Objective research to aid investing decisions

Value Investing Strategy (Strategy Overview)

Allocations for October 2024 (Final)
Cash TLT LQD SPY

Momentum Investing Strategy (Strategy Overview)

Allocations for October 2024 (Final)
1st ETF 2nd ETF 3rd ETF

Technical Analysis Tested on Long-run DJIA Data

| | Posted in: Technical Trading

Does technical analysis work after accounting for luck and trading frictions? More specifically, can traders reliably identify technical rules that generate future net outperformance? In the January 2008 version of their paper entitled “Technical Trading Revisited: Persistence Tests, Transaction Costs, and False Discoveries”, Pierre Bajgrowicz and Olivier Scaillet investigate the economic value of technical trading rules applied to long-run daily Dow Jones Industrial Average (DJIA) data. Their approach includes: (1) a new measure of data snooping bias to distinguish between luck and true forecasting power in backtesting; (2) out-of-sample persistence testing of recently successful trading rules; (3) determination of whether certain trading rules work consistently under specific economic conditions; and, (4) incorporation of trading costs. Using daily DJIA price and volume data for January 1897 through July 2007 to test 7,846 rules (filters, moving averages, support and resistance, channel breakouts and on-balance volume averages), they conclude that:

  • During 1897-1996, before transaction costs, technical analysis appears to be a useful tool for generating outperformance, even after accounting for data snooping.
  • The historical outperformance of technical rules before transaction costs, however, dissipates over time. During January 1997 through July 2007, backtested outperformance disappears across the universe of trading rules. Cheap computing power, falling transaction costs, and growing liquidity may have suppressed patterns in stock returns.
  • Moreover, the backtested success in older data disappears after accounting for comprehensive transaction costs and short sale constraints. For example, the top DJIA trading rule over 1897-1996 generates a mean annual return of 15.9% from an average of 117 transactions per year, implying a breakeven total trading friction of 0.14% per trade.
  • In general, an investor could not have selected ex ante the best future trading rules. The performance of a portfolio of trading rules derived from rolling two-year historical datasets does not persist out-of-sample.
  • Selecting specific trading rules according to business cycle phase (as designated by the National Bureau of Economic Research) does not enhance returns.

In summary, this evidence does not support a belief that technical trading rules reliably generate out-of-sample outperformance after accounting for trading frictions.

Login
Daily Email Updates
Filter Research
  • Research Categories (select one or more)