Optimal critical values of pre-tests when estimating the regression error variance: analytical findings under a general loss structure

This paper re-visits the problem of estimating the regression error variance in a linear multiple regression model after preliminary hypothesis tests for either linear restrictions on the coefficients or homogeneity of variances. There is an extensive literature that discusses these problems, particularly in terms of the sampling properties of the pre-test estimators using various loss functions as the basis for risk analysis. In this paper, a unified framework for analysing the risk properties of these estimators is developed under a general class of loss structures that incorporates virtually all first-order differentiable losses. Particular consideration is given to the choice of critical values for the pre-tests. Analytical results indicate that an α-level substantially higher than those normally used may be appropriate for optimal risk properties under a wide range of loss functions. The paper also generalizes some known analytical results in the pre-test literature and proves other results only previously shown numerically.     

Estimation of regression coefficients after a preliminary test for homoscedasticity

When heteroscedasticity of the variances of disturbances in a regression model is suspected, we perform a preliminary test for homoscedasticity prior to estimation of regression coefficients. According to the result of the pre-test, we use either the ordinary least squares estimator or the two-stage Aitken estimator (2SAE). In this paper, using orthonormal regressors, we derive the mean square error (MSE) of the pre-test estimator and show that the 2SAE is inadmissible when the MSE is used as a criterion. Further, we seek the optimal critical value of the pre-test in the sense of minimizing the average relative risk which is based on the MSE.     

Identifiability of mean‐reverting measurement error with instrumental variable

In the context where one main regressor is measured with error and at least one instrumental variable is available for the correction of measurement error, this paper provides, to the best of our knowledge, a first point‐identification result on the variance of measurement error, the variance of latent variable, and their covariance. We show that the parameters are identified if the regression model is not de facto linear. We illustrate the method in an application to identify mean‐reverting measurement error, a typical issue in reported income where the measurement error of income is negatively correlated with the true income.     

Exact inference for the linear model with groupwise heteroscedastic spherical disturbances

Exact nonparametric inference on a single coefficient in a linear regression model, as considered by Bekker (Working Paper, Department of Economics, University of Groningen, 1997), is elaborated for the case of spherically distributed heteroscedastic disturbances. Instead of approximate inference based on feasible weighted least squares, exact inference is formulated based on partial rotational invariance of the distribution of the vector of disturbances. Thus, classical exact inference based on t-statistics is generalized to exact inference that remains valid in a groupwise heteroscedastic context. The approach is applied to a basic two-sample problem, and to the random- and fixed-effects models for panel data.     

The LIML estimator has finite moments!

The Limited Information Maximum Likelihood estimator of the vector of coefficients of a structural equation in a simultaneous equation model is the vector that defines the linear combination maximizing the effect variance relative to the error variance. If this “eigenvector” solution is normalized by setting a designated coefficient equal to 1, the second-order moment of the estimator may be unbounded. However, the second-order moment is finite if the normalization sets the sample error variance of the linear combination equal to 1.     

Preliminary test estimators of the parameters of simple linear model with measurement error

In this paper, we discuss the properties of preliminary test estimators (PTE) of the parameters of simple linear model with measurement error (ME model) when the slope of the linear model is suspected to be zero. Expressions of the bias, MSE and efficiencies are obtained under conditional as well as unconditional situations with known reliability coefficient. Conditional model results are compared to the standard model without measurement error. We also provide the unconditional model analysis in finite samples. Asymptotic theory under local alternatives is developed when the variance of measurement error or the ratio of the variance of the model error relative to the variance of the measurement error is known. Asymptotic expressions of bias and MSE of the estimators along with their efficiencies are obtained. In every case, it is shown that the measurement error tend to increase the variability of the estimators compared to the estimators without measurement error. Graphs and tables are provided to see these results and to determine optimum level of significance for minimum guaranteed efficiency. Received October 2001 RID="*" ID="*"  A. K. Md. E. Saleh is a Distinguished Research Professor and H. M. Kim is a Ph.D. candidate in the School of Mathematics and Statistics, Carleton University, Ottawa. Acknowledgment. The authors gratefully acknowledge the constructive suggestion of the referees to improve the paper. The research is supported by NSERC grant A3088.     

PRE-TEST ESTIMATION AND TESTING IN ECONOMETRICS: RECENT DEVELOPMENTS

Abstract. This paper surveys a range of important developments in the area of preliminary-test inference in the context of econometric modelling. Both pre-test estimation and pre-test testing are discussed. Special attention is given to recent contributions and results. These include analyses of pre-test strategies under model mis-specification and generalised regression errors; exact sampling distribution results; and pre-testing inequality constraints on the model's parameters. In many cases, practical advice is given to assist applied econometricians in appraising the relative merits of pre-testing. It is shown that there are situations where pre-testing can be advantageous in practice     

Construction of Exact Simultaneous Confidence Bands for a Simple Linear Regression Model

A simultaneous confidence band provides a variety of inferences on the unknown components of a regression model. There are several recent papers using confidence bands for various inferential purposes; see for example, Sun et al. (1999) , Spurrier (1999) , Al‐Saidy et al. (2003) , Liu et al. (2004) , Bhargava & Spurrier (2004) , Piegorsch et al. (2005) and Liu et al. (2007) . Construction of simultaneous confidence bands for a simple linear regression model has a rich history, going back to the work of Working & Hotelling (1929) . The purpose of this article is to consolidate the disparate modern literature on simultaneous confidence bands in linear regression, and to provide expressions for the construction of exact 1 ?α level simultaneous confidence bands for a simple linear regression model of either one‐sided or two‐sided form. We center attention on the three most recognized shapes: hyperbolic, two‐segment, and three‐segment (which is also referred to as a trapezoidal shape and includes a constant‐width band as a special case). Some of these expressions have already appeared in the statistics literature, and some are newly derived in this article. The derivations typically involve a standard bivariate t random vector and its polar coordinate transformation.     

Benchmarking regression algorithms for loss given default modeling

The introduction of the Basel II Accord has had a huge impact on financial institutions, allowing them to build credit risk models for three key risk parameters: PD (probability of default), LGD (loss given default) and EAD (exposure at default). Until recently, credit risk research has focused largely on the estimation and validation of the PD parameter, and much less on LGD modeling. In this first large-scale LGD benchmarking study, various regression techniques for modeling and predicting LGD are investigated. These include one-stage models, such as those built by ordinary least squares regression, beta regression, robust regression, ridge regression, regression splines, neural networks, support vector machines and regression trees, as well as two-stage models which combine multiple techniques. A total of 24 techniques are compared using six real-life loss datasets from major international banks. It is found that much of the variance in LGD remains unexplained, as the average prediction performance of the models in terms of R² ranges from 4% to 43%. Nonetheless, there is a clear trend that non-linear techniques, and in particular support vector machines and neural networks, perform significantly better than more traditional linear techniques. Also, two-stage models built by a combination of linear and non-linear techniques are shown to have a similarly good predictive power, with the added advantage of having a comprehensible linear model component.     

Variance estimation in the change analysis of a linear regression model

We study a “direct test” of Chu and White (1992) proposed for detecting changes in the trend of a linear regression model. The power of this test strongly depends on a suitable estimation of the variance of the error variables involved. We discuss various types of variance estimators and derive their asymptotic properties under the null-hypothesis of “no change” as well as under the alternative of “a change in linear trend”. A small simulation study illustrates the estimators' finite sample behaviour.     

Averaging estimators for autoregressions with a near unit root

This paper uses local-to-unity theory to evaluate the asymptotic mean-squared error (AMSE) and forecast expected squared error from least-squares estimation of an autoregressive model with a root close to unity. We investigate unconstrained estimation, estimation imposing the unit root constraint, pre-test estimation, model selection estimation, and model average estimation. We find that the asymptotic risk depends only on the local-to-unity parameter, facilitating simple graphical comparisons. Our results strongly caution against pre-testing. Strong evidence supports averaging based on Mallows weights. In particular, our Mallows averaging method has uniformly and substantially smaller risk than the conventional unconstrained estimator, and this holds for autoregressive roots far from unity. Our averaging estimator is a new approach to forecast combination.     

A Monte Carlo comparison of traditional and Stein-rule estimators under squared error loss

This paper reports the results of using an orthonormal regression model, a squared error loss measure and Monte Carlo procedures, to compare the risk functions of traditional and Stein-rule pre-test estimators under a variety of conditions. The results of the sampling experiment indicate that the Stein-rule estimators not only dominate the sampling theory estimators used in applied econometric work but that in contrast to conventional wisdom thegains from using members of the Stein-rule family may be quite significant over a large range of the parameter space.     

On new variance approximations for linear models with inequality constraints

In this paper, we examine the estimation of linear models subject to inequality constraints with a special focus on new variance approximations for the estimated parameters. For models with one inequality restriction, the proposed variance formulas are exact. The variance approximations proposed in this paper can be used in regression analysis, Kalman filtering, and balancing national accounts, when inequality constraints are to be incorporated in the estimation procedure.     

Bayesian stochastic search for VAR model restrictions

We propose a Bayesian stochastic search approach to selecting restrictions for vector autoregressive (VAR) models. For this purpose, we develop a Markov chain Monte Carlo (MCMC) algorithm that visits high posterior probability restrictions on the elements of both the VAR regression coefficients and the error variance matrix. Numerical simulations show that stochastic search based on this algorithm can be effective at both selecting a satisfactory model and improving forecasting performance. To illustrate the potential of our approach, we apply our stochastic search to VAR modeling of inflation transmission from producer price index (PPI) components to the consumer price index (CPI).     

Panel data models with spatially correlated error components

In this paper we consider a panel data model with error components that are both spatially and time-wise correlated. The model blends specifications typically considered in the spatial literature with those considered in the error components literature. We introduce generalizations of the generalized moments estimators suggested in Kelejian and Prucha (1999. A generalized moments estimator for the autoregressive parameter in a spatial model. International Economic Review 40, 509–533) for estimating the spatial autoregressive parameter and the variance components of the disturbance process. We then use those estimators to define a feasible generalized least squares procedure for the regression parameters. We give formal large sample results for the proposed estimators. We emphasize that our estimators remain computationally feasible even in large samples.     

Sectoral Shocks and Metropolitan Employment Growth

In this paper, we study the sources of industry employment growth in each of five metropolitan statistical areas (MSAs). The objective is to understand the relative importance of aggregate disturbances versus local sectoral shocks in generating observed employment fluctuations at the MSA level. The empirical evidence presented in this paper derives from structural vector autoregressions (SVARs), estimated for each of the five MSAs. Estimations use monthly employment data covering nine one-digit industrial categories for the period 1951:1–1999:8, as well as two variables that capture the influences of aggregate (i.e., national) shocks on MSAs. We find that within-MSA industry shocks explain considerably more of the forecast error variance in industry employment growth than do aggregate shocks. Sectoral shocks account for between 87 and 94% of the 36-month-ahead forecast error variance. Among individual local sectors, shocks to MSA-specific government, manufacturing, and service sector employment growth are the predominate sources of variability.     

Estimation of variance after a preliminary test of homogeneity and optimal levels of significance for the pre-test

The question of whether to pool two samples in variance estimation is often decided via a preliminary F test. In this paper we show that the optimal pre-test F value is unity for a one- sided alternative, where the objective function is to minimize average relative risk. The outcome is independent of numbers of degrees of freedom in each sample. Optimal significance levels vary somewhat but are close to $\text{1}\text{2}$ for most d.f. and equal to $\text{1}\text{2}$ when numerator and denominator d.f. are equal. The results also apply to regression variance estimation across two data regimes.     

Robust tracking error portfolio selection with worst-case downside risk measures

This paper proposes downside risk measure models in portfolio selection that captures uncertainties both in distribution and in parameters. The worst-case distribution with given information on the mean value and the covariance matrix is used, together with ellipsoidal and polytopic uncertainty sets, to build-up this type of downside risk model. As an application of the models, the tracking error portfolio selection problem is considered. By lifting the vector variables to positive semidefinite matrix variables, we obtain semidefinite programming formulations of the robust tracking portfolio models. Numerical results are presented in tracking SSE50 of the Shanghai Stock Exchange. Compared with the tracking error variance portfolio model and the equally weighted strategy, the proposed models are more stable, have better accumulated wealth and have much better Sharpe ratio in the investment period for the majority of observed instances.     

Air pollution-mortality models: A demonstration of the effects of random measurement error

Errors of measurement have long been recognized as a chronic problem in statistical analysis. Although there is a vast statistical literature of multiple regression models estimating the air pollution-mortality relationship, this problem has been largely ignored. It is well known that pollution measures contain error, but the consequences of this error for regression estimates is not known. We use Lave and Seskin's air pollution model to demonstrate the consequences of random measurement error. We assume a range of 0% to 50% of the variance of the pollution measures is due to error. We find large differences in the estimated effects on mortality of the pollution variables as well as the other explanatory variables once this measurement error is taken into account. These results cast doubt on the usual regression estimates of the mortality effects of air pollution. More generally our results demonstrate the consequences of random measurement error in the explanatory variable of a multiple regression analysis and the misleading conclusions that may result in policy research if this error is ignored.     

A flexible parametric approach for estimating switching regime models and treatment effect parameters

In this paper, we propose a flexible, parametric class of switching regime models allowing for both skewed and fat-tailed outcome and selection errors. Specifically, we model the joint distribution of each outcome error and the selection error via a newly constructed class of multivariate distributions which we call generalized normal mean–variance mixture distributions. We extend Heckman’s two-step estimation procedure for the Gaussian switching regime model to the new class of models. When the distributions of the outcome errors are asymmetric, we show that an additional correction term accounting for skewness in the outcome error distribution (besides the analogue of the well known inverse mill’s ratio) needs to be included in the second step regression. We use the two-step estimators of parameters in the model to construct simple estimators of average treatment effects and establish their asymptotic properties. Simulation results confirm the importance of accounting for skewness in the outcome errors in estimating both model parameters and the average treatment effect and the treatment effect for the treated.