Indicator based trend following and classical chart patterns are two trading methods that were developed in mid 20th century using data from the equity markets mainly and worked well during an extended period of time in those markets due to the presence of autocorrelation. After 1998 things got harder because serial correlation in equity indices decreased due to arbitrage and by 2007 it was mostly gone rendering these methods largely ineffective.
On the above daily chart of S&P 500 from 01/03/1950 to 12/19/2012 the bottom pane is the 1-Lag rolling 120-day autocorrelation of daily arithmetic returns [X(i+1)/X(i) - 1]. It suffices to observe that from 01/1950 to 04/1988 the autocorrelation was very high, especially after 07/1964 and through 04/1988, or for a period of 24 years, with very few and short periods of negative autocorrelation. That was a very good time for technical methods based on trend-following and chart patterns during which some traders might have gotten the impression that these methods are significant although their success was clearly due to the high autocorrelation in the equity markets, i.e. the fact that future prices quite often behaved like past prices. In other markets, like commodity futures and currencies, which started trading in the mid 1970s, these technical methods were later applied and expected to work because they were raised to causation rules by some authors but in reality they failed more often than they worked well and were basically responsible for the high rate of failure of retail traders and even some funds.
The struggle of managed futures performance to generate alpha in the last four years, as reported in recent posts, is also for the most part due to the same effect, i.e. due to the fact that the majority of these funds employ trend following based on methods that were developed in the 1970s and 1980s using induction after an apparent success in the equity markets for a period of two decades before and attributed here to the serial correlation shown on the above S&P 500 chart.
In the EURUSD daily spot exchange rate chart since 01/1998, as shown above, it is clear that before 2005 for the most part the 1-Lag rolling 120-day autocorrelation was negative with only scarce periods of positive values. The autocorrelation values shown on the chart above may not be for the most part significant but this is not the main issue here. What is important to realize is that trend-following and chart patterns, especially those based on weekly or even monthly data, could work well only during a few specific periods. Experienced currency traders knew all along that technical trading methods developed on equity data did not perform well in currency trading because of the unpredictability of those markets. The same is the case for commodities markets. Below is a chart of crude oil futures, continuously adjusted:
In this case too we observe that for an extended period of time from 2003 to 2008, the autocorrelation was overwhelmingly negative and for the most part probably insignificant. There was a period in the early 2000s of positive autocorrelation and then during 2012. These rare periods of positive autocorrelation maybe gave the false impression and even hope to some technical traders that trend-following and chart patterns developed long ago by analyzing equity data do work in other markets too but actually these specific methods were irrelevant and “monkey style trading“, i.e trading with random entries and exits that attempt to follow the trend can work as well. This fact has been demonstrated over and over again by very competent analysts and traders and there is no point in elaborating further on it here.
The same phenomenon can be noticed on the continuous sugar futures chart above. The lack of persistent positive autocorrelation is evident. No wonder why technical signals generated by chart patterns and trend-following systems have failed in this market too in the last two years given that after 2010 the autocorrelation has been overwhelmingly negative, as shown on the above chart.
Most technical methods for longer-term and position trading were developed in the middle part of the 20th century and started being used extensively in the 1980s with the advent of the personal computer that allowed their practical implementation and automation. However, these method were developed primarily using data from the equity market simply because most commodities and currencies were not traded back then in organized exchanges. Their developers falsely attributed the apparent autocorrelation in the equity markets that basically turned profitable any method based on buying on strength to some predictive capacity of these methods. Essentially, they were fooled by randomness and labeled some dubious statistics causal relationships. Later, other technical traders and book authors raised the methods developed for the equity markets to general trading rules for all markets using totally unjustifiable induction. Examples are rules like”if pattern XYZ forms and it is confirmed, then the market will move higher towards a target equal to T” or “if a moving average of X days crosses above a moving average of Y days, then the market will trend up”. However, the conditions that made these rules profitable were not present in the other markets with the same frequency and thus a high percentage of commodity and currency traders that used them in the 1980s and 1990s lost a lot of money to locals. The profitability of the trading methods in the equity markets fell steadily after 1988 and by 2007 it was gone as the autocorrelation of the daily returns was arbitraged out by fast traders. It is quite unlikely that autocorrelation will increase again to levels and frequencies required for the profitability of technical methods that rely on its existence, like trend following and chart pattern trading. High Frequency Trading (HFT) is the main mechanism that is currently in effect to arbitrage out autocorrelation and its profits come mainly from technical traders and funds that trade using methods developed in the 1950s when there were pits, traders calculated moving averages manually and charts were available on a weekly basis. Nowadays, these methods do not have any inherent predictive capacity and have been abandoned by the majority of technical traders in favor methods like scalping, price action trading, medium frequency trading, pairs trading and others, know or unknown to the public, most of which are based on mean-reversion principles rather than on trend following or directional move confirmation like with chart patterns, for example.
Disclosure: no relevant position at the time of this post and no plans to initiate any positions within the next 72 hours..
Charting program: Amibroker (Charts created with AmiBroker – advanced charting and technical analysis software. http://www.amibroker.com/”)