One-step efficient GMM estimation has been developed in the recent papers of Back and Brown (1990), Imbens (1993) and Qin and Lawless (1994). These papers emphasized methods that correspond to using Owen's (1988) method of empirical likelihood to reweight the data so that the reweighted sample obeys all the moment restrictions at the parameter estimates. In this paper we consider an alternative KLIC motivated weighting and show how it and similar discrete reweightings define a class of unconstrained optimization problems which includes GMM as a special case. Such KLIC-- motivated reweightings introduce M auxiliary `tilting' parameters, where M is the number of moments; parameter and overidentification hypotheses can be recast in terms of these tilting parameters. Such tests, when appropriately conditioned on the estimates of the original parameters, are often startlingly more effective than their conventional counterparts. This is apparently due to the local ancillarity of the original parameters for the tilting parameters. 1.

Abstract. We discuss instrumental variables (IV) estimation in the broader context of the generalized method of moments (GMM), and describe an extended IV estimation routine that provides GMM estimates as well as additional diagnostic tests. Stand–alone test procedures for heteroskedasticity, overidentification, and endogeneity in the IV context are also described.

A leading explanation of aggregate stock market behavior suggests that assets are priced as if there were a representative investor whose utility is a power function of the difference between aggregate consumption and a “habit" level, where the habit is some function of lagged and (possibly) contemporaneous consumption. But theory does not provide precise guidelines about the parametric functional relationship between the habit and aggregate consumption. This makes formal estimation and testing challenging; at the same time, it raises an empirical question about the functional form of the habit that best explains asset pricing data. This paper studies the ability of a general class of habit-based asset pricing models to match the conditional moment restrictions implied by asset pricing theory. Our approach is to treat the functional form of the habit as unknown, and to estimate it along with the rest

This paper tests different models of vertical contracting between manufacturers and retailers in the supermarket industry. I estimate demand and use the estimates to compute price-cost margins for retailers and manufacturers under different supply models without observing wholesale prices. I then test which set of margins seems to be compatible with the margins obtained from direct estimates of cost and select the best among the non-nested competing models. The models considered are: (1) a double marginalization pricing model; (2) a vertically integrated model; and (3) a variety of alternative (strategic) supply scenarios, allowing for collusion, non-linear pricing and strategic behavior with respect to private label products. Using data on yogurt sold at several stores in a large urban area of the United States, I find that wholesale prices are close to marginal cost and that retailers have pricing power in the vertical chain. This is consistent with non-linear pricing by the manufacturers or with high bargaining power of the retailers.

Gallant and Tauchen (1996) describe an estimation technique, known as Efficient Method of Moments (EMM), that uses numerical methods to estimate parameters of a structural model by using as matching conditions (moments in the GMM jargon) the gradients of an auxiliary model that fits a subset of variables that may be simulated from the structural model. This paper presents three Monte Carlo experiments to asses the finite sample properties of EMM. The first one compares it with a fully efficient procedure (Maximum Likelihood) by estimating an invertible MA process. The second and third experiments, compare the finite sample properties of the EMM estimators with those of GMM by using stochastic volatility models and consumption-based asset pricing models. The experiments show that the gains in efficiency are impressive; however, given that both EMM and GMM share the same type of objective function, finite sample inference based on asymptotic theory continues to lead, in some cases, to “over rejections ” even though they are not as significant as in GMM.

In a competitive marketplace, the effectiveness of any element of the marketing mix is determined not only by its absolute value, but also by its relative value with respect to the competition. For example, the effectiveness of a price cut in increasing demand is critically related to competitors ’ reaction to the price change. Managers therefore need to know the nature of competitive interactions among firms. In this paper, we take a theory-driven empirical approach to gain a deeper understanding of the competitive pricing behavior in the U.S. auto market. The ability-motivation paradigm posits that a firm needs both the ability and the motivation to succeed in implementing a strategy (Boulding and Staelin 1995). We use arguments from the game-theoretic literature to understand firm motivation and abilities

Abstract not found

We propose a class of new robust Generalized Method of Moments (GMM) tests for endogenous structural breaks. The tests are based on supremum, average and exponential functionals derived from robust GMM estimators with bounded influence function. We study the theoretical local robustness properties of the new tests and show that they imply a uniformly bounded asymptotic sensitivity of size and power under general local deviations from a reference model. We then analyze the finite sample performance of the new robust tests via Monte Carlo simulations, and compare it with that of classical GMM tests for structural breaks. In large samples, we find that the performance of classical asymptotic GMM tests can be quite unstable under slight departures from some given reference distribution. In particular, the loss in power can be substantial in some models. Robust asymptotic tests for structural breaks yield important power improvements both in exactly identified and overidentified model settings. In small samples, bootstrapped versions of the classical and the robust GMM tests provide accurate and stable empirical levels also for quite small sample sizes. However, bootstrapped robust GMM tests are found to provide again a higher finite sample efficiency.

We obtain upper bounds on the (conditional) limiting distributions of identification robust GMM statistics for testing hypothezes that are specified on subsets of the parameters. These upper bounds correspond to the (conditional) limiting distributions that result when the unrestricted parameters are well identified. They lead to more powerful tests than those that result from using projection arguments on tests on all the parameters. The upper bounds only apply when the unrestricted parameters are estimated using the continuous updating estimator. The critical values that result from the upper bounds lead to conservative tests when the unrestricted parameters are not well-identified. The identification robust subset GMM statistics resemble identification statistics when we evaluate them at a value of the hypothesized parameter that is distant from the true one. The power of these statistics is therefore governed by the least identified parameter so a weakly identified parameter implies that the power for tests on any of the parameters is low. 1

this paper, we develop and apply methods for addressing two core problems in evaluating the quality of medical care: determining the effect of medical treatments on patient outcomes, and determining whether and why health outcomes differ across health care providers. Randomized controlled trials are generally regarded as the only definitive method for estimating treatment effects. However, in response to a variety of perceived problems with randomized controlled trials, ranging from their expense and timeliness to their external validity and the "ethical" problems of conducting trials for many commonly-used treatments, some investigators have advocated supplementing clinical-trial results with analyses of various types of observational data. As medical data systems improve, increasingly detailed observational data on large patient populations have become available for potential use in such studies. These data innovations do not eliminate the central problem with observational data: treatments are not randomly assigned across patients and are likely to be related to hard-to-measure factors that also influence outcomes. Two major approaches have been taken to this problem of endogeneity in estimating treatment effects (see McClellan and Noguchi, 2000). One approach has been to collect detailed patient information (e.g. from charts) and control directly for risk factors other than treatment that may influence patient outcomes. This approach is inevitably limited because it is difficult to measure all of the patient preferences, comorbidities, and other characteristics that may influence both treatment choice and outcomes; relative to the many complex patient characteristics that are observed by patients and their health care providers and that may influence treatment choic...