Essays on Technical Analysis and Asset Price Prediction
This thesis consists of three essays tied together with the common thread of technical analysis in asset pricing. They extend the technical analysis literature on data snooping bias, long-range dependence, and mixed frequency technical trading, respectively.
The first essay introduces an aggregate technical trading index by extracting the most relevant forecasting information contained in 7,846 technical trading rules to predict equity risk premium in the U.S. The proposed method significantly outperforms the existing false discovery rate (FDR) method in both in-sample and out-of-sample analysis.
The second essay focuses on the use of macroeconomic variables and technical indicators’ ability to predict equity risk premium. A Bullish Index is introduced to measure the changes in stock market behavior. A positive (negative) shock of the Bullish Index is closely related to strong equity risk premium predictability for forecasts based on macroeconomic variables (technical indicators) for up to six (nine) months.
The third essay studies the forecasting power of technical trading signals generated at various frequency levels with the Mixed Data Sampling model. The proposed aggregated high-frequency technical indicators (daily or weekly) can add value for momentum trading strategies compared to low-frequency (monthly) technical indicators, with a net-of-transactions-costs annualized certainty equivalent return gain up to 1.54%. The results indicate that there is no clear evidence of additional economic value for moving average trading strategies in forecasting monthly U.S equity risk premium.
In conclusion, this thesis provides new methodologies for predicting equity risk premium with technical analysis in the U.S market. The empirical results support the proposed methods with robustness check in all aspects of investing, i.e. the choice of risk preferences, transaction costs, out-of-sample analysis, and subsample period analysis. This thesis provides solid evidence in supporting the use of technical analysis in predicting the stock market movement.
Supervisor(s)Godfrey Charles-Cadogan, Marcel Ausloos
Date of award2023-06-28
Author affiliationDepartment of Economics, Finance and Accounting
Awarding institutionUniversity of Leicester