Sample Separation Information Research Articles

Value of Sample Separation Information in a Sequential Probit Model

We illustrate the estimation and identification of multi-step sequential probit models with and without stepwise sample separation information. The likelihood functions are explicitly derived to ease experimentation with such models. We used data on health, activity limitations, demographic traits and work from the Survey of Income and Program Participation (SIPP) and exactly matched them with Social Security administrative records to showcase our theoretical points. Using a Monte Carlo simulation technique, our results suggest that the correlations in errors across equations may arise due to unobserved individual heterogeneity. Using a novel marginal likelihood approach, we also estimated the above sequential probit model without the sample separation information for the purpose of direct comparison. In terms of both in-sample and jackknife-type out-of-sample predictive analysis, the value of modelling the underlying sequential structure of the determination process in generating correct membership probabilities of belonging to a particular group is confirmed. JEL: C31, C34, C 35, I12, I18

An Empirical Model of Behavioral Heterogeneity

In this paper we propose a new latent class/mixture model (LCM) to determine whether firms behave like profit maximizers or just cost minimizers when there is no additional sample separation information. Since some firms might be maximizing profit while others might minimize cost, the LCM with behavioral heterogeneity can be quite useful. Estimation of the LCM amounts to mixing a cost minimization and a profit maximization model. Using the U.S. airlines data we find that after deregulation about 15% of the airlines are found to be consistent with profit maximizing behavior.

Estimation of the consumption CAPM with imperfect sample separation information

AbstractWe propose a consumption‐based capital asset pricing model consumption (CAPM), in which the pricing kernel is calculated as the average of individuals' intertemporal marginal rates of substitution weighted by the probabilities of holding the asset in question. These probabilities are conditional on available imperfect sample separation information and are estimated simultaneously with the parameters of Euler equations. Using data from the US Consumer Expenditure Survey, we find that the consumption CAPM with probability‐weighted agents yields a more precise estimate of the agent's risk aversion compared with the model, in which the available imperfect information on asset‐holding status is erroneously regarded as a perfect sample separation indicator. Copyright © 2007 John Wiley & Sons, Ltd.

Estimating a mixture of stochastic frontier regression models via the em algorithm: A multiproduct cost function application

Researchers have become increasingly interested in estimating mixtures of stochastic frontiers. Mester (1993), Caudill (1993), and Polachek and Yoon (1987), for example, estimate stochastic frontier models for different regimes, assuming sample separation information is given. Building on earlier work by Lee and Porter (1984), Douglas, Conway, and Ferrier (1995) estimate a stochastic frontier switching regression model in the presence of noisy sample separation information. The purpose of this paper is to extend earlier work by estimating a mixture of stochastic frontiers assuming no sample separation information. This case is more likely to occur in practice than even noisy sample separation information. In order to estimate a mixture of stochastic frontiers with no sample separation information, an EM algorithm to obtain maximum likelihood estimates is developed. The algorithm is used to estimate a mixture of stochastic (cost) frontiers using data on U.S. savings and loans for the years 1986, 1987, and 1988. Statistical evidence is found supporting the existence of a mixture of stochastic frontiers.

Individual heterogeneity and censoring in panel data estimates of tobacco expenditure

AbstractA panel of households is used to test the rational addiction model of Becker et al. (1994). These data raise problems of measurement errors, censoring, and unobservable heterogeneity. We use sample separation information to exclude those households who never purchase tobacco. To deal with the remaining zeros we compare specifications based on infrequency of purchase and on censoring. GMM and system‐GMM are used to deal with errors‐in‐variables and unobservable heterogeneity. Within‐groups two‐step, within‐groups three‐step GMM and Minimum Distance methods are used to allow for censoring. There is evidence that the rational addiction specification is sensitive to unobservable heterogeneity and censoring. Copyright © 2002 John Wiley & Sons, Ltd.

On modelling convergence clubs

A switching regression approach with imperfect sample separation information is used to determine convergence clubs. Regime classification allows one to determine which countries belong to the related convergence clubs. Initial per capita GDP does not provide a perfect sample separation information, but the regimes classification obtained from an endogenous switching model provides unusual results.

On the use of disequilibrium models in applied microeconomic research and the value of sample separation information

The use of disequilibrium models in applied microeconomic research is evaluated. A disequilibrium or switching regime model is used to explainsales levels of individual retail stores. It is investigated whether substantial differences are found if an equilibrium approach is followed instead. Disequilibrium models are known to suffer from the fact that sample separation is unknown. Usually this information is not available. Our sample contains explicit information with respect to the regime to which an observation belongs. Therefore, the value of sample separation information in estimating the disequilibrium model is investigated. Finally, Monte Carlo experiments are conducted to get more insight into these matters.

A smooth likelihood simulator for dynamic disequilibrium models

This article considers the simulation of likelihood functions for dynamic disequilibrium models without sample separation information. A recursive simulation algorithm is proposed. The recursive simulation algorithm is computationally tractable for a class of dynamic disequilibrium models and serially correlated disturbance models. The simulated likelihood functions are smooth in parameters. Monte Carlo studies are provided to demonstrate the computational efficiency of this approach and investigate finite sample properties of simulated likelihood estimators and likelihood ratio test statistics. Regime classification rules based on simulated likelihood are introduced. Finite sample results of the simulated likelihood approach are encouraging.

A smooth likelihood simulator for dynamic disequilibrium models

This article considers the simulation of likelihood functions for dynamic disequilibrium models without sample separation information. A recursive simulation algorithm is proposed. The recursive simulation algorithm is computationally tractable for a class of dynamic disequilibrium models and serially correlated disturbance models. The simulated likelihood functions are smooth in parameters. Monte Carlo studies are provided to demonstrate the computational efficiency of this approach and investigate finite sample properties of simulated likelihood estimators and likelihood ratio test statistics. Regime classification rules based on simulated likelihood are introduced. Finite sample results of the simulated likelihood approach are encouraging.

A Switching Frontier Model for Imperfect Sample Separation Information: With an Application to Constrained Labor Supply

This paper combines frontier functions and switching regressions. This allows economic agents to operate under different efficiency 'regimes,' thus relaxing the assumption that all observations are drawn from the same distribution of inefficiency. The 'switch' is based on sample separation information that is treated first as perfect, then as imperfect (or noisy). Available sample separation information suggests an observation's regime, however the information may not be accurate. By comparing results across alternative specifications of sample separation information as perfect and noisy, this approach provides evidence on the quality of the sample separation information. The technique's usefulness is demonstrated via an application to constrained labor supply. Copyright 1995 by Economics Department of the University of Pennsylvania and the Osaka University Institute of Social and Economic Research Association.

A double‐hurdle model of cigarette consumption

AbstractThis paper shows the importance of the double‐hurdle approach for modelling individuals' cigarette consumption, using data from the UK General Household Survey, and argues that participation and consumption should be treated as separate individual choices. The likelihood function for the full double‐hurdle is derived, and it is shown how restrictions on the stochastic specification of the model and auxillary information, which identifies ex‐smokers, allow it to be decomposed. The empirical results highlight the value of the sample separation information and the need to model starting and quitting as separate decisions.

The value of imperfect sample separation information in mixtures of normal distributions

Abstract This paper conducts a Monte Carlo study in order to evaluate the value of information in a normal mixture model when an imperfect sample separation indicator is available. For a variety of cases, computations are done on the ratios of asymptotic variances of the parameter estimators when sample separation is known versus when only the indicator is available. This study is patterned closely on a paper by Schmidt (1981), and obtains similar results when parameters appearing in Schmidt's model are allowed to vary. The new results here focus on the importance of the probabilities that the regime indicator is correct, for each of the two possible regimes.

Further Results on the Value of Sample Separation Information

Further Results on the Value of Sample Separation Information