We discuss an estimation procedure for continuous-time models based on discrete sampled data with a fixed unit of time between two consecutive observations. Because in general the conditional likelihood of the model cannot be derived, an indirect inference procedure following Gouriéroux, Monfort, and Renault (1993, Journal of Applied Econometrics 8, 85–118) is developed. It is based on simulations of a discretized model. We study the asymptotic properties of this “quasi”-indirect estimator and examine some particular cases. Because this method critically depends on simulations, we pay particular attention to the appropriate choice of the simulation step. Finally, finite-sample properties are studied through Monte Carlo experiments.

The nonparametric and the nuisance parameter approaches to consistently testing statistical models are both attempts to estimate topological measures of distance between a parametric and a nonparametric fit, and neither dominates in experiments. This topological unification allows us to greatly extend the nuisance parameter approach. How and why the nuisance parameter approach works and how it can be extended bear closely on recent developments in artificial neural networks. Statistical content is provided by viewing specification tests with nuisance parameters as tests of hypotheses about Banach-valued random elements and applying the Banach central limit theorem and law of iterated logarithm, leading to simple procedures that can be used as a guide to when computationally more elaborate procedures may be warranted.

Many estimators and tests are of the form of a ratio of quadratic forms in normal variables. Excepting a few very special cases little is known about the density or distribution of these ratios, particularly if we allow for noncentrality in the quadratic forms. This paper assumes this generality and derives saddlepoint approximations for this class of statistics. We first derive and prove the existence of an exact inversion based on the joint characteristic function. Then the saddlepoint algorithm is applied and the leading term found, and analytic justification of the asymptotic nature of the approximation is given. As an illustration we consider the calculation of sizes and powers of F-tests, where a new exact result is found.

Econometric Theory is proud to announce the winning article for the Tjalling C. Koopmans Econometric Theory Prize over the period 1994–1996 (inclusive). The prize is jointly supported by the publishers, Cambridge University Press, and Mrs. Truus Koopmans. It is named in honor of Tjalling C. Koopmans, the 1975 Nobel Laureate in Economic Science. The selection of the winning article was made by the Advisory Board of the Journal, and all articles published in Econometric Theory (1994–1996) were candidates for the prize, except those that were authored or coauthored by the Editor and members of the Advisory Board. The prize is accompanied by a financial award of $500 to each of the winning authors.

This paper updates Hall's (1987, Econometric Theory 3, 171–194; 1990, Econometric Theory 6, 1–16) rankings of academic institutions by publication activity in econometrics over the period 1989–1995. Fifteen leading international journals that publish econometrics articles are used to provide the database. The rankings are based on standardized page counts of articles published in these journals over the stated period. Separate rankings are provided for theoretical contributions in econometrics and for all contributions in econometrics. An “Econometricians' Hall of Fame” is developed listing the top 100 individual producers of theoretical econometrics and all contributions in econometrics research in the 15 journals considered. Separate rankings are also provided both for institutions and for individuals according to all econometrics publication activity in Econometrica, Econometric Theory, and Journal of Econometrics.

We study the convergence in distribution of M-estimators over a convex kernel. Under convexity, the limit distribution of M-estimators can be obtained under minimal assumptions. We consider the case when the limit is arbitrary, not necessarily normal. If some Taylor expansions hold, the limit distribution is stable. As an application, we examine the limit distribution of M-estimators for the multivariate linear regression model. We obtain the distributional convergence of M-estimators for the multivariate linear regression model for a wide range of sequences of regressors and different types of conditions on the sequence of errors.

This paper investigates the behavior of the maximum likelihood estimator of a Gaussian autoregressive moving average model with a unit root in the moving average polynomial. The results are primarily of interest in testing hypotheses that involve moving average unit roots as, for example, when testing for stationarity of a series.

Diffuse priors lead to pathological posterior behavior when used in Bayesian analyses of simultaneous equation models (SEM's). This results from the local nonidentification of certain parameters in SEM's. When this a priori known feature is not captured appropriately, it results in an a posteriori favoring of certain specific parameter values that is not the consequence of strong data information but of local nonidentification. We show that a proper consistent Bayesian analysis of a SEM explicitly has to consider the reduced form of the SEM as a standard linear model on which nonlinear (reduced rank) restrictions are imposed, which result from a singular value decomposition. The priors/posteriors of the parameters of the SEM are therefore proportional to the priors/posteriors of the parameters of the linear model under the condition that the restrictions hold. This leads to a framework for constructing priors and posteriors for the parameters of SEM's. The framework is used to construct priors and posteriors for one, two, and three structural equation SEM's. These examples together with a theorem, showing that the reduced forms of SEM's accord with sets of reduced rank restrictions on standard linear models, show how Bayesian analyses of generally specified SEM's can be conducted.

A semiparametric likelihood method is proposed for the estimation of sample selection models. The method is a two-step semiparametric scoring estimation procedure based on an index restriction and kernel estimation. Under some regularity conditions, the estimator is square-root n-consistent and asymptotically normal. The estimator is also asymptotically efficient in the sense that its asymptotic covariance matrix attains the semiparametric efficiency bound under the index restriction. For the binary choice sample selection model, it also attains the efficiency bound under the independence assumption. This method can be applied to the estimation of general sample selection models.

Many unit root and cointegration tests require an estimate of the spectral density function at frequency zero of some process. Commonly used are kernel estimators based on weighted sums of autocovariances constructed using estimated residuals from an AR(1) regression. However, it is known that with substantially correlated errors, the OLS estimate of the AR(1) parameter is severely biased. In this paper, we first show that this least-squares bias induces a significant increase in the bias and mean-squared error (MSE) of kernel-based estimators. We then consider a variant of the autoregressive spectral density estimator that does not share these shortcomings because it bypasses the use of the estimate from the AR(1) regression. Simulations and local asymptotic analyses show its bias and MSE to be much smaller than those of a kernel-based estimator when there is strong negative serial correlation. We also include a discussion about the appropriate choice of the truncation lag.

This paper considers adaptive maximum likelihood estimation of reduced rank vector error correction models. It is shown that such models can be asymptotically efficiently estimated even in the absence of knowledge of the shape of the density function of the innovation sequence, provided that this density is symmetric. The construction of the estimator, involving the nonparametric kernel estimation of the unknown density using the residuals of a consistent preliminary estimator, is described, and its asymptotic distribution is derived. Asymptotic efficiency gains over the Gaussian pseudo–maximum likelihood estimator are evaluated for elliptically symmetric innovations.

This paper provides a polynomial factorization theorem that is then used to extend the characterization parts of the parametric representation theorems of Johansen (1992, Econometric Theory 8, 188–202) for vector autoregressive processes integrated of up to order 2. A characterization theorem is provided in the general case of an I(d) process. For the discussion of the complicated polynomial cointegration properties of such processes, the case of an I(3) process is considered as an example.

The distribution of sums of squared random variables with heavy-tailed distributions is investigated. Considering random variables in the domain of attraction of a stable Paretian law we derive the limiting distribution as the degrees of freedom approach infinity. The finite-degrees-of-freedom behavior for stable Paretian variates is simulated. Response surface techniques are employed to compactly summarize the simulation results for a relevant range of significance levels.

In this paper we consider a semiparametric version of the test for seasonal unit roots suggested by Hylleberg, Engle, Granger, and Yoo (1990, Journal of Econometrics 44, 215–238). The asymptotic theory is based on the analysis of a simple regression problem, and the results apply to tests at any given frequency in the range (0,π]. Monte Carlo simulations suggest that the test may have more power than the parametric test of Hylleberg et al. (1990). On the other hand, the semiparametric version suffers from severe size distortions in some situations.

Given an observation of a discrete-time process {Yi,i = 0...n} assumed to be Markov, stationary, and time reversible, we develop a (conservative) test procedure of embeddability by a continuous-time reversible Markov process. The test statistic is derived from a set of moment inequality restrictions implied by the spectral properties of such continuous-time processes. Most interesting is that the embeddability condition of interest is a direct extension of the well-known embeddability problem by a two-state Markov chain. Empirical experiments show that the embeddability hypothesis is rejected more frequently for exchange rate daily data than for stock indices data.

This paper deals with the asymptotic properties of quasi-maximum likelihood estimators for multivariate heteroskedastic models. For a general model, we give conditions under which strong consistency can be obtained; unlike in the current literature, the assumptions on the existence of moments of the error term are weak, and no study of the various derivatives of the likelihood is required. Then, for a particular model, the multivariate GARCH model with constant correlation, we describe the set of parameters where these conditions hold.

We consider a linear regression model with regression parameters (θ1,...,θp) and error variance parameter σ2. Our aim is to find a confidence interval with minimum coverage probability 1 − α for a parameter of interest θ1 in the presence of nuisance parameters (θ2,...,θp,σ2). We consider two confidence intervals, the first of which is the standard confidence interval for θ1 with coverage probability 1 − α. The second confidence interval for θ1 is obtained after a variable selection procedure has been applied to θp. This interval is chosen to be as short as possible subject to the constraint that it has minimum coverage probability 1 − α. The confidence intervals are compared using a risk function that is defined as a scaled version of the expected length of the confidence interval. We show that, subject to certain conditions including that [(dimension of response vector) − p] is small, the second confidence interval is preferable to the first when we anticipate (without being certain) that |θp|/σ is small. This comparison of confidence intervals is shown to be mathematically equivalent to a corresponding comparison of prediction intervals.

In this paper, we study the bias-corrected test developed in Fan (1994). It is based on the integrated squared difference between a kernel estimator of the unknown density function of a random vector and a kernel smoothed estimator of the parametric density function to be tested under the null hypothesis. We provide an alternative asymptotic approximation of the finite-sample distribution of this test by fixing the smoothing parameter. In contrast to the normal approximation obtained in Fan (1994) in which the smoothing parameter shrinks to zero as the sample size grows to infinity, we obtain a non-normal asymptotic distribution for the bias-corrected test. A parametric bootstrap procedure is proposed to approximate the critical values of this test. We show both analytically and by simulation that the proposed bootstrap procedure works. Consistency and local power properties of the bias-corrected test with a fixed smoothing parameter are also discussed.

In this article we present a new method for computing the theoretical autocovariance function of an autoregressive moving average model. The importance of our theorem is that it yields two interesting results: First, a closed-form solution is derived in terms of the roots of the autoregressive polynomial and the parameters of the moving average part. Second, a sufficient condition for the lack of model redundancy is obtained.

This paper considers tests for structural change of the cointegrating vector and the adjustment vector in the error correction model with an unknown change point. This paper derives new tests for structural change, which are applicable to maximum likelihood estimation. Our tests for structural change of the cointegrating vector have the same nonstandard asymptotic distributions that have been found by Hansen (1992a, Journal of Business and Economic Statistics 10, 321–335). In contrast, the tests on the adjustment vector have the same asymptotic distributions that have been found by Andrews and Ploberger (1994, Econometrica 62, 1383–1414) for models with stationary variables. Asymptotic critical values are provided.