Why and How Systematic Strategies Decay

Mispricing and Risk Premia in Currency Markets

“…The first approach examines predictor profits in periods before and after the dissemination of research publicizing the trading strategies. If profits reflect mispricing and publication leads to investors learning about strategies and trading on them to exploit mispricing, currency excess return predictability should decline post publication (Falck et al, 2021;McLean and Pontiff, 2016;Chordia et al, 2014;Schwert, 2003;Cochrane, 1999). Consistent with mispricing as a source of predictability, we show that risk-adjusted payoffs associated with currency strategies significantly decrease after the academic research has been published and that post-publication declines are greater for strategies with economically or statistically larger in-sample profits and smaller limits to arbitrage.…”

“…15 The publication effect and the interaction terms involving in-sample profits are always negative and significant for profits gross and net of transaction costs using alternative samples with different sets of currencies (Table A9). 16 Tests using a combined proxy as in Falck et al (2021) also show no evidence of overfitting.…”

“…In contrast, trading profits should not change after publication if they reflect compensation for risk, conditional on no fundamental change in the risk-return trade-off or pricing of risk (McLean and Pontiff, 2016;Schwert, 2003;Chordia et al, 2014;Cochrane, 1999). If currency excess return predictability originates solely from in-sample statistical bias or data mining, predictability should not exist in the out-of-sample period (Falck et al, 2021;McLean and Pontiff, 2016;Schwert, 2003;Cochrane, 1999;Fama 1998Fama , 1991. 12 Profits of individual predictors are generally positive and significant over the full sample period before accounting for transaction costs as documented in the literature, while net profits are naturally smaller (Table A6).…”

“…In the same vein, our approach is non-discretionary with regards to the sample of currency predictors and the inclusion of potentially riskbased predictors. In line with prior asset pricing literature, the focus of our paper is on the crosssection of predictors similar to Falck et al (2021), Engelberg et al (2020Engelberg et al ( , 2018, Guo et al (2020), Calluzzo et al (2019), McLean and Pontiff (2016), and Chordia et al (2014).…”

Discontinued Positive Feedback Trading and the Decline of Momentum Profitability

“…In the second part of the analysis, we zoom in on a narrow window around the June 3 Prior work has identified many other mechanisms that may cause factor returns to decline: changes in liquidity (Khandani and Lo, 2011;Chordia, Subrahmanyam, and Tong, 2014;Lee and Ogden, 2015), and increased arbitrage activity (Marquering, Nisser, and Valla, 2006;Green, Hand, and Soliman, 2011;Hanson and Sunderam, 2013;McLean and Pontiff, 2016;Calluzzo, Moneta, and Topaloglu, 2019;Cho, 2020). Several studies propose that some factors may result from possible data-mining or overfitting (Harvey, Liu, and Zhu, 2016;Harvey, 2017;Hou, Xue, and Zhang, 2020;Huang, Song, and Xiang, 2020b;Falck, Rej, and Thesmar, 2021).…”