This paper examines several practical issues regarding the implementation of the Phillips and Hansen fully modified least squares (FMLS) method for the estimation of a cointegrating vector. Various versions of this method arise by selecting between standard and prewhitened kernel estimation and between parametric and nonparametric automatic bandwidth estimators and also among alternative kernels. A Monte Carlo study is conducted to investigate the finite-sample properties of the alternative versions of the FMLS procedure. The results suggest that the prewhitened kernel estimator of Andrews and Monahan (1992, Econometrica 60, 953–966) in which the bandwidth parameter is selected via the nonparametric procedure of Newey and West (1994, Review of Economic Studies 61, 631–653) minimizes the second-order asymptotic bias effects.

High breakdown point estimators in regression are robust against gross contamination in the regressors and also in the errors; the least median of squares (LMS) estimator has the additional property of packing the majority of the sample most tightly around the estimated regression hyperplane in terms of absolute deviations of the residuals and thus is helpful in identifying outliers. Asymptotics for a class of high breakdown point smoothed LMS estimators are derived here under a variety of conditions that allow for time series applications; joint limit processes for several smoothed estimators are examined. The limit process for the LMS estimator is represented via a generalized Gaussian process that defines the generalized derivative of the Wiener process.

We introduce a kernel-based estimator of the density function and regression function for data that have been grouped into family totals. We allow for a common intrafamily component but require that observations from different families be independent. We establish consistency and asymptotic normality for our procedures. As usual, the rates of convergence can be very slow depending on the behavior of the characteristic function at infinity. We investigate the practical performance of our method in a simple Monte Carlo experiment.

This paper investigates the limiting properties of the Canova and Hansen test, testing for the null hypothesis of no unit root against seasonal unit roots, under a sequence of local alternatives with the model extended to have seasonal dummies and trends or no deterministic term and also only seasonal dummies. We derive the limiting distribution of the test statistic and its characteristic function under local alternatives. We find that the local limiting power is an inverse function of the spectral density at frequency π (π/2) when we test against a negative unit root (annual unit roots). We also theoretically show that the local limiting power of the Canova and Hansen test against a negative unit root (annual unit roots) does not increase when the true process has annual unit roots (a negative unit root) but not a negative unit root (annual unit roots), which has been observed in Monte Carlo simulations in such research as Caner (1998, Journal of Business and Economic Statistics 16, 349–356), Canova and Hansen (1995, Journal of Business and Economic Statistics 13, 237–252), and Hylleberg (1995, Journal of Econometrics 69, 5–25).

In this paper, I derive the efficiency bound of the structural parameter in a linear simultaneous equations model with many instruments. The bound is derived by applying a convolution theorem to Bekker's (1994, Econometrica 62, 657–681) asymptotic approximation, where the number of instruments grows to infinity at the same rate as the sample size. Usual instrumental variables estimators with a small number of instruments are heuristically argued to be efficient estimators in the sense that their asymptotic distribution is minimal. Bayesian estimators based on parameter orthogonalization are heuristically argued to be inefficient.

Consistency of kernel estimators of the long-run covariance matrix of a linear process is established under weak moment and memory conditions. In addition, it is pointed out that some existing consistency proofs are in error as they stand.

The problem addressed in this paper is to test the null hypothesis that a time series process is uncorrelated up to lag K in the presence of statistical dependence. We propose an extension of the Box–Pierce Q-test that is asymptotically distributed as chi-square when the null is true for a very general class of dependent processes that includes non-martingale difference sequences. The test is based on a consistent estimator of the asymptotic covariance matrix of the sample autocorrelations under the null. The finite sample performance of this extension is investigated in a Monte Carlo study.

This paper considers the problem of choosing the number of bootstrap repetitions B to use with the BCa bootstrap confidence intervals introduced by Efron (1987, Journal of the American Statistical Association 82, 171–200). Because the simulated random variables are ancillary, we seek a choice of B that yields a confidence interval that is close to the ideal bootstrap confidence interval for which B = ∞. We specify a three-step method of choosing B that ensures that the lower and upper lengths of the confidence interval deviate from those of the ideal bootstrap confidence interval by at most a small percentage with high probability.

Phoebus J. Dhrymes is one of the best known econometricians of the last 40 years. He has made substantial contributions to econometric theory through articles in leading journals and by way of a series of outstanding texts on the foundations and methods of econometrics. His early research began with an applied econometric focus on problems of production and investment. His later contributions concentrated on the foundations of econometric methodology, including systems of simultaneous equations. Throughout the econometrics community, Dhrymes is well known for his influential textbooks, some of which have been translated into several languages. His 1970 book Econometrics: Statistical Foundations and Applications provided an accessible and rigorous foundation for both students and teachers of econometrics. His subsequent books have continued to treat foundational issues and have tracked new areas of econometric interest through to his 1998 book Time Series, Unit Roots, and Cointegration. Reading his books reveals Dhrymes as a teacher, synthesizer, and master expositor. As he says in the interview that follows, “my books are not typical textbooks. I perceive them more as books that bridge the gap between ordinary textbooks and journal articles and as filters that distill and synthesize the wisdom of many contributors to the subject. On this score I was influenced in my writing by the way I learn when studying by myself.”

In this paper we study nonparametric estimation of regression quantiles for time series data by inverting a weighted Nadaraya–Watson (WNW) estimator of conditional distribution function, which was first used by Hall, Wolff, and Yao (1999, Journal of the American Statistical Association 94, 154–163). First, under some regularity conditions, we establish the asymptotic normality and weak consistency of the WNW conditional distribution estimator for α-mixing time series at both boundary and interior points, and we show that the WNW conditional distribution estimator not only preserves the bias, variance, and, more important, automatic good boundary behavior properties of local linear “double-kernel” estimators introduced by Yu and Jones (1998, Journal of the American Statistical Association 93, 228–237), but also has the additional advantage of always being a distribution itself. Second, it is shown that under some regularity conditions, the WNW conditional quantile estimator is weakly consistent and normally distributed and that it inherits all good properties from the WNW conditional distribution estimator. A small simulation study is carried out to illustrate the performance of the estimates, and a real example is also used to demonstrate the methodology.

Kernel smoothing techniques free the traditional parametric estimators of volatility from the constraints related to their specific models. In this paper the nonparametric local exponential estimator is applied to estimate conditional volatility functions, ensuring its nonnegativity. Its asymptotic properties are established and compared with those for the local linear estimator. It theoretically enables us to determine when the exponential is expected to be superior to the linear estimator. A very strong and novel result is achieved: the exponential estimator is asymptotically fully adaptive to unknown conditional mean functions. Also, our simulation study shows superior performance of the exponential estimator.

This paper establishes an invariance principle applicable for the asymptotic analysis of sieve bootstrap in time series. The sieve bootstrap is based on the approximation of a linear process by a finite autoregressive process of order increasing with the sample size, and resampling from the approximated autoregression. In this context, we prove an invariance principle for the bootstrap samples obtained from the approximated autoregressive process. It is of the strong form and holds almost surely for all sample realizations. Our development relies upon the strong approximation and the Beveridge–Nelson representation of linear processes. For illustrative purposes, we apply our results and show the asymptotic validity of the sieve bootstrap for Dickey–Fuller unit root tests for the model driven by a general linear process with independent and identically distributed innovations. We thus provide a theoretical justification on the use of the bootstrap Dickey–Fuller tests for general unit root models, in place of the testing procedures by Said and Dickey and by Phillips.

This paper introduces structural equations that do not satisfy the rank and/or order condition(s) for identification but still are identifiable. These equations are called seemingly unidentified structural equations. The key to the identifiability of these equations is that the right-hand-side endogenous variables undergo structural changes with respect to the exogenous and/or predetermined variables. To estimate the seemingly unidentified structural equations, this paper uses the classical minimum distance (MD) estimator and the principal components instrumental variables (PCIV) estimator. The PCIV estimator is different from conventional IV estimators for structural equations in that nonlinear functions of exogenous and/or predetermined variables are used as instruments. Simulation results comparing the estimator efficiency of the MD and PCIV estimators are reported in this paper. The results indicate that the MD and PCIV estimators are complementary to each other. The estimation methods proposed in this paper are applied to the Japanese export and GDP data to study the effect of export growth rate on that of GDP.

Modern time series econometrics involves a diversity of models. In addition to the more traditional vector autoregressive (VAR) and autoregressive moving average (ARMA) systems, cointegration and unit root models are in widespread use for macroeconomic data, nonlinear and non-Gaussian models are popular for financial data, and long memory models are becoming more common in both macroeconomic and financial applications. Much econometric thought relates to issues of estimation and hypothesis testing, and so, in the absence of a usable finite sample theory (as is the case for the models just mentioned), an enormous amount of effort has been given to developing adequate asymptotics for statistical inference. There is often a lag between the introduction of a new model and the development of an asymptotic theory. In consequence, applied econometricians sometimes have to estimate time series models for which no asymptotic theory is available. For instance, multivariate generalized autoregressive conditional heteroskedasticity (GARCH) models have been in use in empirical research for a while, and practitioners have been using asymptotic normality of estimators in this model even though a theoretical justification is not available.

This paper considers generalized method of moment–type estimators for which a criterion function is minimized that is not the “standard” quadratic distance measure but instead is a general Lp distance measure. It is shown that the resulting estimators are root-n consistent but not in general asymptotically normally distributed, and we derive the limit distribution of these estimators. In addition, we prove that it is not possible to obtain estimators that are more efficient than the “usual” L2-GMM estimators by considering Lp-GMM estimators. We also consider the issue of the choice of the weight matrix for Lp-GMM estimators.

We consider a multiple mismeasured regressor errors-in-variables model where the measurement and equation errors are independent and have moments of every order but otherwise are arbitrarily distributed. We present parsimonious two-step generalized method of moments (GMM) estimators that exploit overidentifying information contained in the high-order moments of residuals obtained by “partialling out” perfectly measured regressors. Using high-order moments requires that the GMM covariance matrices be adjusted to account for the use of estimated residuals instead of true residuals defined by population projections. This adjustment is also needed to determine the optimal GMM estimator. The estimators perform well in Monte Carlo simulations and in some cases minimize mean absolute error by using moments up to seventh order. We also determine the distributions for functions that depend on both a GMM estimate and a statistic not jointly estimated with the GMM estimate.

In Econometric Theory, vol. 18, no. 4, p. 1008, the phrase “Solution, proposed by O. Linton and D. Kristensen” should be changed to “Solution, proposed by O. Linton, D. Kristensen, and G. Dhaene.