Statik ve dinamik varlık fiyatlama modellerinin İstanbul Menkul Kıymetler Borsası'nda gözlemlenen anomalilere cevapları

This paper documents various explanations for cross-sectional financial market anomalies observed in the Istanbul Stock Exchange. In this framework, we test unconditional and conditional asset pricing models against non-risk firm characteristics by using data on individual securities. In conditional asset pricing models, the factor loadings are allowed to vary with market capitalization, book-to-market ratio and a business cycle indicator. As a business cycle indicator for Turkey, we use the composite leading indicator which is calculated by Central Bank of the Republic of Turkey.In order to examine the relationship between individual stock returns and non-risk firm characteristics such as size, book-to-market, past returns and liquidity, we apply Fama-Macbeth type regressions in two steps. In the first step, excess stock returns are regressed on factor loadings that vary with size, book-to-market and the macroeconomic variable. Specifically, we use five models: 1) CAPM, 2) Fama-French three factor model, 3) Fama-Frech augmented by momentum factor, 4) Fama-French augmented by illiquidity factor and 5) Fama-French augmented by momentum and illiquidity factor. In the second step, we run cross-sectional regressions in which risk adjusted returns calculated in step 1 are used as dependent variable and regressed on firm specific characteristics. We use monthly returns, and size, book-to-market, past returns and illiquidity for the time period covering February 1997 to April 2008. The results suggest that dynamic versions of multi-factor asset pricing models are able to capture the impact of firm-specific characteristics on expected stock returns but book-to-market effect persists in the static version of all models.