Bias correcting adjustment coefficients in a cointegrated VAR with known cointegrating vectors

The maximum likelihood estimator of the adjustment coefficient in a cointegrated vector autoregressive model (CVAR) is generally biased. For the case where the cointegrating vector is known in a first-order CVAR with no intercept, we derive a condition for the unbiasedness of the maximum likelihood estimator of the adjustment coefficients, and provide a simple characterization of the bias in case this condition is violated. A feasible bias correction method is shown to virtually eliminate the bias over a large part of the parameter space.     

An MPEC estimator for misclassification models

In this paper, we propose a constrained maximum likelihood estimator for misclassification models, by formulating the estimation as an MPEC (Mathematical Programming with Equilibrium Constraints) problem. Our approach improves the numerical accuracy and avoids the singularity problem. Monte Carlo simulations confirm that the proposed estimator reduces bias and standard deviation of the estimator, especially when the sample is small/medium and/or the dimension of latent variable is large.     

Inflation, unemployment, and the time consistency of the US monetary policy

This paper verifies the performance of the Barro and Gordon (1983) model to explain the US inflation since the early 1950s. We divide the period from 1951:2 to 2010:2 according to each chairman of the Federal Reserve (FED). In addition, we consider aggregated periods, represented by pre-Volcker, Volcker-Greenspan, Greenspan-Bernanke, and whole sample. A genetic algorithm of stochastic search is applied to reduce the sensitivity of the maximum likelihood estimator to the initial parameter values. Surprisingly, our results show that the time consistency problem explains the US inflation during the Greenspan chairmanship at the FED.     

The incidental parameter problem in a non-differentiable panel data model

We consider a panel quantile model with fixed effects. It is shown that the maximum likelihood estimator is numerically equivalent to the least absolute deviations estimator of the differenced model, and as a consequence, there is no incidental parameter problem.     

On Liu estimators for the logit regression model

This paper introduces a shrinkage estimator for the logit model which is a generalization of the estimator proposed by Liu (1993) for the linear regression. This new estimation method is suggested since the mean squared error (MSE) of the commonly used maximum likelihood (ML) method becomes inflated when the explanatory variables of the regression model are highly correlated. Using MSE, the optimal value of the shrinkage parameter is derived and some methods of estimating it are proposed. It is shown by means of Monte Carlo simulations that the estimated MSE and mean absolute error (MAE) are lower for the proposed Liu estimator than those of the ML in the presence of multicollinearity. Finally the benefit of the Lie estimator is shown in an empirical application where different economic factors are used to explain the probability that municipalities have net increase of inhabitants.     

Convenient estimators for the panel probit model: Further results

Bertschek and Lechner (1998) propose several variants of a GMM estimator based on the period specific regression functions for the panel probit model. The analysis is motivated by the complexity of maximum likelihood estimation and the possibly excessive amount of time involved in maximum simulated likelihood estimation. But, for applications of the size considered in their study, full likelihood estimation is actually straightforward, and resort to GMM estimation for convenience is unnecessary. In this note, we reconsider maximum likelihood based estimation of their panel probit model then examine some extensions which can exploit the heterogeneity contained in their panel data set. Empirical results are obtained using the data set employed in the earlier study. Helpful comments and suggestions by Irene Bertschek and Michael Lechner are gratefully acknowledged. This paper has also benefited from comments by two anonymous referees and from seminar participants at the Center for Health Economics at the University of York. Any remaining errors are the responsibility of the author.     

Fractional integration with drift: estimation in small samples

We examine the finite-sample behavior of estimators of the order of integration in a fractionally integrated time-series model. In particular, we compare exact time-domain likelihood estimation to frequency-domain approximate likelihood estimation. We show that over-differencing is of critical importance for time-domain maximum-likelihood estimation in finite samples. Overdifferencing moves the differencing parameter (in the over-differenced model) away from the boundary of the parameter space, while at the same time obviating the need to estimate the drift parameter. The two estimators that we compare are asymptotically equivalent. In small samples, however, the time-domain estimator has smaller mean squared error than the frequency-domain estimator. Although the frequency-domain estimator has larger bias than the time-domain estimator for some regions of the parameter bias, it can also have smaller bias. We use a simulation procedure which exploits the approximate linearity of the bias function to reduce the bias in the time-domain estimator.     

Approximate Whittle analysis of fractional cointegration and the stock market synchronization issue

I consider a bivariate stationary fractional cointegration system and I propose a quasi-maximum likelihood estimator based on the Whittle analysis of the joint spectral density of the regressor and errors. This allows to estimate jointly all parameters of interest of the model. I lead a Monte Carlo experiment to investigate the finite sample properties of this estimator when integration orders are less than 1/2. However, it is not so easy for practitioners to identify whether or not the observed time series are stationary. This issue is investigated by extending the numerical analysis to mean-reverting non-stationary region of the parameter space, although the proposed estimator is not theoretically designed to handle this case. The results display good finite sample properties in both cases, stationary and non-stationary. Thereby, it reveals that making a wrong decision on the stationarity of raw series does not lead to an erroneous conclusion. An application to the stock market synchronization is proposed to illustrate the empirical relevance of this estimator.     

Maximum likelihood estimation for vector autoregressions with multivariate stochastic volatility

This paper analyzes the maximum likelihood estimation for vector autoregressions with stochastic volatility. The stochastic volatility is modeled following Uhlig (1997). The asymptotic distribution of the maximum likelihood estimate is discussed under mild regularity conditions. The maximum likelihood estimate can be obtained via an iterative method. In that case, the maximum likelihood estimate becomes the iteratively reweighted least squares estimate analyzed in Rubin (1983). The iteratively reweighted least squares estimate is computationally much simpler than the Bayesian method offered by Uhlig (1997).     

Consumption and Government Spending Substitutability Revisited: Evidence from Taiwan

Bailey (1971) first documented the idea that there may be a degree of substitutability of the relationship between government spending and private consumption. In this paper, this issue is embedded in a Markov–switching framework where the relationship is subject to shifting between two different regimes. To control small–sample bias, the bootstrap maximum likelihood estimator is used. Evidence from Taiwan indicates that the crowding–in effect dominated the pre–1980 period; the substitutability dominates the post–1980 period. It renders unconvincing the Keynesian plea for expansionary fiscal policy of Taiwan since the 1980s. A Mundell–Fleming approach is proposed to explain this dating.     

A simple and efficient test for the Pareto law

This note presents a simple and locally optimal test statistic for the Pareto law. The test is based on the Lagrange multiplier principle and can be computed easily once the maximum likelihood estimator of the scale parameter of the Pareto density has been obtained. A Monte Carlo exercise shows the good small sample properties of the test under the null of the Pareto law and also its power against some sensible and interesting alternatives. In addition, the proposed test is compared to a goodness of fit test which is powerful against more or less all alternatives. Eventually, a simple application to urban economics is performed.     

Modified Quasi‐Likelihood Ratio Test for Regime Switching

In this paper we propose a modified quasi‐likelihood ratio test of the null hypothesis of one regime against the alternative of two regimes in Markov regime‐switching models. The asymptotic distribution of the proposed test statistic is a simple function of Gaussian random variables, and the inference is no more complicated than in the standard case. Our simulations show that the proposed test has good finite sample size and power that are comparable to the quasi‐likelihood ratio test of Cho and White. We apply our test to stock returns and Japanese policy functions.     

CONSISTENT VARIANCE OF THE LAPLACE‐TYPE ESTIMATORS: APPLICATION TO DSGE MODELS

The Laplace‐type estimator has become popular in applied macroeconomics, in particular for estimation of dynamic stochastic general equilibrium (DSGE) models. It is often obtained as the mean and variance of a parameter's quasi‐posterior distribution, which is defined using a classical estimation objective. We demonstrate that the objective must be properly scaled; otherwise, arbitrarily small confidence intervals can be obtained if calculated directly from the quasi‐posterior distribution. We estimate a standard DSGE model and find that scaling up the objective may be useful in estimation with problematic parameter identification. It this case, however, it is important to adjust the quasi‐posterior variance to obtain valid confidence intervals.     

A THREE‐STEP PROCEDURE FOR ESTIMATING AND TESTING COINTEGRATED ARMAX MODELS*

To deal with a variety of inferential problems on non‐stationary cointegrated time series, this paper proposes a computationally feasible method based on the Whittle likelihood and examines its performance. For the empirical application of our method, the paper investigates three sets of Japanese and US monetary and financial time‐series data. To evaluate the p‐value of the likelihood ratio statistic, we propose an approximation procedure based on the gamma distribution and the accompanying Laguerre expansion for reducing the computational burden. We also provide a numerical procedure for the asymptotic covariance matrix of the Whittle estimator.     

Estimation of Structural Nonlinear Errors‐in‐Varibles Models by Simulated Least‐Squares Method

This article proposes a simulation approach to obtain least‐squares or generalized least‐squares estimators of structural nonlinear errors‐in‐variables models. The proposed estimators are computationally attractive because they do not need numerical integration nor huge numbers of simulations per observable. In addition, the asymptotic covariance matrix of the estimator has a simple decomposition that may be used to guide selection of appropriate simulation sizes. The method is also useful for models with missing data or imperfect surrogate covariates, where application of conventional least‐squares and maximum‐likelihood methods is restricted by numerical multidimensional integrations.     

The asymptotic efficiency of semiparametric estimators for censored linear regression models

This paper presents numerical comparisons of the asymptotic mean square estimation errors of semiparametric generalized least squares (SGLS), quantite, symmetrically censored least squares (SCLS), and tobit maximum likelihood estimators of the slope parameters of censored linear regression models with one explanatory variable. The results indicate that the SCLS estimator is less efficient than the other two semiparametric estimators. The SGLS estimator is more efficient than quantile estimators when the tails of the distribution of the random component of the model are not too thick and the probability of censoring is not too large. The most efficient semiparametric estimators usually have smaller mean square estimation errors than does the tobit estimator when the random component of the model is not normally distributed and the sample size is 500–1,000 or more.     

Efficient estimation of partially linear varying coefficient models

In this paper, we consider the problem of estimating a semiparametric partially linear varying coefficient model. We derive the semiparametric efficiency bound for the asymptotic variance of the finite-dimensional parameter estimator. We also propose an efficient estimator for estimating the finite-dimensional parameter of the model. Simulation results show substantial efficiency gain of our proposed estimator over a conventional estimator as considered in Ahmad et al. (2005).     

SEMINONPARAMETRIC MAXIMUM LIKELIHOOD ESTIMATION OF CONDITIONAL MOMENT RESTRICTION MODELS*

This article studies estimation of a conditional moment restriction model with the seminonparametric maximum likelihood approach proposed by Gallant and Nychka (Econometrica 55 (March 1987), 363–90). Under some sufficient conditions, we show that the estimator of the finite dimensional parameter θ is asymptotically normally distributed and attains the semiparametric efficiency bound and that the estimator of the density function is consistent under L₂ norm. Some results on the convergence rate of the estimated density function are derived. An easy to compute covariance matrix for the asymptotic covariance of the θ estimator is presented.     

Testing of the mean reversion parameter in continuous time models

In this paper we use the approximate bias expressions developed in Yu (2012) and Bao et al. (2013) to improve the testing of the ordinary least squares or quasi-maximum likelihood estimator of the mean reversion parameter in continuous time models. We follow the approach given in Iglesias and Phillips (2005) and Chambers (2013), where if we bias correct the estimated mean reversion parameter, we can improve on the small sample properties of the testing procedure. Simulation results confirm the usefulness of this approach using a t$t$-statistic in this setting in the near unit root situation when the mean reversion parameter is approaching its lower bound. Therefore we always recommend bias correcting when applying a t$t$-statistic in practice in this context.     

Comparison of SML and GMM estimators for the random coefficient logit model using aggregate data

A simulated maximum likelihood (SML) estimator for the random coefficient logit model using aggregate data is found to be more efficient than the widely used generalized method of moments estimator (GMM) of Berry et?al. (Econometrica 63:841?C890, 1995). In particular, the SML estimator is better than the GMM estimator in recovery of heterogeneity parameters which are often of central interest in marketing research. With the GMM estimator, the analyst must determine what moment conditions to use for parameter identification, especially the heterogeneity parameters. With the SML estimator, the moment conditions are automatically determined as the gradients of the log-likelihood function, and these are the most efficient ones if the model is correctly specified. Another limitation of the GMM estimator is that the product market shares must be strictly positive while the SML estimator can handle zero market share observations. Properties of the SML and GMM estimators are demonstrated in simulated data and in data from the US photographic film market.