Paper: arXiv 2405.10920
Authors: Shuxin Guo, Qiang Liu
Abstract
We study the data-generating processes for factors expressed in return differences, which the literature on time-series asset pricing seems to have overlooked. For the factors’ data-generating processes or long-short zero-cost portfolios, a meaningful definition of returns is impossible; further, the compounded market factor (MF) significantly underestimates the return difference between the market and the risk-free rate compounded separately. Surprisingly, if MF were treated coercively as periodic-rebalancing long-short (i.e., the same as size and value), Fama-French three-factor (FF3) would be economically unattractive for lacking compounding and irrelevant for suffering from the small “size of an effect.” Otherwise, FF3 might be misspecified if MF were buy-and-hold long-short. Finally, we show that OLS with net returns for single-index models leads to inflated alphas, exaggerated t-values, and overestimated Sharpe ratios (SR); worse, net returns may lead to pathological alphas and SRs. We propose defining factors (and SRs) with non-difference compound returns.
Complexity vs Empirical Score
- Math Complexity: 7.5/10
- Empirical Rigor: 2.0/10
- Quadrant: Lab Rats — high math complexity, limited empirical testing
Why this score: The paper is dense with advanced mathematical concepts like geometric Brownian motions, compounding, and log returns, but it lacks empirical backtesting, datasets, or statistical metrics, focusing instead on theoretical critiques of existing models.
Research Flowchart
flowchart TD
A["Research Goal<br>Data-generating process for factors?"] --> B{"Methodology<br>Compare Factor Definitions"}
B --> C["Define Factor: Net Return Difference<br>Used in FF3 & Size/Value"]
B --> D["Define Factor: Compounded Return<br>Used for Market Factor"]
C --> E["Computational Process<br>Standard OLS Analysis"]
D --> F["Computational Process<br>Compounding Separately<br>vs. Jointly"]
E --> G["Outcome A: Net Returns<br>Inflated Alphas & SRs<br>Pathological Results"]
F --> H["Outcome B: Compounded Returns<br>Market Factor Underestimates<br>Return Difference"]
G --> I["Key Finding<br>Factors require Non-Difference<br>Compound Returns"]
H --> I
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes defining factors and Sharpe ratios with non‑difference compound (log) returns instead of excess net returns.
Source sentence
We propose defining factors (and SRs) with non-difference compound returns.
Assumes 3
- Cash can earn (or must be financed with) risk‑free interest rates with compounding. market
Source sentence
Critically, let’s do not forget that the cash can earn (or has to be financed with) risk-free interest rates with compounding for simplicity.9 Thus, the terminal value: 𝑆0 + 𝜇̅𝑆0 ∑∏(1 + 𝑟𝑠) 𝑛−1 𝑠=𝑡 𝑛−1 𝑡=1 + 𝜇̅𝑆0 (3) is not equal to 𝑆𝑛, where 𝑟𝑡 is the one-period risk-free interest rate at the beginning of period t.
- Using net returns in OLS leads to inflated alphas, t‑values and Sharpe ratios. model
Source sentence
Finally, we show that OLS with net returns for single-index models leads to inflated alphas, exaggerated t-values, and overestimated Sharpe ratios (SR); worse, net returns may lead to pathological alphas and SRs.
- The average behavior of a buy‑and‑hold portfolio is correctly measured by log returns. model
Source sentence
Based on these findings, we suggest that both the market factor and Sharpe ratio should be defined with (buy-and-hold) log returns instead of excess net returns.
Data
- S&P 500 index monthly adjusted closes market-prices public
Source sentence
The monthly adjusted closes from January 1, 1960 to January 1, 2020 are downloaded from Yahoo!Finance.
Costs
Not applicable
Finds 4
- Compounded net returns overestimate the index’s terminal value by 89%, an egregious error. performance
Source sentence
Unfortunately, if compounded, the mean of net returns overestimates the index’s terminal value by 89%, an egregious error that cannot be neglected.
- OLS with net returns inflates alphas, t‑values and Sharpe ratios. statistical
Source sentence
Finally, we show that OLS with net returns for single-index models leads to inflated alphas, exaggerated t-values, and overestimated Sharpe ratios (SR); worse, net returns may lead to pathological alphas and SRs. Through numerical analyses, we further show that using net returns in OLS will result in inflated t-values, exaggerated alphas, and 6 overestimated Sharpe ratios.
- The arithmetic mean, when compounded, overestimates the terminal value by 89% and is unsuitable for measuring average return of a buy‑and‑hold portfolio. theoretical
Source sentence
In contrast, the arithmetic mean, if compounded, overestimates the terminal value by 89% and is unsatisfactory.
- Log returns provide correct average behavior for both buy‑and‑hold and long‑short portfolios. theoretical
Source sentence
Based on these findings, we suggest that both the market factor and Sharpe ratio should be defined with (buy-and-hold) log returns instead of excess net returns. As discussed previously, the average behavior of a buy-and-hold portfolio can only be measured correctly by log returns.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 14 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (9)
- “Options, Futures, and Other Derivatives” (2016) — link
- Common risk factors in the returns on stocks and bonds (1993) — link
- Modern Portfolio Theory and Investment Analysis. (1982) — link
- Volatility‐Managed Portfolios (2017) — link
- Size and value in China (2019) — link
- Editor’s Choice Digesting Anomalies: An Investment Approach — link
- … and the Cross-Section of Expected Returns (2015) — link
- A five-factor asset pricing model (2014) — link
and 1 more
- Continuous-Time Finance. (1991) — link
0 of 9 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.