Temporal optimisation of signals emitted automatically by securities exchange indicators

• Published in: Cuadernos de gestión 2020-01, Vol.20 (3), p.61-71
• Main Authors: Martín-García, Rodrigo, Ventura Pérez, Enrique, Arguedas-Sanz, Raquel
• Format: Article
• Language: English
• Subjects: Algorithms; análisis técnico; bolsa de valores; cruce ema; Decision making; ema; Equity; estrategia de trading; Investments; Literature reviews; optimal lags; Pattern recognition; Prices; Profitability; retardos óptimos; rsi; Securities markets; Stock exchanges; stock market; technical analysis; trading strategy
• ISSN: 1131-6837; 1988-2157
• DOI: 10.5295/cdg.170851rm

Stock exchange indicators deliver buy/sell signals that enable analysts to improve the results of a strategy based strictly on fundamental analysis. Nonetheless, since the automatic implementation of signals as they appear may not yield optimal returns, the present paper analysed the suitability of using a series of technical indicators as guidance for portfolio results. A second aim pursued was to study how delaying the implementation of indicator signals may enhance profitability. A simulation was performed for the years 2005-2016 using the most representative index for the Spanish stock exchange, the IBEX35 and all its constituent securities, along with seven indicators (RoC, RSI, SMA, EMA, MACD, Bollinger bands and Stochastic Oscillator) and a total of 81 combinations of buy/sell lag times. The definition of three non-overlapping sub-periods to guarantee the reliability of the findings yielded a total of 61 236 simulated portfolios. The conclusion drawn from the results was that for certain combinations of indicators, delaying the implementation of buy/sell signals improves returns. More specifically, optimal lag times identified for RSI and EMA signals were shown to deliver statistically significant improvements in portfolio returns, irrespective of the period studied. Those findings were consistent the results of an alternative simulation in which the five securities that were both the most liquid and had the greatest impact on the index were not considered, to rule out the possible effect of the relative weight of securities on either portfolio returns or their normalisation.