АНАЛІЗ ЕФЕКТИВНОСТІ МЕТОДІВ ТЕХНІЧНОГО АНАЛІЗУ ПРИ ПРОГНОЗУВАННІ ФОНДОВИХ ІНДЕКСІВ

Abstract

The article analyzes the profitability of technical analysis methods for the seven stock indices during the last ten years. According to the analysis, the profitability of technical analysis has increased recently due to changes in market conditions. However, the efficiency of technical analysis methods was much lower during 2010-2018. The analysis showed that technical analysis methods demonstrated best results on the Chinese, Indian, and Hong Kong stock indices, the worst – on the American, European, and Japanese stock indices. However, the stability of these methods is quite low: their profitability varies greatly with the change of the sample. The issue of aggregation of technical analysis signals and ARIMA-model signals is also considered in this paper. The optimal parameters for the technical analysis methods were selected by testing on historical data; optimal ARIMA models were selected for each index. For 3 out of 7 indices the optimal model is WN (white noise). Most technical analysis methods showed poor results on the American (S&P 500) and European (Euronext 100) stock indices (except for the last two years). The results can be used to develop trading strategies.The analysis showed that technical analysis methods demonstrated best results on the Chinese, Indian, and Hong Kong stock indices, the worst – on the American, European, and Japanese stock indices. However, the stability of these methods is quite low: their profitability varies greatly with the change of the sample. The issue of aggregation of technical analysis signals and ARIMA-model signals is also considered in this paper. The optimal parameters for the technical analysis methods were selected by testing on historical data; optimal ARIMA models were selected for each index. For 3 out of 7 indices the optimal model is WN (white noise). Most technical analysis methods showed poor results on the American (S&P 500) and European (Euronext 100) stock indices (except for the last two years). The results can be used to develop trading strategies.