Handbook of Empirical Economics and Finance _1 pptx

Giri Computational Methods in Statistics and Econometrics, Hisashi Tanizaki Applied Sequential Methodologies: Real-World Examples with Data Analysis, edited by Nitis Mukhopadhyay, Sujay

Empirical Economics and Finance

The EM Algorithm and Related Statistical Models, edited by Michiko Watanabe and

Kazunori Yamaguchi

Multivariate Statistical Analysis, Second Edition, Revised and Expanded, Narayan C Giri

Computational Methods in Statistics and Econometrics, Hisashi Tanizaki

Applied Sequential Methodologies: Real-World Examples with Data Analysis, edited by

Nitis Mukhopadhyay, Sujay Datta, and Saibal Chattopadhyay

Handbook of Beta Distribution and Its Applications, edited by Arjun K Gupta and

Saralees Nadarajah

Item Response Theory: Parameter Estimation Techniques, Second Edition, edited by Frank B Baker and Seock-Ho Kim

Statistical Methods in Computer Security, edited by William W S Chen

Elementary Statistical Quality Control, Second Edition, John T Burr

Data Analysis of Asymmetric Structures, Takayuki Saito and Hiroshi Yadohisa

Mathematical Statistics with Applications, Asha Seth Kapadia, Wenyaw Chan, and Lemuel Moyé Advances on Models, Characterizations and Applications, N Balakrishnan, I G Bairamov, and

O L Gebizlioglu

Survey Sampling: Theory and Methods, Second Edition, Arijit Chaudhuri and Horst Stenger

Statistical Design of Experiments with Engineering Applications, Kamel Rekab and

Muzaffar Shaikh

Quality by Experimental Design, Third Edition, Thomas B Barker

Handbook of Parallel Computing and Statistics, Erricos John Kontoghiorghes

Statistical Inference Based on Divergence Measures, Leandro Pardo

A Kalman Filter Primer, Randy Eubank

Introductory Statistical Inference, Nitis Mukhopadhyay

Handbook of Statistical Distributions with Applications, K Krishnamoorthy

A Course on Queueing Models, Joti Lal Jain, Sri Gopal Mohanty, and Walter Böhm

Univariate and Multivariate General Linear Models: Theory and Applications with SAS,

Second Edition, Kevin Kim and Neil Timm

Randomization Tests, Fourth Edition, Eugene S Edgington and Patrick Onghena

Design and Analysis of Experiments: Classical and Regression Approaches with SAS,

Leonard C Onyiah

Analytical Methods for Risk Management: A Systems Engineering Perspective,

Paul R Garvey

Confidence Intervals in Generalized Regression Models, Esa Uusipaikka

Introduction to Spatial Econometrics, James LeSage and R Kelley Pace

Acceptance Sampling in Quality Control, Edward G Schilling and Dean V Neubauer

Applied Statistical Inference with MINITAB®, Sally A Lesik

Nonparametric Statistical Inference, Fifth Edition, Jean Dickinson Gibbons and Subhabrata Chakraborti Bayesian Model Selection and Statistical Modeling, Tomohiro Ando

Handbook of Empirical Economics and Finance, Aman Ullah and David E A Giles

No claim to original U.S Government works

Printed in the United States of America on acid-free paper

10 9 8 7 6 5 4 3 2 1

International Standard Book Number-13: 978-1-4200-7036-1 (Ebook-PDF)

This book contains information obtained from authentic and highly regarded sources Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S Copyright Law, no part of this book may be reprinted, reproduced, ted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers.

transmit-For permission to photocopy or use material electronically from this work, please access www.copyright com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400 CCC is a not-for-profit organization that provides licenses and registration for a variety of users For organizations that have been granted a photocopy license by the CCC,

a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used

only for identification and explanation without intent to infringe.

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com

Preface ix

About the Editors xv

List of Contributors xvii

1 Robust Inference with Clustered Data 1

A Colin Cameron and Douglas L Miller 2 Efficient Inference with Poor Instruments: A General Framework 29

Bertille Antoine and Eric Renault 3 An Information Theoretic Estimator for the Mixed Discrete Choice Model .71

Amos Golan and William H Greene 4 Recent Developments in Cross Section and Panel Count Models 87

Pravin K Trivedi and Murat K Munkin 5 An Introduction to Textual Econometrics 133

Stephen Fagan and Ramazan Genc¸ay 6 Large Deviations Theory and Econometric Information Recovery 155

Marian Grend´ar and George Judge 7 Nonparametric Kernel Methods for Qualitative and Quantitative Data 183

Jeffrey S Racine 8 The Unconventional Dynamics of Economic and Financial Aggregates 205

Karim M Abadir and Gabriel Talmain 9 Structural Macroeconometric Modeling in a Policy Environment 215 Martin Fukaˇc and Adrian Pagan

vii

viii Contents

10 Forecasting with Interval and Histogram Data: Some

Financial Applications 247 Javier Arroyo, Gloria Gonz´alez-Rivera, and Carlos Mat´e

11 Predictability of Asset Returns and the Efficient

15 Spatial Panels 435 Badi H Baltagi

16 Nonparametric and Semiparametric Panel Econometric

Models: Estimation and Testing 455 Liangjun Su and Aman Ullah

Index 499

Econometrics originated as a branch of the classical discipline of ical statistics At the same time it has its foundation in economics where itbegan as a subject of quantitative economics While the history of the quanti-tative analysis of both microeconomic and macroeconomic behavior is long,the formal of the sub-discipline of econometrics per se came with the estab-lishment of the Econometric Society in 1932, at a time when many of themost significant advances in modern statistical inference were made by JerzyNeyman, Egon Pearson, Sir Ronald Fisher, and their contemporaries All ofthis led to dramatic and swift developments in the theoretical foundations

mathemat-of econometrics, followed by commensurate changes that took place in theapplication of econometric methods over the ensuing decades From time totime these developments have been documented in various ways, includ-ing various “handbooks.” Among the other handbooks that have been pro-

duced, The Handbook of Applied Economic Statistics (1998), edited by Aman Ullah and David E A Giles, and The Handbook of Applied Econometrics and

Statistical Inference (2002), edited by Aman Ullah, Alan T K Wan, and Anoop

Chaturvedi (both published by Marcel Dekker), took as their general themethe over-arching importance of the interface between modern econometricsand mathematical statistics

However, the data that are encountered in economics often have unusualproperties and characteristics These data can be in the form of micro (cross-section), macro (time-series), and panel data (time-series of cross-sections).While cross-section data are more prevalent in the applied areas of micro-economics, such as development and labor economics, time-series data arecommon in finance and macroeconomics Panel data have been used exten-sively in recent years for policy analysis in connection with microeconomic,macroeconomic and financial issues Associated with each of these types ofdata are various challenging problems relating to model specification, estima-tion, and testing These include, for example, issues relating to simultaneityand endogeneity, weak instruments, average treatment, censoring, functionalform, nonstationarity, volatility and correlations, cointegration, varying co-efficients, and spatial data correlations, among others All these complex-ities have led to several developments in the econometrics methods andapplications to deal with the special models arising In fact many advanceshave taken place in financial econometrics using time series, in labor eco-nomics using cross section, and in policy evaluations using panel data In theface of all these developments in the economics and financial econometrics,

the motivation behind this Handbook is to take stock of the subject matter of

empirical economics and finance, and where this research field is likely tohead in the near future Given this objective, various econometricians who

ix

x Preface

are acknowledged international experts in their particular fields were missioned to guide us about the fast, recent growing research in economics

com-and finance The contributions in this Hcom-andbook should prove to be useful

for researchers, teachers, and graduate students in economics, finance, ology, psychology, political science, econometrics, statistics, engineering, andthe medical sciences

soci-The Handbook contains sixteen chapters that can be divided broadly into

the following three parts:

1 Micro (Cross-Section) Models

2 Macro and Financial (Time-Series) Models

3 Panel Data Models

Part I of the Handbook consists of chapters dealing with the statistical issues

in the analysis of econometric models analysis with the cross-sectional dataoften arising in microeconomics The chapter by Cameron and Miller reviewsmethods to control for regression model error that is correlated within groups

or clusters, but is uncorrelated across groups or clusters The importance ofthis stems from the fact that failure to control for such clustering can lead to

an understatement of standard errors, and hence an overstatement of cal significance, as emphasized most notably in empirical studies by Moultonand others These may lead to misleading conclusions in empirical and policywork Cameron and Miller emphasize OLS estimation with statistical infer-ence based on minimal assumptions regarding the error correlation process,but they also review more efficient feasible GLS estimation, and the adaptation

statisti-to nonlinear and instrumental variables estimastatisti-tors Trivedi and Munkin haveprepared a chapter on the regression analysis of empirical economic modelswhere the outcome variable is in the form of non-negative count data Countregressions have been extensively used for analyzing event count data thatare common in fertility analysis, health care utilization, accident modeling,insurance, and recreational demand studies, for example Several special fea-tures of count regression models are intimately connected to discreteness andnonlinearity, as in the case of binary outcome models such as the logit and pro-bit models The present survey goes significantly beyond the previous suchsurveys, and it concentrates on newer developments, covering both the prob-ability models and the methods of estimating the parameters of these models

It also discusses noteworthy applications or extensions of older topics other chapter is by Fagan and Gen¸cay dealing with textual data econometrics.Most of the empirical work in economics and finance is undertaken using cat-egorical or numerical data, although nearly all of the information available todecision-makers is communicated in a linguistic format, either through spo-ken or written language While the quantitative tools for analyzing numericaland categorical data are very well developed, tools for the quantitative anal-ysis of textual data are quite new and in an early stage of development Ofcourse, the problems involved in the analysis of textual data are much greaterthan those associated with other forms of data Recently, however, researchhas shown that even at a coarse level of sophistication, automated textual

An-processing can extract useful knowledge from very large textual databases.This chapter aims to introduce the reader to this new field of textual econo-metrics, describe the current state-of-the-art, and point interested researcherstoward useful public resources.

In the chapter by Golan and Greene an information theoretic estimator is veloped for the mixed discrete choice model used in applied microeconomics.They consider an extension of the multinomial model, where parametersare commonly assumed to be a function of the individual’s socio-economiccharacteristics and of an additive term that is multivariate distributed (notnecessarily normal) and correlated This raises a complex problem of deter-mining large number of parameters, and the current solutions are all based

de-on simulated methods A complementary approach for handling an determined estimation problem is to use an information theoretic estimator,

under-in which (and unlike the class of simulated estimators) the underdetermunder-inedproblem is converted into a constrained optimization problem where all of theavailable information enters as constraints and the objective functional is anentropy measure A friendly guide for applying it is presented The chapter byRacine looks into the issues that arise when we are dealing with data on eco-nomic variables that have nonlinear relationship of some unknown form Suchmodels are called nonparametric Within this class of models his contributionemphasizes the case where the regression variables include both continuousand discrete (categorical) data (nominal or ordinal) Recent work that ex-plores the relationship between Bayesian and nonparametric kernel methods

is also emphasized The last two chapters in Part I are devoted to exploringsome theoretical contributions Grend´ar and Judge introduce fundamentallarge deviations theory, a subfield of probability theory, where the typicalconcern is about the asymptotic (large sample) behavior, on a logarithmicscale, of a probability of a given event The results discussed have impli-cations for the so-called maximum entropy methods, and for the samplingdistributions for both nonparametric maximum likelihood and empirical like-lihood methods Finally, Antoine and Renault consider a general frameworkwhere weaker patterns of identification may arise in a model Typically, thedata generating process is allowed to depend on the sample size However,contrary to what is usually done in the literature on weak identification, theysuggest not to give up the goal of efficient statistical inference: even fragileinformation should be processed optimally for the purpose of both efficientestimation and powerful testing These insights provide a new focus that isespecially needed in the studies on weak instruments

Part II of the Handbook contains chapters on time series models extensively

used in empirical macroeconomics and finance The chapter by Fukaˇc andPagan looks at the development of macro-econometric models over the pastsixty years, especially those that have been used for analyzing policy options.They classify them in four generations of models, giving extremely usefuldetails and insights of each generation of models with their designs, the way

in which parameters were quantified, and how they were evaluated Abadirand Talmain explore an issue existing in many macroeconomic and aggregate

xii Preface

financial time-series Specifically, the data follow a nonlinear long-memoryprocess that requires new econometric tools to analyze them This is becauselinear ARIMA modeling, often used in standard empirical work, is not consis-tent with the real world macroeconomic and financial data sets In view of thisAbadir and Talmain have explored econometric aspects of nonlinear model-ing guided by economic theory The chapter by Ludvigson and Ng developsthe relationship between bond excess premiums and the macroeconomy byconsidering factors augmented panel regression of 131 months Macroeco-nomic factors are found to have statistically significant predictive power forexcess bond returns Also, they show that forecasts of excess bond returns (orbond risk premia) are countercyclical This implies that investors are compen-sated for risks associated with recessions In another chapter Pesaran exploresthe predictability of asset returns and the empirical and theoretical basis ofthe efficient market hypothesis (EMH) He first overviews the statistical prop-erties of asset returns at different frequencies and considers the evidence onreturn predictability, risk aversion and market efficiency The chapter thenfocuses on the theoretical foundation of the EMH, and shows that marketefficiency could coexist with heterogeneous beliefs and individual irrational-ity provided that individual errors are cross-sectionally weakly dependent,but at times of market euphoria or gloom these individual errors are likely tobecome cross-sectionally strongly dependent, so that the collective outcomecould display significant departures from market efficiency In deviation withthe above chapters in this part, which deal with the often used classical pointdata estimation, Arroyo, Gonz´alez-Rivera and Mat´e review the statistical lit-erature on the regression analysis and forecasting with the interval-valuedand histogram-valued data sets that are increasingly becoming available ineconomics and finance Measures of dissimilarities are presented which help

us to evaluate forecast errors from different methods They also provide plications relating to forecasting the daily interval low/high prices of theS&P500 index, and the weekly cross-sectional histogram of the returns to theconstituents of the S&P500 index

ap-Part III of the Handbook contains chapters on the types of panel data and

spa-tial models which are increasingly becoming important in analyzing complexeconomic behavior and policy evaluations While there has been an extensivegrowth of the literature in this area in recent years, at least two issues haveremained underdeveloped One of them relates to the econometric issuesthat arise when analyzing panel models that contain time-series dynamicsthrough the presence of lagged dependent variables Hsiao, in his chapter,reviews the literature on dynamic panel data models in the presence of unob-served heterogeneity across individuals and over time, from three perspec-tives: fixed vs random effects specification; additive vs multiplicative effects;and the maximum likelihood vs methods of moments approach On the otherhand, Su and Ullah, in their chapter, explore the often ignored issue of thenonlinear functional form of panel data models by adopting both nonpara-metric and semiparametric approaches In their review they focus on therecent developments in the econometrics of conventional panel data models

with a one-way error component structure; partially linear panel data els; varying coefficient panel data models; nonparametric panel data modelswith multi-factor error structure; and nonseparable nonparametric panel datamodels Within the framework of panel data or purely cross-sectional datasets we also have the issues that arise when the dependence across cross-sectional units is related to location and distance, as is often found in studies

mod-in regional, urban, and agricultural economics The chapter by Baltagi dealswith this area of study and it introduces spatial error component regressionmodels, and the associated methods of estimation and testing He also dis-cusses some of the issues related to prediction using such models, and studiesthe performance of various panel unit root tests when spatial correlation ispresent Finally, the chapter by Lee and Yu studies the maximum likelihoodestimation of spatial dynamic panel data where both the cross-section andtime-series observations are large A new estimation method, based on a par-ticular data transformation approach, is proposed which may eliminate timedummy effects and unstable or explosive components A bias correction pro-cedure for these estimators is also suggested

In summary, this Handbook brings together both review material and new

methodological and applied results which are extremely important to the rent and future frontiers in empirical economics and finance The emphasis

cur-is on the inferential cur-issues that arcur-ise in the analyscur-is of cross-sectional, series, and panel data–based empirical models in economics and finance and

time-in related discipltime-ines In view of this, the contents and scope of the Handbook

should have wide appeal We are very pleased with the final outcome and weowe a great debt to the authors of the various chapters for their marveloussupport and cooperation in the preparation of this volume We are also mostgrateful to Damaris Carlos and Yun Wang, University of California, Riverside,for the efficient assistance that they provided Finally, we thank the fine edi-torial and production staff at Taylor & Francis, especially David Grubbs andSuzanne Lassandro, for their extreme patience, guidance, and expertise

Aman Ullah David E A Giles

About the Editors

Aman Ullah is a Distinguished Professor and Chair in the Department of

Economics at the University of California, Riverside A Fellow of the Journal

of Econometrics and the National Academy of Sciences (India), he is the author

and coauthor of 8 books and over 125 professional papers in economics, metrics, and statistics He is also an Associate Fellow of CIREQ, Montreal,Research Associate of Info-Metrics Institute, Washington, and Senior Fellow

econo-of the Rimini Centre for Economic Analysis, Italy Precono-ofessor Ullah has been

a coeditor of the journal Econometric Reviews, and he is currently a member

of the editorial boards of Econometric Reviews, Journal of Nonparametric

Statis-tics, Journal of Quantitative Economics, Macroeconomics and Finance in Emerging Market Economies, and Empirical Economics, among others Dr Ullah received

the Ph.D degree (1971) in economics from the Delhi School of Economics,University of Delhi, India

David E.A Gilesis a Professor in the Department of Economics at the versity of Victoria, British Columbia, Canada He is the author of numerousjournal articles and book chapters, and author and coauthor of five books

Uni-including the book Seemingly Unrelated Regression Equations Models (Marcel Dekker, Inc.) He has served as an editor of New Zealand Economic Papers as well as an associate editor of Journal of Econometrics and Econometric Theory.

He has been the North American editor of the Journal of International Trade and

Economic Development since 1996, and he is currently associate editor of munications in Statistics and a member of the editorial boards of the Journal of Quantitative Economics, Statistical Papers, and Economics Research International,

Com-among others Dr Giles received the Ph.D degree (1975) in economics fromthe University of Canterbury, Christchurch, New Zealand

xv