4-Factor Model for Overnight Returns

Abstract

We propose a 4-factor model for overnight returns and give explicit definitions of our 4 factors. Long horizon fundamental factors such as value and growth lack predictive power for overnight (or similar short horizon) returns and are not included. All 4 factors are constructed based on intraday price and volume data and are analogous to size (price), volatility, momentum and liquidity (volume). Historical regressions a la Fama and MacBeth (1973) suggest that our 4 factors have sizable serial t-statistic and appear to be relevant predictors for overnight returns. We check this by using our 4-factor model in an explicit intraday mean-reversion alpha.

Combining Alphas via Bounded Regression

“…Thus, as expected, adding the liquidity bounds has the diversification effect on the portfolios, so the Sharpe ratios are substantially improved -as usual, at the expense of (slightly) lowering paper ROC and CPS. Note that, even with tight liquidity bounds, the 4 style factors prc, mom, hlv and vol of [60] add value, further validating the 4-factor model of [60].…”

“…To illustrate the use of the algorithm, we have employed it to construct portfolios for intraday mean-reversion alphas with the loadings matrix Λ iA in the following 5 incarnations: i) intercept only (so K = 1); ii) BICS (Bloomberg Industry Classification System) sectors; iii) BICS industries; iv) BICS sub-industries; and v) the 4 style factors prc, mom, hlv and vol of [60] plus BICS sub-industries. The regression weights are the inverse sample variances: z i = 1/C ii (see below).…”

“…P&L graphs for the intraday alphas discussed in Section 2.5.3, with a summary in Table2. Bottom-to-top-performing: i) Regression over intercept only, ii) regression over BICS sectors, iii) regression over BICS industries, iv) regression over BICS sub-industries, and v) regression over 4 style factors prc, mom, hlv and vol of[60] plus BICS sub-industries. The investment level is $10M long plus $10M short.…”

ETF Risk Models

“…12 The number of relevant style factors is even fewer in short-horizon risk models for use, e.g., in short-horizon quant trading applications [Kakushadze, 2015a], where it is 4 (or even fewer).…”

ETF Risk Models

“…9 The (often misconstrued) "shrinkage" method[Ledoit and Wolf, 2004] is nothing but a special type of statistical risk models; see[Kakushadze, 2016],[Kakushadze and Yu, 2017] for details. 10 E.g., BICS (Bloomberg Industry Classification System), GICS (Global Industry Classification Standard), ICB (Industry Classification Benchmark), SIC (Standard Industrial Classification), etc.11 The number of relevant style factors is even fewer in short-horizon risk models for use, e.g., in short-horizon quant trading applications[Kakushadze, 2015a], where it is 4 (or even fewer). 12 More precisely, this holds in the regression of 𝛹 𝛼 over 𝑥 𝛼 , demeaned 𝑦 𝛼 , and demeaned 𝑧 𝛼 .…”