Cheng Hsiao - Academia.edu (original) (raw)

Papers by Cheng Hsiao

The Oxford Handbook of Panel Data, 2015

This Chapter reviews the recent literature on dynamic panel data models with a short time span an... more This Chapter reviews the recent literature on dynamic panel data models with a short time span and a large cross-section. Throughout the discussion we consider linear models with additional endogenous covariates. First we give a broad overview of available inference methods placing emphasis on GMM. We next discuss in more detail the assumption of mean stationarity underlying the system GMM estimator. We discuss causes of deviations from mean stationarity, their consequences and tests for mean stationarity. Prepared for the Oxford Handbook on Panel Data (editor: Badi H. Baltagi), Oxford University Press. We would like to thank Artūras Juodis, two anonymous referees and the Editor for helpful comments and suggestions.

This paper provides a selective survey of the panel macroeconometric techniques that focus on con... more This paper provides a selective survey of the panel macroeconometric techniques that focus on controlling the impact of “unobserved heterogeneity” across individuals and over time to obtain valid inference for “structures” that are common across individuals and over time. We consider issues of (i) estimating vector autoregressive models; (ii) testing of unit root or cointegration; (iii) statistical inference for dynamic simultaneous equations models; (iv) policy evaluation; and (v) aggregation and prediction.

SSRN Electronic Journal, 2015

We examine the asymptotic properties of IV, GMM or MLE to estimate dynamic panel data models when... more We examine the asymptotic properties of IV, GMM or MLE to estimate dynamic panel data models when either N or T or both are large. We show that the Anderson and Hsiao (1981, 1982) simple instrumental variable estimator (IV) or maximizing the likelihood function with initial value distribution properly treated (quasi-maximum likelihood estimator) is asymptotically unbiased when either N or T or both tend to infinity. On the other hand, the QMLE mistreating the initial value as fixed is asymptotically unbiased only if N is fixed and T is large. If both N and T are large and N T → c (c = 0, c < ∞) as T → ∞, it is asymptotically biased of order N T. We also explore the source of the bias of the Arellano and Bond (1991) type GMM estimator. We show that it is asymptotically biased of order T N if T N → c (c = 0, c < ∞) as N → ∞ even if we restrict the number of instruments used. Monte Carlo studies show that whether an estimator is asymptotically biased or not has important implications on the actual size of the conventional t-test.

SSRN Electronic Journal, 2014

We study the identi…cation and estimation of panel dynamic simultaneous equations models. We show... more We study the identi…cation and estimation of panel dynamic simultaneous equations models. We show that the presence of time-persistent individual-speci…c e¤ects does not lead to changes in the identi…cation conditions of traditional Cowles Commission dynamic simultaneous equations models. However, the limiting properties of the estimators depend on the way the cross-section dimension, N , or the time series dimension, T , goes to in…nity. We propose three limited information estimator: panel simple instrumental variables (PIV), panel generalized two stage least squares (PG2SLS), and panel limited information maximum likelihood estimation (PLIML). We show that they are all asymptotically unbiased independent of the way of how N or T tends to in…nity. Monte Carlo studies are conducted to compare the performance of the PLIML, PIV, PG2SLS, the Arellano-Bond type generalized method of moments and the Akashi-Kunitomo least variance ratio estimator and to demonstrate the sensitivity of statistical inference to the asymptotic bias of an estimator.

SSRN Electronic Journal, 2014

We propose a panel data approach to disentangle the impact of "one treatment" from the "other tre... more We propose a panel data approach to disentangle the impact of "one treatment" from the "other treatment" when the observed outcomes are subject to both treatments. We use the Great Hanshin-Awaji earthquake that took place on January 17, 1995 to illustrate our methodology. We find that there were no persistent earthquake effects. The observed persistent effects are due to structural change in Hyogo prefecture.

SSRN Electronic Journal, 2014

This paper provides a selective survey of the panel macroeconometric techniques that focus on con... more This paper provides a selective survey of the panel macroeconometric techniques that focus on controlling the impact of "unobserved heterogeneity" across individuals and over time to obtain valid inference for "structures" that are common across individuals and over time. We consider issues of (i) estimating vector autoregressive models; (ii) testing of unit root or cointegration; (iii) statistical inference for dynamic simultaneous equations models; (iv) policy evaluation; and (v) aggregation and prediction.

The Singapore Economic Review, 2009

This paper reviews the literature on the impact of marijuana decriminalization policy on marijuan... more This paper reviews the literature on the impact of marijuana decriminalization policy on marijuana smoking prevalence. Due to mixed findings in the existing studies, we attempt to find a common basis to explain the different results across papers. The main purpose is to provide a coherent background as to what outcomes may be expected from certain type of data, econometric models, and explanatory variables. If possible, we also try to provide the explanation as to why certain results are found.

International Economic Review, 1978

This paper is concerned with the estirrtion of the parameters in a dynamic simultaneous equation ... more This paper is concerned with the estirrtion of the parameters in a dynamic simultaneous equation model with stationary disturbances under the assimption that the variables are subj ect to random measurement errors. The conditions under which the parameters are identified are stated. An asymptotically efficient frequency-domain class of instrumental variables estimators is suggested. The procedure consists of two basic steps. The first step transforms the model in such a way that the observed exogenous variables are asymptotically orthogonal to the residual terms. The second step involves an iterative procedure like that of Robinson [13].

AStA Advances in Statistical Analysis, 2011

We would like to thank the referees and the editor for helpful comments and suggestions in an ear... more We would like to thank the referees and the editor for helpful comments and suggestions in an early version. Many thanks to Annick Vignes for providing us the dataset. We also thank

Semiparametric estimation methods are used for models which are partly parametric and partly nonp... more Semiparametric estimation methods are used for models which are partly parametric and partly nonparametric; typically the parametric part is an underlying regression function which is assumed to be linear in the observable explanatory variables, while the nonparametric component involves the distribution of the model's 'error terms'. Semiparametric methods are particularly useful for limited dependent variable models (for example, the binary response or censored regression models), since fully parametric specifications for those models yield inconsistent estimators if the parametric distribution of the errors is misspecified.

The New Palgrave Dictionary of Economics

The New Palgrave Dictionary of Economics

Restricted until 6 April 2008. This dissertation estimates the average treatment effects of a mot... more Restricted until 6 April 2008. This dissertation estimates the average treatment effects of a mother's full-time employment on children's body mass index (BMI) and likelihood of becoming overweight. The matched mother-child data from the 2000 wave of the U.S. National Longitudinal Survey of Youth 79 (NLSY79) are used. In the first part of the dissertation, the econometric methods correcting the bias from "selection on observables," including control function and matching based on propensity score, are applied to perform the estimation. In the second part, the econometric methods correcting the bias from "selection on unobservables," including maximum likelihood and semiparametric approaches, are used to conduct the estimation. It is concluded that, on average, the group of children with full-time working mothers have significantly higher BMI and a greater likelihood of becoming overweight.

Monetary and and Economic Studies, 2002

We use annual Japanese prefecture data on income, population, demand deposits, and saving deposit... more We use annual Japanese prefecture data on income, population, demand deposits, and saving deposits from 1992 to 1997 to investigate the issue of whether there exists a stable money demand function under the low interest rate policy. The evidence appears to support the contention that there does exist a stable money demand function with long-run income elasticity greater than one for M2 and less than one for Ml. Furthermore, we find that Japan's money demand is sensitive to interest rate changes. However, there is no evidence of the presence of a liquidity trap.

This paper provides a review of linear panel data models with slope heterogeneity, introduces var... more This paper provides a review of linear panel data models with slope heterogeneity, introduces various types of random coefficients models and suggests a common framework for dealing with them. It considers the fundamental issues of statistical inference of a random coefficients formulation using both the sampling and Bayesian approaches. The paper also provides a review of heterogeneity, simultaneous equation random coefficient models, and the more recent developments in the area of cross-sectional dependence in panel data models.

SSRN Electronic Journal

This Discussion Paper is issued under the auspices of the Centre's research programme in INSTITUT... more This Discussion Paper is issued under the auspices of the Centre's research programme in INSTITUTIONS AND ECONOMIC PERFORMANCE. Any opinions expressed here are those of the author(s) and not those of the Centre for Economic Policy Research. Research disseminated by CEPR may include views on policy, but the Centre itself takes no institutional policy positions. The Centre for Economic Policy Research was established in 1983 as a private educational charity, to promote independent analysis and public discussion of open economies and the relations among them. It is pluralist and non-partisan, bringing economic research to bear on the analysis of medium-and long-run policy questions. Institutional (core) finance for the Centre has been provided through major grants from the Economic and Social Research Council, under which an ESRC Resource Centre operates within CEPR; the Esmée Fairbairn Charitable Trust; and the Bank of England. These organizations do not give prior review to the Centre's publications, nor do they necessarily endorse the views expressed therein. These Discussion Papers often represent preliminary or incomplete work, circulated to encourage discussion and comment. Citation and use of such a paper should take account of its provisional character.

SSRN Electronic Journal

We use a quasi-likelihood function approach to clarify the role of initial values and the relativ... more We use a quasi-likelihood function approach to clarify the role of initial values and the relative size of the cross-section dimension N and the time series dimension T in the asymptotic distribution of dynamic panel data models with the presence of individual-speci…c e¤ects. We show that the quasi-maximum likelihood estimator (QMLE) treating initial values as …xed constants is asymptotically biased of order q N T 3 as T goes to in…nity for a time series models and asymptotically biased of order q N T for a model that also contains other covariates that are correlated with the individual-speci…c e¤ects. Using Mundlak-Chamberlain approach to condition the e¤ects on the covariates can reduce the asymptotic bias to the order of q N T 3 , provided the data generating processes for the covariates are homogeneous across cross-sectional units. On the other hand, the QMLE combining the Mundlak-Chamberlain approach with the proper treatment of initial value distribution is asymptotically unbiased if N goes to in…nity whether T is …xed or goes to in…nity. Monte Carlo studies are conducted to demonstrate the importance of properly treating initial values in getting valid statistical inference. The results also suggest that when using the conditional approach to get around the issue of incidental parameters, in …nite sample it is perhaps better to follow Mundlak's (1978) suggestion to simply condition the individual This paper was stimulated by the private communication with Jushan Bai. We would like to thank him for pointing out the terms that cancel the impact of individual-speci…c e¤ects and the impact of correlation between the errors of the equation and the lagged dependent variables. We would also like to thank the editor Oliver Linton, an associate editor, two anonymous referees and Elie Tamer for helpful comments. Partial research support by China NSF #71103004 and #71631004 to the …rst author is also gratefully acknowledged.

Journal of Applied Econometrics

We consider panel parametric, semiparametric and nonparametric methods of constructing counterfac... more We consider panel parametric, semiparametric and nonparametric methods of constructing counterfactuals. We show through extensive simulations that no method is able to dominate other methods in all circumstances, since the true data‐generating process is typically unknown. We therefore also suggest a model‐averaging method as a robust method to generate counterfactuals. As an illustration of the sensitivity of counterfactual construction, we reexamine the impact of California's Tobacco Control Program on per capita cigarette consumption and election day registration (EDR) laws on voters' turnout by different methods.

Econometrics

This paper considers methods of estimating a static correlated random coefficient model with pane... more This paper considers methods of estimating a static correlated random coefficient model with panel data. We mainly focus on comparing two approaches of estimating unconditional mean of the coefficients for the correlated random coefficients models, the group mean estimator and the generalized least squares estimator. For the group mean estimator, we show that it achieves Chamberlain (1992) semi-parametric efficiency bound asymptotically. For the generalized least squares estimator, we show that when T is large, a generalized least squares estimator that ignores the correlation between the individual coefficients and regressors is asymptotically equivalent to the group mean estimator. In addition, we give conditions where the standard within estimator of the mean of the coefficients is consistent. Moreover, with additional assumptions on the known correlation pattern, we derive the asymptotic properties of panel least squares estimators. Simulations are used to examine the finite sam...

The Oxford Handbook of Panel Data, 2015

This Chapter reviews the recent literature on dynamic panel data models with a short time span an... more This Chapter reviews the recent literature on dynamic panel data models with a short time span and a large cross-section. Throughout the discussion we consider linear models with additional endogenous covariates. First we give a broad overview of available inference methods placing emphasis on GMM. We next discuss in more detail the assumption of mean stationarity underlying the system GMM estimator. We discuss causes of deviations from mean stationarity, their consequences and tests for mean stationarity. Prepared for the Oxford Handbook on Panel Data (editor: Badi H. Baltagi), Oxford University Press. We would like to thank Artūras Juodis, two anonymous referees and the Editor for helpful comments and suggestions.

This paper provides a selective survey of the panel macroeconometric techniques that focus on con... more This paper provides a selective survey of the panel macroeconometric techniques that focus on controlling the impact of “unobserved heterogeneity” across individuals and over time to obtain valid inference for “structures” that are common across individuals and over time. We consider issues of (i) estimating vector autoregressive models; (ii) testing of unit root or cointegration; (iii) statistical inference for dynamic simultaneous equations models; (iv) policy evaluation; and (v) aggregation and prediction.

SSRN Electronic Journal, 2015

We examine the asymptotic properties of IV, GMM or MLE to estimate dynamic panel data models when... more We examine the asymptotic properties of IV, GMM or MLE to estimate dynamic panel data models when either N or T or both are large. We show that the Anderson and Hsiao (1981, 1982) simple instrumental variable estimator (IV) or maximizing the likelihood function with initial value distribution properly treated (quasi-maximum likelihood estimator) is asymptotically unbiased when either N or T or both tend to infinity. On the other hand, the QMLE mistreating the initial value as fixed is asymptotically unbiased only if N is fixed and T is large. If both N and T are large and N T → c (c = 0, c < ∞) as T → ∞, it is asymptotically biased of order N T. We also explore the source of the bias of the Arellano and Bond (1991) type GMM estimator. We show that it is asymptotically biased of order T N if T N → c (c = 0, c < ∞) as N → ∞ even if we restrict the number of instruments used. Monte Carlo studies show that whether an estimator is asymptotically biased or not has important implications on the actual size of the conventional t-test.

SSRN Electronic Journal, 2014

We study the identi…cation and estimation of panel dynamic simultaneous equations models. We show... more We study the identi…cation and estimation of panel dynamic simultaneous equations models. We show that the presence of time-persistent individual-speci…c e¤ects does not lead to changes in the identi…cation conditions of traditional Cowles Commission dynamic simultaneous equations models. However, the limiting properties of the estimators depend on the way the cross-section dimension, N , or the time series dimension, T , goes to in…nity. We propose three limited information estimator: panel simple instrumental variables (PIV), panel generalized two stage least squares (PG2SLS), and panel limited information maximum likelihood estimation (PLIML). We show that they are all asymptotically unbiased independent of the way of how N or T tends to in…nity. Monte Carlo studies are conducted to compare the performance of the PLIML, PIV, PG2SLS, the Arellano-Bond type generalized method of moments and the Akashi-Kunitomo least variance ratio estimator and to demonstrate the sensitivity of statistical inference to the asymptotic bias of an estimator.

SSRN Electronic Journal, 2014

We propose a panel data approach to disentangle the impact of "one treatment" from the "other tre... more We propose a panel data approach to disentangle the impact of "one treatment" from the "other treatment" when the observed outcomes are subject to both treatments. We use the Great Hanshin-Awaji earthquake that took place on January 17, 1995 to illustrate our methodology. We find that there were no persistent earthquake effects. The observed persistent effects are due to structural change in Hyogo prefecture.

SSRN Electronic Journal, 2014

This paper provides a selective survey of the panel macroeconometric techniques that focus on con... more This paper provides a selective survey of the panel macroeconometric techniques that focus on controlling the impact of "unobserved heterogeneity" across individuals and over time to obtain valid inference for "structures" that are common across individuals and over time. We consider issues of (i) estimating vector autoregressive models; (ii) testing of unit root or cointegration; (iii) statistical inference for dynamic simultaneous equations models; (iv) policy evaluation; and (v) aggregation and prediction.

The Singapore Economic Review, 2009

This paper reviews the literature on the impact of marijuana decriminalization policy on marijuan... more This paper reviews the literature on the impact of marijuana decriminalization policy on marijuana smoking prevalence. Due to mixed findings in the existing studies, we attempt to find a common basis to explain the different results across papers. The main purpose is to provide a coherent background as to what outcomes may be expected from certain type of data, econometric models, and explanatory variables. If possible, we also try to provide the explanation as to why certain results are found.

International Economic Review, 1978

This paper is concerned with the estirrtion of the parameters in a dynamic simultaneous equation ... more This paper is concerned with the estirrtion of the parameters in a dynamic simultaneous equation model with stationary disturbances under the assimption that the variables are subj ect to random measurement errors. The conditions under which the parameters are identified are stated. An asymptotically efficient frequency-domain class of instrumental variables estimators is suggested. The procedure consists of two basic steps. The first step transforms the model in such a way that the observed exogenous variables are asymptotically orthogonal to the residual terms. The second step involves an iterative procedure like that of Robinson [13].

AStA Advances in Statistical Analysis, 2011

We would like to thank the referees and the editor for helpful comments and suggestions in an ear... more We would like to thank the referees and the editor for helpful comments and suggestions in an early version. Many thanks to Annick Vignes for providing us the dataset. We also thank

Semiparametric estimation methods are used for models which are partly parametric and partly nonp... more Semiparametric estimation methods are used for models which are partly parametric and partly nonparametric; typically the parametric part is an underlying regression function which is assumed to be linear in the observable explanatory variables, while the nonparametric component involves the distribution of the model's 'error terms'. Semiparametric methods are particularly useful for limited dependent variable models (for example, the binary response or censored regression models), since fully parametric specifications for those models yield inconsistent estimators if the parametric distribution of the errors is misspecified.

The New Palgrave Dictionary of Economics

The New Palgrave Dictionary of Economics

Restricted until 6 April 2008. This dissertation estimates the average treatment effects of a mot... more Restricted until 6 April 2008. This dissertation estimates the average treatment effects of a mother's full-time employment on children's body mass index (BMI) and likelihood of becoming overweight. The matched mother-child data from the 2000 wave of the U.S. National Longitudinal Survey of Youth 79 (NLSY79) are used. In the first part of the dissertation, the econometric methods correcting the bias from "selection on observables," including control function and matching based on propensity score, are applied to perform the estimation. In the second part, the econometric methods correcting the bias from "selection on unobservables," including maximum likelihood and semiparametric approaches, are used to conduct the estimation. It is concluded that, on average, the group of children with full-time working mothers have significantly higher BMI and a greater likelihood of becoming overweight.

Monetary and and Economic Studies, 2002

We use annual Japanese prefecture data on income, population, demand deposits, and saving deposit... more We use annual Japanese prefecture data on income, population, demand deposits, and saving deposits from 1992 to 1997 to investigate the issue of whether there exists a stable money demand function under the low interest rate policy. The evidence appears to support the contention that there does exist a stable money demand function with long-run income elasticity greater than one for M2 and less than one for Ml. Furthermore, we find that Japan's money demand is sensitive to interest rate changes. However, there is no evidence of the presence of a liquidity trap.

This paper provides a review of linear panel data models with slope heterogeneity, introduces var... more This paper provides a review of linear panel data models with slope heterogeneity, introduces various types of random coefficients models and suggests a common framework for dealing with them. It considers the fundamental issues of statistical inference of a random coefficients formulation using both the sampling and Bayesian approaches. The paper also provides a review of heterogeneity, simultaneous equation random coefficient models, and the more recent developments in the area of cross-sectional dependence in panel data models.

SSRN Electronic Journal

This Discussion Paper is issued under the auspices of the Centre's research programme in INSTITUT... more This Discussion Paper is issued under the auspices of the Centre's research programme in INSTITUTIONS AND ECONOMIC PERFORMANCE. Any opinions expressed here are those of the author(s) and not those of the Centre for Economic Policy Research. Research disseminated by CEPR may include views on policy, but the Centre itself takes no institutional policy positions. The Centre for Economic Policy Research was established in 1983 as a private educational charity, to promote independent analysis and public discussion of open economies and the relations among them. It is pluralist and non-partisan, bringing economic research to bear on the analysis of medium-and long-run policy questions. Institutional (core) finance for the Centre has been provided through major grants from the Economic and Social Research Council, under which an ESRC Resource Centre operates within CEPR; the Esmée Fairbairn Charitable Trust; and the Bank of England. These organizations do not give prior review to the Centre's publications, nor do they necessarily endorse the views expressed therein. These Discussion Papers often represent preliminary or incomplete work, circulated to encourage discussion and comment. Citation and use of such a paper should take account of its provisional character.

SSRN Electronic Journal

We use a quasi-likelihood function approach to clarify the role of initial values and the relativ... more We use a quasi-likelihood function approach to clarify the role of initial values and the relative size of the cross-section dimension N and the time series dimension T in the asymptotic distribution of dynamic panel data models with the presence of individual-speci…c e¤ects. We show that the quasi-maximum likelihood estimator (QMLE) treating initial values as …xed constants is asymptotically biased of order q N T 3 as T goes to in…nity for a time series models and asymptotically biased of order q N T for a model that also contains other covariates that are correlated with the individual-speci…c e¤ects. Using Mundlak-Chamberlain approach to condition the e¤ects on the covariates can reduce the asymptotic bias to the order of q N T 3 , provided the data generating processes for the covariates are homogeneous across cross-sectional units. On the other hand, the QMLE combining the Mundlak-Chamberlain approach with the proper treatment of initial value distribution is asymptotically unbiased if N goes to in…nity whether T is …xed or goes to in…nity. Monte Carlo studies are conducted to demonstrate the importance of properly treating initial values in getting valid statistical inference. The results also suggest that when using the conditional approach to get around the issue of incidental parameters, in …nite sample it is perhaps better to follow Mundlak's (1978) suggestion to simply condition the individual This paper was stimulated by the private communication with Jushan Bai. We would like to thank him for pointing out the terms that cancel the impact of individual-speci…c e¤ects and the impact of correlation between the errors of the equation and the lagged dependent variables. We would also like to thank the editor Oliver Linton, an associate editor, two anonymous referees and Elie Tamer for helpful comments. Partial research support by China NSF #71103004 and #71631004 to the …rst author is also gratefully acknowledged.

Journal of Applied Econometrics

We consider panel parametric, semiparametric and nonparametric methods of constructing counterfac... more We consider panel parametric, semiparametric and nonparametric methods of constructing counterfactuals. We show through extensive simulations that no method is able to dominate other methods in all circumstances, since the true data‐generating process is typically unknown. We therefore also suggest a model‐averaging method as a robust method to generate counterfactuals. As an illustration of the sensitivity of counterfactual construction, we reexamine the impact of California's Tobacco Control Program on per capita cigarette consumption and election day registration (EDR) laws on voters' turnout by different methods.

Econometrics

This paper considers methods of estimating a static correlated random coefficient model with pane... more This paper considers methods of estimating a static correlated random coefficient model with panel data. We mainly focus on comparing two approaches of estimating unconditional mean of the coefficients for the correlated random coefficients models, the group mean estimator and the generalized least squares estimator. For the group mean estimator, we show that it achieves Chamberlain (1992) semi-parametric efficiency bound asymptotically. For the generalized least squares estimator, we show that when T is large, a generalized least squares estimator that ignores the correlation between the individual coefficients and regressors is asymptotically equivalent to the group mean estimator. In addition, we give conditions where the standard within estimator of the mean of the coefficients is consistent. Moreover, with additional assumptions on the known correlation pattern, we derive the asymptotic properties of panel least squares estimators. Simulations are used to examine the finite sam...