testing the efficient market hypothesis .testing the efficient market hypothesis of the helsinki

Download TESTING THE EFFICIENT MARKET HYPOTHESIS .Testing the Efficient Market Hypothesis of the Helsinki

Post on 18-Sep-2018

213 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • Research Reports

    Kansantaloustieteen laitoksen tutkimuksia No. 91:2002

    Dissertationes Oeconomicae

    Carolina Sierimo

    TESTING THE EFFICIENT MARKET HYPOTHESIS OFTHE HELSINKI STOCK EXCHANGE

    Further empirical evidence based on nonlinear models

    ISBN 952-10-0426-6 (nid.)ISBN 952-10-0430-4 (pdf)

    http//:ethesis.helsinki.fiYliopistopainoHelsinki 2002

  • HELSINGIN YLIOPISTO )HELSINGFORS UNIVERSITET )UNIVERSITY OF HELSINKITiedekunta/Osasto ) Fakultet/Sektion ) Faculty Laitos ) Institution ) Department

    Faculty of Social Sciences Department of EconomicsTekij ) Frfattare ) Author

    Carolina SierimoTyn nimi ) Arbetets titel ) Title

    Testing the Efficient Market Hypothesis of the Helsinki Stock ExchangeFurther empirical evidence based on nonlinear modelsOppiaine ) Lromne ) Subject

    EconomicsTyn laji ) Arbetets art ) Level Aika ) Datum ) Month and year Sivumr ) Sidoantal) Number of pages

    Thesis 2002 165Tiivistelm ) Referat ) Abstract

    In an Efficient Market prices adjust instantaneously toward their fundamental values, andtrading volume contains no information about future price developments. However, thesequential arrival of information model implies a positive causal relation between absolutestock returns and trading volume in either direction, and the mixture model suggests apositive causal relation running from volume to absolute returns. Empirical evidence supportsthese theories, since large movements in stock prices typically take place on days with hightrading volume. Here, we consider the relationship between stock return and trading volumeusing a nonlinear framework. First, nonlinearity in the Helsinki Stock Exchange data is testedusing a STAR-model. Next, we compute nonlinear impulse responses; first to check thestability of the models, and second to scrutinize the persistence of shocks in return andvolume series. Linear Granger causality tests indicate bidirectional causality between returnsand volume. By contrast, the nonlinear causality tests suggest that only in a few cases canvolume be used to forecast returns. Thus, the empirical findings give only slight support to themixture model. Finally, we exploit the outcome of causality tests to specify a STVEC modelin order to take into account the influence of the composition effect and common persistenceon the results. We conclude that causality runs mainly from returns to trading volume,corroborating the positive feedback trading hypothesis.

    There is a large body of empirical evidence that stock markets perform poorly duringinflationary periods. Several explanations have been offered for this so-called anomaly.Here, we test the claim that the spurious negative correlation between stock returns andinflation is due to counter-cyclical monetary policy. We identify various regimes in theFinnish data using the MS-VAR model. When necessary, the existing long run relation-shipsbetween the model variables are incorporated using the MS-VEC model. Using alternativesets of explanatory variables including measures for monetary policy stringency, we concludethat the sign of the relation between returns and inflation depends especially on the timehorizon chosen. In monthly models the statistically significant contempo-raneous correlationsbetween returns and inflation are always negative, but positive in the case of quarterly data.Stocks thus seem to be a good hedge against inflation in the long run. To be more specific,stocks seem to maintain protection against purely monetary inflation but fail to provide ahedge against inflation arising from real output shocks. Last, this is tested using the regime-dependent impulse response function.Avainsanat ) Nyckelord ) Keywords

    empirical study, nonlinear models, Efficient Market Hypothesis, stock exchange - Finland,returns - trading volume, returns - inflation, returns - monetary policySilytyspaikka ) Frvaringsstlle ) Where deposited

    Social Science LibraryMuita tietoja ) vriga uppgifter ) Additional information

  • Habe nun, ach! Philosophie, Juristerei, Theologie und Medizin und leider auchVolkswirtschaft durchaus studiert, mit heiem Bemhn. Da steh' ich nun, ich armer Tor, undbin so klug als wie zuvor! Heie Magister, heie Doktor gar, und ziehe schon and die zehen

    Jahr' herauf, herab und quer und krumm meinen Schler an der Nase herum- Und sehe, da wir nichts wissen knnen!Das will mir schier das Herz verbrennen.

    Aus dem FaustJohann Wolfgang von Goethe

    1749 - 1832

  • Foreword

    This thesis is dedicated to my grandmothers Aune Hakkarainen and Eeva Hytnen. Both of themwere very strong women working their way through many obstacles. I hope I have inherited atleast some of their strength.

    Professor Matti Virn taught me to persist in doing empirical work while I worked as a researchassistant for him. Without his example, I would probably have given up this work at some point.I also want to thank Professor Seppo Honkapohja for his personal support (henkinen tuki). Iwill not even try to describe it as we would probably end up arguing about the exact definition,in any case. My special thanks go to Seppos wife! as who else would have taught him to quarrelin such a pleasant, civilized way? Without the care of the Departmental Administrator RitvaTervinen, the Department of Economics of the University of Helsinki would have probablyceased to exist. I am sure that I will also need her valuable help and knowledge about everythinggoing on in the academic world in the future.

    Outi Yli-Renko has kept me in touch with real life during the long years I have been working onthis thesis. She has patiently listened to all my troubles during our long walks on Sundays. I ammost grateful for these breaks from research, and I would like to repay this somehow at somepoint in our lives. Elina, Minna, Sirkka, and Virve have also always been there when needed.Finally, I want to thank the man of my life for a broken hearth.

    This thesis was completed during my visit to Jaana Keitaanrantas home in Monteverde, Rome.Though the distance between Rome and Helsinki is great, it has not stopped our friendship.Through E-mail she encouraged me in my work when I have had doubts about whether it wasever going to be completed.

  • Acknowledgements

    I wish to express my gratitude to all who have supported me in the course of this study.

    Chapter two of this thesis has benefitted from the help of Professor Timo Tersvirta and his teamat the Stockholm School of Economics. I also want to thank Dr. Dick van Dijk (ErasmusUniversity Rotterdam) for his comments and Professor Markku Rahiala (University of Oulu) forhis help. Furthermore, Anna Persson, Joakim Skalin, Stefan Lundbergh (all three were workingas Ph. D. students at the time of this study at the Stockholm School of Economics) and Dr. vanDijk have provided me with various GAUSS procedures for use, and helped me in modifyingthem to suit my own needs.

    The computations reported in Chapter 3 were carried out using Dr. Hans-Martin Krolzigs(Institute of Economics and Statistics, University of Oxford) MS-VAR for Ox (Doornik, 1998and Hendry & Doornik, 1999), and the regime-dependent impulse responses computations werebased on the Ox procedures by Ehrmann, M., Ellison, M. and N. Valla (European UniversityInstitute). I am also grateful for all the help I have received from Hans-Martin Krolzig andMarianne Sensier, Ph. D. (School of Economic Studies, Universityof Manchester) while workingon this Chapter. Special thanks go to Eija Toivanen (Bank of Finland), who has helped me inupdating my data set.

    I am also grateful to my official examiners, Professor Johan Knif (Swedish School of Economicsand Business Administration, Vaasa) and Professor Seppo Pynnnen (University of Vaasa), fortheir constructive comments and criticism.

    Roderick McConchie, Ph. D., has been of invaluable help in improving my English. His variousquestions pointed out some of the weaknesses of the manuscript as well.

    Any errors that may remain are my own.

    Financial support from the Yrj Jahnsson Foundation is gratefully acknowledged.

    Helsinki 26.3.2002

    Carolina Sierimo

  • Contents

    1. Introduction...........................................................................................................................71.1 The motivation of this study..............................................................................................71.2 New approaches in testing the Efficient Market Hypothesis...........................................11

    1.2.1 The motivation for using STAR, STAR-GARCH and MS-VAR models..............111.2.2 Nonlinearity and structural breaks...........................................................................17

    1.3 The generalized impulse response function....................................................................181.4 Introduction to Chapter 2................................................................................................211.5 Introduction to Chapter 3................................................................................................221.6 Contribution....................................................................................................................25

    References .................................................................................................................................28

    2. Modeling nonlinearity in the stock return ! trading volume relationship............332.1 Theoretical background for the stock return ! trading volume relationship..................33

    2.1.1 The no-trade theorem ...........................................................................................362.1.2 Linear dependences ..............................................................................................362.1.

Recommended

View more >