Are Chartists Artists? The Determinants and Profitability of Recommendations Based on Technical AnalysisVW Staff
Are Chartists Artists? The Determinants and Profitability of Recommendations Based on Technical Analysis
Utrecht University – Utrecht University School of Economics
February 22, 2016
1. Introduction The relevance of recommendations published by security analysts has been subject to extensive academic research. The larger part of the literature is directed towards recommendations on the basis of fundamental analysis. Technical analysts represent a different category. They believe that past stock prices and trading volume may show patterns that indicate future trends. If that were true, price patterns on the stock market1 would contradict weak-form market efficiency, which states that all information from historical data is already incorporated in current prices.
Tools based on technical analysis (TA) are widely available to investors. Many brokers offer TA functionalities to their clients, and investors can furthermore rely on commercial charting packages offered by professional vendors. It is therefore not surprising that TA is broadly used among investors. For the Netherlands, Hoffmann et al. (2010) showed that the number of private investors using TA was larger than the number of investors relying on fundamental analysis. The use of TA is not limited to private investors only. For professional investors, Carter and Van Auken (1990) and Menkhoff (2010) found that 35 percent and 87 percent, respectively, considered TA to be important for trading decisions.
Most of the research regarding the profitability of TA focuses on the usefulness of individual trading rules (i.e., trading rules based on one single method), of which many exist. Common trading rules rely on moving averages and on trading range breakouts (Brock et al., 1992). These rules are mostly applied on observed stock prices; past trading volume is generally only used as a secondary tool (Sullivan et al., 1999). Although some studies support the value of TA to some extent (e.g., Lo et al., 2000; Wong et al., 2003; Chong and Ng, 2008; and Metghalchi et al., 2008), many others did not find any evidence that TA can be used to generate abnormal returns (e.g., Kwon and Kish, 2002; Tian et al., 2002; Lento et al., 2007; Marshall and Cahan, 2005; Marshall et al., 2009; Schulmeister, 2009). Confronted with academic criticism of their methodology, technicians occasionally respond that technical analysis is an art rather than a science, as also stated by DeMark (1994: xi): “Technical analysis has always had more art than science to it”. This suggests that technicians take into account more than simple trading rules when formulating investment recommendations. Therefore, in order to address this ‘art’-component of TA, not trading rules but TA-based recommendations published by specialized technicians should be studied, particularly because the ‘art’-aspect of a technical analyst is likely to transcend the pure TA rules. Two major questions are relevant here: first, are recommendations associated with positive abnormal returns, and second, to what extent do these recommendations differ from signals derived from technical trading rules?
Evaluations of recommendations issued by technical analysts are relatively scarce and evidence is mixed. Cowles (1933) was the first to analyze recommendations published by technicians. He found that this type of recommendation published in the Wall Street Journal underperformed a buy-and-hold strategy. Brown et al. (1998) applied different statistical methods to Cowles’ dataset and found that these recommendations in fact yielded risk-adjusted abnormal returns. Dawson (1985) analyzed recommendations issued by a Singapore investment advisory firm. He found that the recommended stocks did not outperform the market. Dawson (1985: 183) added that “from an optimal research perspective more than one investment advisor should be included”. However, no other TA sources were available at that time. The existing studies (Cowles, 1933; Dawson, 1985; and Brown et al., 1998) have several limitations: the number of considered recommendations is small, and the recommendations are published by only a limited number of technical analysts. Furthermore, the short-term profitability of TA has not been tested in these papers while Menkhoff (2010) reported that TA was most frequently used for investment decisions with a horizon of just some weeks. 2