has investigated whether there are variations from the “mean” return to education across the population with mixed results. We use an instrumental variables estimator for quantile regression on a sample of twins to estimate an entire family of returns to education at different quantiles of the conditional distribution of wages while addressing simultaneity and measurement error biases. We test whether there is individual heterogeneity in returns to education and find that: more able individuals obtain more schooling perhaps due to lower marginal costs and/or higher marginal benefits of schooling and that higher ability individuals (those further to the right in the conditional distribution of wages) have higher returns to schooling consistent with a non-trivial interaction between schooling and unobserved abilities in the generation of earnings. The estimated returns are never lower than 9 percent and can be as high as 13 percent at the top of the conditional distribution of wages but they vary significantly only along the lower to middle quantiles. Our findings may have meaningful implications for the design of educational policies.     

Rainy day politics. An instrumental variables approach to the effect of parties on political outcomes

Scholars have grappled with the question of how parties affect policy. Here I propose and test an instrumental variable approach using rainfall. In Norwegian municipal elections, potential left wing voters are likely to abstain from voting with election day rain, whereas the opposite holds for right-wingers. Then rainfall provides an exogenous source of variation, and hence an instrument, for the party composition of the municipal council. A strengthening of the right wing parties due to rainfall shifts expenditures toward education, but reduces total spending. This also shows that political competition does not drive party platforms to converge.     

Human capital,innovation and productivity in Colombian enterprises: a structural approach using instrumental variables

ABSTRACT

In this paper we explore the R&D–innovation–productivity linkage for the Colombian manufacturing industry, paying special attention to the role of human capital. Using data from two firm-level surveys, the Survey of Development and Technological Innovation (EDIT) and the Annual Manufacturing Survey (EAM), we extend the model of Crépon, Duguet, and Mairesse [1998. ‘Research, Innovation and Productivity: An Econometric Analysis at the Firm Level.’ Economics of Innovation and New Technology 7 (2): 115–158] (hereafter CDM) by including human capital at the investment decision stage. We implement an instrumental variable methodology to correct the potential endogeneity that may arise with the inclusion of human capital. Our results suggest that human capital has a causal effect on research and development (R&D) investment decisions, the innovation behavior of the firm, and increases the labor productivity of the firm. The conclusions highlight the relevance of human capital in the surrounding literature which stands in contrast to prior work that has not included this variable.     

A selective review of the first 20 years of instrumental variables models in health-services research and medicine

Abstract

Background:

The method of instrumental variables (IV) is useful for estimating causal effects. Intuitively, it exploits exogenous variation in the treatment, sometimes called natural experiments or instruments. This study reviews the literature in health-services research and medical research that applies the method of instrumental variables, documents trends in its use, and offers examples of various types of instruments.     

Modified Hausman tests for inefficient estimators

Widely-used estimators such as the generalized method of moments (GMM) and quasi-maximum likelihood (QML) are not efficient in general. This study shows how the Hausman specification test may easily be corrected to be used with inefficient estimators. It introduces a related test that has better power, and it points out the relationship to the GMM criterion test of specification of overidentified models. A brief simulation illustrates the results.     

Inference approaches for instrumental variable quantile regression

We consider asymptotic and finite sample confidence bounds in instrumental variables quantile regressions of wages on schooling with relatively weak instruments. We find practically important differences between the asymptotic and finite sample interval estimates.     

Estimation of an endogenous switching regression model with discrete dependent variables: Monte-Carlo analysis and empirical application of three estimators

The performances of alternative two-stage estimators for the endogenous switching regression model with discrete dependent variables are compared, with regard to their usefulness as starting values for maximum likelihood estimation. This is especially important in the presence of large correlation coefficients, in which case maximum likelihood procedures have difficulties to converge. Monte-Carlo simulations indicate that an estimator that corrects for conditional heteroskedasticity of the residuals is superior in almost all instances, and especially when maximum likelihood is problematic. This result is also obtained in an empirical example in which off-farm work participation equations of farm women are conditional on farm work participation status. First version received: July 1995/final version received: March 1998