after correction for transaction costs, are mainly found for the Asian, Latin American, Middle East and Russian stock market indices. For example, the best recursive optimizing and testing procedure generates an excess return over the buy-and-hold of 43.52, 32.42, 20.99, 12.61, 7.50 and 4.88% yearly for the Argentinean Merval, after implementing 0, 0.10, 0.25, 0.50, 0.75 and 1% transaction costs. However, for the US, Japanese and most Western European stock market indices the recursive out-of-sample forecasting procedure does not show to be profitable, after implementing transaction costs.
If the Sharpe ratio criterion is used for selecting the best strategy during the training period, then the Sharpe ratio of the best recursive optimizing and testing procedure in excess of the Sharpe ratio of the buy-and-hold benchmark is on average 0.0544, 0.0419, 0.0298, 0.0164, 0.0086 and 0.0052 in the case of 0, 0.10, 0.25, 0.50, 0.75 and 1% costs per trade (see second to last row of table 5.15B). As for the mean return selection criterion, the best recursive optimizing and testing procedure does not generate excess Sharpe ratios over the buy-and-hold for the US and most Western European indices in the presence of transaction costs. The best results are mainly found for the Latin American, Egyptian and Asian stock market indices.
For comparison, the last rows in tables 5.15A and 5.15B show the average over the results of the best strategies selected by the mean return or Sharpe ratio criterion in sample for each index tabulated. As can be seen, clearly the results of the best strategies selected in sample are better than the results of the best recursive out-of-sample forecasting procedure.
For the cases that the best strategy in the optimizing period is selected by the mean return and Sharpe ratio criterion respectively, tables 5.16A and 5.16B show for the 0 and 0.50% transaction cost cases the estimation results of the Sharpe-Lintner CAPM (see equation 5.1), where the return in US Dollars of the best recursive optimizing and testing procedure in excess of the US risk-free interest rate is regressed against a constant α and the return of the local main stock market index in US Dollars in excess of the US risk-free interest rate. Tables 5.17A, B summarize the CAPM estimation results for the two possible choices of the market portfolio and for all transaction cost cases by showing the number of indices for which significant estimates of α or β are found at the 10% significance level. Estimation is done with Newey-West (1987) heteroskedasticity and autocorrelation consistent (HAC) standard errors.
If the local main stock market index is taken to be the market portfolio and if the best strategy in the training period is selected by the mean return criterion, then in the case of zero transaction costs it can be seen in table 5.17A that for 37 indices a significantly positive estimate of α is found. As can be seen in table 5.16A, mainly for the US, Japan