Impact of global and domestic macroeconomic variables on Borsa Istanbul Stock returns

Abstract

This study investigates the impact of global and domestic macroeconomic variables on the monthly returns of stocks traded in Borsa Istanbul during 1998- 2011 period. 13 variables are chosen as the candidates of source of variations in stock returns: BIST 100 index, one month lagged series of BIST 100 index, realized volatility of BIST 100 index, price-to-earnings (P/E) ratio, trade volume, S&P 500 index, inflation, foreign trade balance, industrial production index, oil prices, gold prices, foreign exchange rate and interest rate on short term government bonds. Taking the return of BIST 100 index as the dependent variable, 5 of the remaining 12 variables, namely, S&P 500 index, trade volume, industrial production index, foreign exchange rate and interest rate on short term government bonds are selected as the best model by using an algorithm based on model selection criteria, AIC and LASSO. These variables are regressed on return of BIST 100 index and residual of this regression (market residual) is regarded as the 6th factor, then the explanatory power of this 6-factor model is analyzed on Fama-French size and book-to-market portfolios. Market residual, S&P 500 index and trade volume are found to have a statistically significant explanatory power on all portfolios, whereas industrial production index and foreign exchange rate are found to be effective with significance level of above 5%. On the other hand, interest rate on short term government bonds has no significant effect on stock returns.