Colin McKenzie

Bio: Colin McKenzie is an academic researcher from Keio University. The author has contributed to research in topics: Estimator & Underwriting. The author has an hindex of 14, co-authored 107 publications receiving 907 citations. Previous affiliations of Colin McKenzie include London School of Economics and Political Science & Australian National University.

Variable Addition and Lagrange Multiplier Tests for Linear and Logarithmic Regression Models

Abstract: The purpose of this paper is to examine the properties of various tests of linear and logarithmic (or log-linear) regression models. The test procedures may be categorized as follows: (1) tests that exploit the fact that the two models are intrinsically non-nested; (2) tests based on the Box-Cox data transformation; and (3) diagnostic tests of functional form misspecification against an unspecified alternative. The small-sample properties of several tests are investigated through a Monte Carlo experiment, as is their robustness to non-normality of the errors. Copyright 1988 by MIT Press.

On Efficient Estimation and Correct Inference in Models with Generated Regressors: a General Approach

Abstract: Two approaches have been developed for deriving the properties of efficiency and consistency of standard errors of two step estimators of linear models containing current or lagged unobserved expectations of a single variable. One method is based on the derivatives of the likelihood function and information matrix, while the other uses the true covariance matrix of the disturbance vector when unknown parameters or variables are replaced by corresponding estimates. In this paper, the second approach is extended to cases where the structural equation is nonlinear and the model contains expectations of more than one variable or expectations of future variables. The properties of a frequently used estimator to deal with missing observations problems, a model involving a variance as an explanatory variable, and a recently developed estimator for autoregressive moving average models can be easily derived using the results of the paper. Methods for improving the efficiency of two step estimators are outlined.

When are two step estimators efficient

Abstract: Kruskal's theorem is used to provide simple and elegant alternative derivations of the efficiency of some two step estimators (2SE) for models containing anticipated and unanticipated variables. Several new results are established: 2SE is not efficient for a structural equation with current and lagged values of both anticipated and unanticipated variables; 2SE is always efficient for the parameter associated with the current unanticipated variable, and for the parameter associated with the lagged unanticipated variable if there is no lagged dependent variable in the expectations equation; the inclusion of additional regressors in the structural equation and contemporaneous correlation of the structural and expectations errors can both be analysed in a straightforward manner; the single-equation generalized least squares estimator can be as efficient as the systems maximum likelihood estimator.

Alternative forms and properties of the score test

Abstract: TWO different issues relating to the score (S) test are investigated. Firstly, we study the finite sample properties of a number of asymptotically equivalent forms of the S test. From our simulation results we observe that these forms can behave very differently in finite samples. Secondly, we investigate the power properties of the S test and find that it compares favorably to those of the likelihood ratio (LR) test although the former does not use information about the precise forms of the alternatives.

Econometric Analysis of Cross Section and Panel Data

Abstract: The second edition of this acclaimed graduate text provides a unified treatment of two methods used in contemporary econometric research, cross section and data panel methods. By focusing on assumptions that can be given behavioral content, the book maintains an appropriate level of rigor while emphasizing intuitive thinking. The analysis covers both linear and nonlinear models, including models with dynamics and/or individual heterogeneity. In addition to general estimation frameworks (particular methods of moments and maximum likelihood), specific linear and nonlinear methods are covered in detail, including probit and logit models and their multivariate, Tobit models, models for count data, censored and missing data schemes, causal (or treatment) effects, and duration analysis. Econometric Analysis of Cross Section and Panel Data was the first graduate econometrics text to focus on microeconomic data structures, allowing assumptions to be separated into population and sampling assumptions. This second edition has been substantially updated and revised. Improvements include a broader class of models for missing data problems; more detailed treatment of cluster problems, an important topic for empirical researchers; expanded discussion of "generalized instrumental variables" (GIV) estimation; new coverage (based on the author's own recent research) of inverse probability weighting; a more complete framework for estimating treatment effects with panel data, and a firmly established link between econometric approaches to nonlinear panel data and the "generalized estimating equation" literature popular in statistics and other fields. New attention is given to explaining when particular econometric methods can be applied; the goal is not only to tell readers what does work, but why certain "obvious" procedures do not. The numerous included exercises, both theoretical and computer-based, allow the reader to extend methods covered in the text and discover new insights.

Unequally spaced panel data regressions with ar(1) disturbances

Abstract: This paper deals with the estimation of unequally spaced panel data regression models with AR(1) remainder disturbances. A feasible generalized least squares (GLS) procedure is proposed as a weighted least squares that can handle a wide range of unequally spaced panel data patterns. This procedure is simple to compute and provides natural estimates of the serial correlation and variance components parameters. The paper also provides a locally best invariant test for zero first-order serial correlation against positive or negative serial correlation in case of unequally spaced panel data.

Modeling Ordered Choices: A Primer

Abstract: It is increasingly common for analysts to seek out the opinions of individuals and organizations using attitudinal scales such as degree of satisfaction or importance attached to an issue Examples include levels of obesity, seriousness of a health condition, attitudes towards service levels, opinions on products, voting intentions, and the degree of clarity of contracts Ordered choice models provide a relevant methodology for capturing the sources of influence that explain the choice made amongst a set of ordered alternatives The methods have evolved to a level of sophistication that can allow for heterogeneity in the threshold parameters, in the explanatory variables (through random parameters), and in the decomposition of the residual variance This book brings together contributions in ordered choice modeling from a number of disciplines, synthesizing developments over the last fifty years, and suggests useful extensions to account for the wide range of sources of influence on choice