ISTANBUL STOCK EXCHANGE

Istanbul Stock Exchange

Table of Contents

INTRODUCTION3

FAMA-FRENCH'S 3 FACTOR MODEL5

CARHART'S 4 FACTOR MODEL6

DATA AND METHODOLOGY11

FINDINGS AND COMPARISON17

Fama-French's 3 factor model17

Carhart's 4 factor model25

Comparison31

CONCLUSIONS32

REFERENCES35

Istanbul Stock Exchange

Introduction

Previous empirical research on financial markets finds evidence against the classical capital asset pricing model (CARHART) of Sharpe (1964) and Lintner (1965). Asset returns are characterized by skewness and significant leptokurtosis providing evidence against the normality assumption. Furthermore, additional factors such as size and value, which are called Fama-French factors in the literature, have been shown to be significant determinants of excess returns in financial markets. Previous empirical literature on emerging financial markets takes these issues into consideration and analyses the importance of Fama-French factors as well as higher comoments.

This article uses data from the Istanbul Stock Exchange (ISE) and adds to the existing literature on emerging markets in three dimensions. First, when compared to previous studies on ISE, we use an extended data set and include all stocks that are listed in ISE during 1996-2005. Second, this article is a first attempt at understanding the relative importance of coskewness in explaining the variation of excess returns in ISE. We initially estimate a traditional CARHART and then examine the incremental effect of Fama-French factors as well as coskewness on the variation of portfolio excess returns. Third, in contrast to previous studies on ISE, we perform the multivariate test of Gibbons et al. (1989) to investigate whether CARHART is sufficient to explain the expected returns or multifactor models adding small minus big (SMB), high minus low (HML) and/or coskewness to the market factor would remove any pricing bias captured by the intercept term. The multivariate test investigates whether the pricing biases are jointly equal to zero; furthermore, it reveals whether the market portfolio or a linear combination of factor portfolios lies on the minimum variance boundary. We also run cross-sectional regressions following Fama-MacBeth (1973) as well as full information maximum likelihood (FIML) method to uncover the incremental power of coskewness over CARHART and Fama-French factors.

Our estimations indicate the presence of a significant relationship between average excess returns and coskewness in ISE, especially for size and industry portfolios. We argue that coskewness is able to account for the size premium, i.e. small market capitalization stocks have higher returns than big ones. While a two-factor model incorporating coskewness has more explanatory power than the traditional CARHART in cross-sectional regressions, that power diminishes as we include Fama-French factors.

Fama-French's 3 factor model

A factor model that expands on the capital asset pricing model (CARHART) by adding size and value factors in addition to the market risk factor in CARHART. This model considers the fact that value and small cap stocks outperform markets on a regular basis. By including these two additional factors, the model adjusts for the outperformance tendency, which is thought to make it a better tool for evaluating manager performance.

The three factor model is motivated by the empirical finding that size and the ratio of book-to-market equity have consistent and significant explanatory power for US stock returns at the ...