To avoid such non-representative backtesting in the future, the authors advocate for simple smart beta methodologies and more controls and guidelines in backtesting methodology to limit upward biases. To limit the impact of transaction costs, the authors recommend avoiding smart beta strategies with high turnover rates and in low-liquidity areas of the markets. Most importantly, investors should keep in mind that live and future returns will probably be lower than the returns produced by backtests.
Research Affiliates still holds that backtesting is a useful tool to frame forward-looking expectations and to gain an understanding of the risks associated with an investing strategy. However, the smart beta universe may rely too heavily on backtesting.
For the analysis, Research Affiliates examined the performance of 125 U.S. smart beta indexes upon which ETFs are based, excluding sector indexes and indexes with less than one year of backtest or live return data.