Technical Indicators Predict Cross-Sectional Expected Stock Returns

Technical analysis is a form of investment valuation that analyses past prices to predict future price action. Numerous studies document that technical analysis can predict individual stock returns. However, much less is known about how technical indicators explain the variation in the cross-section of stock returns. A research paper by Zeng, Marshall, Nguyen, and Visaltanachoti (2021) fills this gap and investigates the performance of 14 well-documented technical indicators in explaining cross-sectional stock returns. The studied technical indicators include moving averages, momentum, and on-balance volume with different lengths. The authors utilize the indicators jointly in cross-sectional regressions, which allows them to incorporate information from 14 technical indicators simultaneously. The resultant estimated smoothed OLS (SOLS) model is used to predict cross-sectional stock returns. The study results have shown that the SOLS model with joint 14 technical indicators generates positive cross-sectional and time-series out-of-sample R-squared, and its performance is robust over time. Additionally, it outperforms the three-factor model of Fama and French (1993) with lower errors. Naturally, an accurate forecasting model, such as the SOLS model, can be utilized to develop a profitable trading strategy.

Fundamental reason

Trend-following strategies based on technical indicators generate buy (sell) signals in a positive (negative) market trend attested by a recent increase (decrease) in stock price. Existing literature documents that technical indicators represent a handy tool for predicting stock returns. For example, Zhu and Zhou (2009) provide theoretical support on how technical analysis adds value to the asset allocation by investing between a riskless bond and predictable individual stocks. In a recent study, Zeng, Marshall, Nguyen, and Visaltanachoti (2021) find that technical indicators show significant predictive ability in forecasting time-varying individual stock returns, especially for the high limits of arbitrage firms. Accordingly, a model combining the information from multiple trend-following technical indicators simultaneously can accurately identify a trend in stock prices and explain the variation of stock returns in a cross-section.

Get Premium Strategy Ideas & Pro Reporting

  • Unlocked Screener & 200+ Advanced Charts
  • 600+ uncommon trading strategy ideas
  • New strategies on a bi-weekly basis
  • 1000+ links to academic research papers
  • 400+ out-of-sample backtests
  • Design multi-factor multi-asset portfolios
Markets Traded
equities

Backtest period from source paper
1932-2020

Confidence in anomaly's validity
Strong

Indicative Performance
41.91%

Notes to Confidence in Anomaly's Validity

Notes to Indicative Performance

Table 3, Panel B, period 1932:01-2020:12, annualized mean monthly return of 2.96%


Period of Rebalancing
Monthly

Estimated Volatility
11.78%

Notes to Period of Rebalancing

Notes to Estimated Volatility

Table 3, Panel B, period 1932:01-2020:12, annualized standard deviation of 3.40%


Number of Traded Instruments
1000

Maximum Drawdown

Notes to Number of Traded Instruments

The exact number of traded instruments depends on individual needs for diversification.


Notes to Maximum drawdown

not stated


Complexity Evaluation
Complex strategy

Sharpe Ratio
2.73

Notes to Complexity Evaluation

Region
United States

Financial instruments
stocks

Simple trading strategy

The investment universe consists of all firms from the CRSP database listed on NYSE, AMEX, and NASDAQ. Firstly, exclude all firms with less than 60 monthly return observations. Secondly, construct 14 firm-level technical indicators based on three trend-following strategies (moving average, momentum, and volume-based indicators). The first strategy is based on the moving average rule, which forms the trading signals by comparing the two moving averages with different lengths. The second strategy is based on the momentum trading rule, which generates the trading signals by comparing the current stock price with its level n months ago. The third strategy is based on the “on-balance” volume rule, which generates the trading signals by evaluating the changes in stock trading volume. For a detailed description of the technical indicators’ construction, see section 2.2. Thirdly, each month t regress the return of each stock i on 14 technical indicators from month t-1, using a fixed window of the latest 60 monthly observations to estimate the return over the next month (see equations 5 and 6). To mitigate the overfitting problem, take the time-series average of the cross-sectional OLS estimated coefficients applying a 60-month smoothing window (see equations 7a, 7b, and 7c). At the end of each month, sort all stocks into value-weighted deciles based on their estimated returns in the next month. Buy the top decile (stocks with the highest expected returns) and sell the bottom decile (stocks with the lowest expected returns). The resulting long-short portfolio is value-weighted and rebalanced monthly.

Hedge for stocks during bear markets

Not known - Overall, the strategy has an insignificant beta (equal to zero), but a more detailed analysis should be performed (primarily focusing on crises).

Source paper
Out-of-sample strategy's implementation/validation in QuantConnect's framework (chart+statistics+code)
Other papers

Get Quantpedia Premium or Pro

  • Unlocked Screener & 200+Advanced Charts
  • 600+ uncommon trading strategy ideas
  • New strategies on a bi-weekly basis
  • 1000+ links to academic research papers
  • 400+ out-of-sample backtests
  • Design multi-factor multi-asset portfolios

Subscribe for Newsletter

Be first to know, when we publish new content


    logo
    The Encyclopedia of Quantitative Trading Strategies

    Log in

    GET NOW
    BLACK FRIDAY DEAL
    GET NOW
    BLACK FRIDAY DEAL