Empirical asset pricing via machine learning: evidence from the European stock market
Open accessThe DOI matches its Crossref record; 94 of the 94 references it cites were found in a registry or an index; 6 of the 42 sources its claims cite were checked for support.
Abstract
This paper evaluates the predictive performance of machine learning methods in forecasting European stock returns. Compared to a linear benchmark model, interactions and nonlinear effects help improve the predictive performance. But machine learning models must be adequately trained and tuned to overcome the high dimensionality problem and to avoid overfitting. Across all machine learning methods, the most important predictors are based on price trends and fundamental signals from valuation ratios. However, the models exhibit substantial variation in statistical predictive performance that translate into pronounced differences in economic profitability. The return and risk measures of long-only trading strategies indicate that machine learning models produce sizeable gains relative to our benchmark. Neural networks perform best, also after accounting for transaction costs. A classification-based portfolio formation, utilizing a support vector machine that avoids estimating stock-level expected returns, performs even better than the neural network architecture.
Overview
CoChat Summary
TL;DR
This paper evaluates machine learning methods for forecasting European stock returns against a linear Fama–MacBeth benchmark using 22 firm-level predictors, and finds that interactions and nonlinear effects improve predictive performance when models are adequately trained and tuned. Neural networks perform best among expected-return-based approaches, but a classification-based support vector machine that avoids estimating stock-level expected returns performs even better.
Key findings
- Machine learning models beat the linear benchmark on out-of-sample predictive R² (full sample): ols_full −0.52%, ols_pars 0.57%, elanet 1.21%, pcr 1.24%, pls 1.08%, rf 1.18%, gbrt 1.14%, nn_1 1.23%. Predictive slopes were closest to 1 for rf (1.03), gbrt (1.07), and nn_1 (0.98), versus 0.65 for ols_pars and 0.45 for ols_full.
- The most important predictors across models are based on recent price trends (e.g., short-term reversal, momentum) and fundamental valuation signals (e.g., earnings-to-price, book-to-market); tree-based models concentrate weight on few variables, while dimension-reducing models spread importance more widely.
- Long-only equal-weighted portfolios: nn_1 produced a monthly H–L spread of 3.26% (return ratio 1.04, annualized Sharpe 3.89, t-stat 13.96); the ols_pars portfolio reached a terminal value of €70.29 versus €107.18 for nn_1 before transaction costs, and €17.02 versus €26.01 after applying a conservative 57 bps monthly cost.
- The SVM classification approach slightly outperformed nn_1: accuracy 12.63% vs. 12.10% (random classifier 10.04%), and higher terminal values (€113.57 vs. €107.18 equal-weighted before costs; €27.57 vs. €26.01 after costs).
Methods
- Sample: all firms publicly listed in one of the nineteen Eurozone countries as of December 2020, monthly data from Thomson Reuters Datastream, January 1990–December 2020; average universe of 832 firms (range 56 to 1,012); average monthly excess return 0.51% (SD 4.55%); minimum €25 mill. market cap and at least 50 qualifying firms per month.
- Predictors: 22 firm-level characteristics from Drobetz et al. (2019), plus 44 country/SIC dummies (parsimonious set = 66 covariates) and, in the full set, two-way interactions and second/third-order polynomials (341 covariates); characteristics winsorized at 1%/99% and rank-mapped to (−1, +1).
- Models: OLS (parsimonious and full), elastic net, PCR, PLS, random forests, gradient boosted regression trees, feedforward neural networks (nn_1–nn_5), and a support vector machine for classification-based portfolio formation; rolling-window time-series cross-validation with 8 years training, 2 years validation, 1 year testing, annual re-estimation; evaluation via predictive slopes, predictive R², modified Diebold–Mariano tests, decile portfolio sorts, conditional CAPM/FF5M alphas, and long-only portfolio performance net of transaction costs.
Limitations
- The paper states that neural networks are computationally intensive and can be specified in innumerable architectures, so the authors pre-specify five models and treat nn_1–nn_5 as "a conservative lower bound" for neural network performance; they also note that deeper networks deteriorated quickly, possibly due to overfitting on a relatively small dataset.
- The paper notes that the DM test's weak error dependence assumption may be questionable due to potentially strong cross-sectional return dependence, and that pre-estimation variable selection based on full-sample importance metrics could introduce foresight bias.
- Evident from the text: the applied 57 bps monthly transaction cost is described as an overly conservative discount that pushes most value-weighted portfolios below the benchmark, and the authors caution that statistical and economic performance may be only weakly associated (citing Leitch and Tanner 1991).
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures










Verification
Claims the paper makes
5 Valid · 6 Not valid · 27 Not checked · 194 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Although the transaction cost models of Bollerslev et al. (2018) and Koijen et al. (2018) are quite conservative, we follow an even more cautious approach and deduct 57 bps from each portfolio's monthly excess returns.”
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Investment portfolio performance
What the sources say
- Supportedcites Bollerslev et al. (2018) source
Section 5.3 (p. 35) says Bollerslev et al. use the full bid-ask spread rather than the customary half-spread for their benchmark transaction-cost estimate, allowing for potentially higher costs from market impact. That supports describing their cost assumption as conservative; this verdict addresses only Bollerslev et al., not the separately cited Koijen et al. paper. (LLM, conf=0.86)
- Partly supportedcites Koijen et al. (2018) sourceExperimental — accuracy not yet measured
Section 4.2 calls Koijen et al.'s transaction-cost estimate conservative for options because the quoted spreads likely overstate realistic costs. The Introduction describes its estimates for other asset classes as realistic, so the blanket characterization of its transaction-cost model as conservative is broader than the paper supports. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This is our benchmark model, and it is able to explain a substantial percentage of the cross-sectional variation in European stock returns.”
§ Empirical asset pricing via machine learning: evidence from the European stock market > Introduction
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Although they use the FM regressions model that is re-estimated on a monthly basis, the ols_pars model provides nearly identical predictions in our sample-splitting and reestimation setting.”
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Ordinary least squares
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Across all machine learning methods, the most important predictors are based on price trends and fundamental signals from valuation ratios.”
- Also stated in: § Introduction · plain
“The most influential predictors are based on recent price trends, e.g., short-term reversal and stock momentum, and fundamental signals from valuation ratios, e.g., earnings-to-price and book-to-market ratios.”
§ Empirical asset pricing via machine learning: evidence from the European stock market > Abstract
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § Introduction · plain
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, the models exhibit substantial variation in statistical predictive performance that translate into pronounced differences in economic profitability.”
- Also stated in: § Introduction · plain
“Despite these commonalities, our forecast models exhibit differences in statistical predictive performance (as indicated by predictive slopes, predictive R 2 metrics, and Die- bold and Mariano (1995) test statistics), which also translate into marked differences in economic profitability.”
§ Empirical asset pricing via machine learning: evidence from the European stock market > Abstract
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § Introduction · plain
99 more claims
Create a free account to see all 232 claims and every source verdict.
- It may state two separate claims as one. 63
- Probably not a substantial claim. 62
- The quote only partly states this claim. 26
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Machine learning method: Support vector machines
“Smaller misclassification costs c allow us to ignore more of the misclassified support vectors, while continuing to fit optimal hyperplanes.”
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Appendix 1
“svm c {0.01, 0.05, 0.1, 0.5, 1} Vector influence γ {0.01, 0.05, 0.1, 0.5, 1} Misclassification costs”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Investment portfolio performance
“Most importantly, the numbers of the SVM-based model are slightly superior to that of the neural network for both weighting schemes.”
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Conclusion
“Before transaction costs Terminal value [€] 70.29 107.18 113.576.00 8.40 9.07 Excess return annualized [%] 22.45 24.93 25.28 8.90 10.67 11.07 Std annualized [%] 18.20 17.82 18.36 17.39 18.50 18.85 Maximum drawdown [%] 51.24 50.12 51.80 44.25 48.87 45.87 Sharpe ratio 1.24 1.41 1.38 0.51 0.58 0.59”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Machine learning method: Support vector machines
“Note that the aggregate accuracy can equivalently be computed as the number of true classifications (cases in the cells on the confusion matrix's left diagonal) divided by the number of total classifications (cases in any cell of the confusion matrix).”
§ Avg Std Definition Excess return [%] 0.51 4.55 Return in excess of the 3-month EURIBOR rate (scaled to the 1-month horizon) 1 size 13.22 0.48 Log market value of equity at the end of the prior month 2 bm -0.55 0.18 Log book value of equity minus log market value of equity at the end of the prior month 3 operatingprofitability 0.06 0.01 Operating profit divided by average total assets in the prior fiscal year 4 totalassetgrowth 0.07 0.04 Log total asset growth in the prior fiscal year 5 ret_2_12 0.09 0.20 Excess return from month -12 to month -2 6 issues_1_36 0.09 0.02 Log growth in split-adjusted shares outstanding from month -36 to month -1 7 accrualschange -0.04 0.01 Working capital accruals (based on Sloan 1996) in the prior fiscal year 8 roa 0.05 0.01 Income before extraordinary items divided by average total assets (return on assets) in the prior fiscal year 9 investment 1.02 0.08 Capital expenditures divided by net sales and revenues in the prior fiscal year 10 ret_1 0.01 0.05 Excess return of the prior month 11 ret_12_36 0.09 0.31 Excess return from month -36 to month -12 12 dy 0.03 0.01 Dividends per share over the prior 12 months divided by price at the end of the prior month 13 beta 0.83 0.10 Market beta, estimated from weekly returns from month -36 to month -1 14 vola 0.32 0.06 Standard deviation, estimated from weekly returns from month -12 to month -1 15 turnover 0.02 0.01 Average monthly turnover volume from month -12 to month -1 16 debttoprice 0.84 0.32 Short-term plus long-term debt in the prior fiscal year divided by market capitalization at the end of the prior month 17 salestoprice 2.16 0.60 Net sales and revenues in the prior fiscal year divided by market capitalization at the end of the prior month 18 cftoprice 0.14 0.04 Net income plus depreciation, depletion and amortization in the prior fiscal year divided by market capitalization at the end of the prior month 19 earningstoprice 0.04 0.02 Net income in the prior fiscal year divided by market capitalization at the end of the prior month 20 issues_1_12 0.02 0.01 Log growth in split-adjusted shares outstanding from month -12 to month -1 21 fcdispersion -2.05 0.18 Log standard deviation of I/B/E/S EPS forecasts minus log average absolute value of I/B/E/S EPS forecasts from month -12 to month -1 22 grossprofitability 0.30 0.03 Net sales and revenues minus costs of goods sold divided by average total assets in the prior fiscal year Sample size 832 Number of firms with non-missing observations for the excess return and each of the twenty-two characteristics > Machine learning method: Support vector machines
“Aggregate metrics are then calculated as the weighted average of each class' individual metric, taking the number of realized cases within each class as weights.”
Citations
Cites
Breiman, L. (2001) · Machine Learning · doi:10.1023/a:1010933404324 · Cited by 133,538 · cited 1×
“Random forests (rf) modify Breiman’s (2001) traditional bagging approach.”
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818
Common risk factors in the returns on stocks and bonds
Fama, E. F., & French, K. R. (1993) · Journal of Financial Economics · doi:10.1016/0304-405x(93)90023-5 · Cited by 25,754 · cited 2×
“Most prominently, Fama and French (1993) propose a multifactor alternative to the CAPM and find that it is better at explaining cross-sectional variation in expected returns than the CAPM.”
BackgroundCAPITAL ASSET PRICES: A THEORY OF MARKET EQUILIBRIUM UNDER CONDITIONS OF RISK*
Sharpe, W. F. (1964) · The Journal of Finance · doi:10.1111/j.1540-6261.1964.tb02865.x · Cited by 19,441 · cited 1×
“For example, the Capital Asset Pricing Model (CAPM) introduced by Sharpe (1964), Lintner (1965), and Mossin (1966) posits that, in equilibrium, a stock’s expected return is solely driven by its sensitivity to a systematic risk factor, i.e., the market risk.”
On Persistence in Mutual Fund Performance
Carhart, M. M. (1997) · The Journal of Finance · doi:10.1111/j.1540-6261.1997.tb03808.x · Cited by 17,283
Filders, R. (1992) · International Journal of Forecasting · doi:10.1016/0169-2070(92)90130-2 · Cited by 16,165
Risk, Return, and Equilibrium: Empirical Tests
Fama, E. F., & MacBeth, J. D. (1973) · Journal of Political Economy · doi:10.1086/260061 · Cited by 14,258 · cited 1×
“This approach typically runs Fama and MacBeth (1973) regressions (FM regressions).”
MethodBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S. (2024) · arXiv (Cornell University) · doi:10.57702/o9raffed · Cited by 14,214
Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency
JEGADEESH, N., & TITMAN, S. (1993) · The Journal of Finance · doi:10.1111/j.1540-6261.1993.tb04702.x · Cited by 11,584 · cited 2×
“explain differences in expected returns across stocks as a function of stock-level characteristics, such as size, value, and momentum (Fama and French, 1993; Jegadeesh and Titman, 1993; Lewellen, 2015).”
BackgroundEnsemble methods in machine learning
Dietterich, T. G. (2000) · Lecture Notes in Computer Science · doi:10.1007/3-540-45014-9_1 · Cited by 8,408
A five-factor asset pricing model
Fama, E. F., & French, K. R. (2015) · Journal of Financial Economics · doi:10.1016/j.jfineco.2014.10.010 · Cited by 7,852 · cited 2×
“We calculate alphas of the decile portfolios relative to the CAPM and the Fama and French (2015) five-factor model (FF5M), which we extend here by using Carhart’s (1997) momentum factor.”
MethodThe arbitrage theory of capital asset pricing
Ross, S. A. (1976) · Journal of Economic Theory · doi:10.1016/0022-0531(76)90046-6 · Cited by 7,009 · cited 1×
“Other examples include Ross’ (1976) asset pricing theory (APT) or Merton’s (1973) intertemporal CAPM (ICAPM).”
An Intertemporal Capital Asset Pricing Model
Merton, R. C. (1973) · Econometrica · doi:10.2307/1913811 · Cited by 6,508 · cited 1×
“For example, the Capital Asset Pricing Model (CAPM) introduced by Sharpe (1964), Lintner (1965), and Mossin (1966) posits that, in equilibrium, a stock’s expected return is solely driven by its sensitivity to a systematic risk factor, i.”
MethodBy Force of Habit: A Consumption‐Based Explanation of Aggregate Stock Market Behavior
Campbell, J. Y., & Cochrane, J. H. (1999) · Journal of Political Economy · doi:10.1086/250059 · Cited by 5,210 · cited 5×
“This translates into high persistence over time, which in turn suggests that any predictability in monthly excess returns is likely to extend to longer horizons (Campbell and Cochrane 1999; Cochrane 2008). Second, many of the characteristics are constructed similarly, i.e., issuance metrics, or incorporate similar fir…”
InfluentialDiebold, F., & Mariano, R. (1994) · National Bureau of Economic Research · doi:10.3386/t0169 · Cited by 5,185
Diebold, F. X., & Mariano, R. S. (1995) · Journal of Business & Economic Statistics · doi:10.1080/07350015.1995.10524599 · Cited by 5,130
Equilibrium in a Capital Asset Market
Mossin, J. (1966) · Econometrica · doi:10.2307/1910098 · Cited by 4,910 · cited 1×
“For example, the Capital Asset Pricing Model (CAPM) introduced by Sharpe (1964), Lintner (1965), and Mossin (1966) posits that, in equilibrium, a stock’s expected return is solely driven by its sensitivity to a systematic risk factor, i.e., the market risk.”
MethodHansen, L. K., & Salamon, P. (1990) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/34.58871 · Cited by 4,260 · cited 2×
“Fifth, we adopt the ensemble approach proposed by Hansen and Salamon (1990) and Dietterich (2000), and use ten independent seeds (s = 1, .”
MethodPopular Ensemble Methods: An Empirical Study
Opitz, D., & Maclin, R. (1999) · Journal of Artificial Intelligence Research · doi:10.1613/jair.614 · Cited by 2,953
Empirical Asset Pricing via Machine Learning
Gu, S., Kelly, B., & Xiu, D. (2020) · The Review of Financial Studies · doi:10.1093/rfs/hhaa009 · Cited by 2,545
SECURITY PRICES, RISK, AND MAXIMAL GAINS FROM DIVERSIFICATION*
Lintner, J. (1965) · The Journal of Finance · doi:10.1111/j.1540-6261.1965.tb02930.x · Cited by 2,452 · cited 1×
“For example, the Capital Asset Pricing Model (CAPM) introduced by Sharpe (1964), Lintner (1965), and Mossin (1966) posits that, in equilibrium, a stock’s expected return is solely driven by its sensitivity to a systematic risk factor, i.e., the market risk.”
MethodPresidential Address: Discount Rates
COCHRANE, J. H. (2011) · The Journal of Finance · doi:10.1111/j.1540-6261.2011.01671.x · Cited by 2,201 · cited 2×
“TS models typically conduct time series regressions of broad aggregate portfolio returns on a number of macroeconomic variables, including valuation ratios and interest rates, and on technical indicators (Cochrane 2011; Rapach and Zhou 2013).”
Background… and the Cross-Section of Expected Returns
Harvey, C. R., Liu, Y., & Zhu, H. (2016) · Review of Financial Studies · doi:10.1093/rfs/hhv059 · Cited by 2,119
Do Stock Prices Fully Reflect Information in Accruals and Cash Flows About Future Earnings?
Sloan, R. G. (1996) · The Accounting Review · doi:10.2308/tar-9608042309 · Cited by 1,956
Hypothesis Testing with Efficient Method of Moments Estimation
Newey, W. K., & West, K. D. (1987) · International Economic Review · doi:10.2307/2526578 · Cited by 1,844 · cited 4×
“17 The DM test statistic for comparing column model j and row model k is is the time series average of these differences, and ̂𝜎 d kj is the Newey and West (1987) standard error of d kj , t + 1 (with four lags to account for possible hetero-scedasticity and autocorrelation).”
InfluentialMeasuring Fund Strategy and Performance in Changing Economic Conditions
FERSON, W. E., & SCHADT, R. W. (1996) · The Journal of Finance · doi:10.1111/j.1540-6261.1996.tb02690.x · Cited by 1,759
Practical neural network recipes in C
Masters, T. D. (1993) · Cited by 1,636
FAMA, E. F., & FRENCH, K. R. (2008) · The Journal of Finance · doi:10.1111/j.1540-6261.2008.01371.x · Cited by 1,578 · cited 1×
“This is in line with Fama and French (2008), who find that the smallest 20% of stocks comprise only three percent of aggregate market capitalization.”
ResultDoes Academic Research Destroy Stock Return Predictability?
MCLEAN, R. D., & PONTIFF, J. (2016) · The Journal of Finance · doi:10.1111/jofi.12365 · Cited by 1,566 · cited 1×
“In line with Drobetz et al. (2019), we use a set of twenty-two firm-level predictors that are commonly used in the asset pricing literature (Chordia et al. 2017; Harvey et al. 2016; Lewellen 2015; McLean and Pontiff 2016).”
MethodThe Dog That Did Not Bark: A Defense of Return Predictability
Cochrane, J. H. (2008) · Review of Financial Studies · doi:10.1093/rfs/hhm046 · Cited by 1,227
Handbook of Economic Forecasting
(2013) · Handbook of Economic Forecasting · doi:10.1016/b978-0-444-53683-9.00023-2 · Cited by 1,210
Data-Snooping Biases in Tests of Financial Asset Pricing Models
Lo, A. W., & MacKinlay, A. C. (1990) · Review of Financial Studies · doi:10.1093/rfs/3.3.431 · Cited by 1,134 · cited 2×
“According to Gu et al. (2020), using large datasets mitigates sample selection or data snooping biases (Lo and MacKinlay 1990) and also can help avoid model overfitting by increasing the ratio of observation count to parameter count.”
Commonality in the determinants of expected stock returns
Haugen, R. A., & Baker, N. L. (1996) · Journal of Financial Economics · doi:10.1016/0304-405x(95)00868-f · Cited by 1,065 · cited 1×
“Haugen and Baker (1996), Lewellen (2015), and Drobetz et al. (2019) examine composite estimates of expected excess returns obtained from cross-sectional FM regressions.”
Hou, K., Xue, C., & Zhang, L. (2020) · The Review of Financial Studies · doi:10.1093/rfs/hhy131 · Cited by 989 · cited 1×
“1 Various studies, however, report violations of this assumption (see, e.g., Hou et al. 2020, for a comprehensive list of asset pricing anomalies) and study alternative asset pricing models.”
International tests of a five-factor asset pricing model
Fama, E. F., & French, K. R. (2017) · Journal of Financial Economics · doi:10.1016/j.jfineco.2016.11.004 · Cited by 927
Dittmar, R. F. (2002) · The Journal of Finance · doi:10.1111/1540-6261.00425 · Cited by 838 · cited 1×
“Following Dittmar (2002), we classify them into two subcategories.”
InfluentialCharacteristics are covariances: A unified model of risk and return
Kelly, B. T., Pruitt, S., & Su, Y. (2019) · Journal of Financial Economics · doi:10.1016/j.jfineco.2019.05.001 · Cited by 809
Kozak, S., Nagel, S., & Santosh, S. (2020) · Journal of Financial Economics · doi:10.1016/j.jfineco.2019.06.008 · Cited by 750
International Stock Return Predictability: What Is the Role of the United States?
RAPACH, D. E., STRAUSS, J. K., & ZHOU, G. (2013) · The Journal of Finance · doi:10.1111/jofi.12041 · Cited by 678 · cited 2×
“Freyberger et al. (2020) and Rapach et al. (2013) also use the lasso framework.”
BackgroundMethodA Taxonomy of Anomalies and Their Trading Costs
Novy-Marx, R., & Velikov, M. (2016) · Review of Financial Studies · doi:10.1093/rfs/hhv063 · Cited by 658
Dissecting Characteristics Nonparametrically
Freyberger, J., Neuhierl, A., & Weber, M. (2020) · The Review of Financial Studies · doi:10.1093/rfs/hhz123 · Cited by 619
Rapach, D., & Zhou, G. (2013) · Handbook of Economic Forecasting · doi:10.1016/b978-0-444-53683-9.00006-2 · Cited by 613
Shanken, J. (1990) · Journal of Econometrics · doi:10.1016/0304-4076(90)90095-b · Cited by 597 · cited 2×
“T h e s a m p l e i n c l ud e s a ll fi r m s t h a t w e r e pu b li c l y li s t e d i n on e o f t h e n i n e t ee n E u r o z on e c oun t r i e s i n a ny g i v e n m on t h du r i ng t h e J a nu a r y 1990 –”
BackgroundMethodAsset Pricing Models and Financial Market Anomalies
Avramov, D., & Chordia, T. (2006) · Review of Financial Studies · doi:10.1093/rfs/hhj025 · Cited by 523
The Cross Section of Expected REIT Returns
Chui, A. C. W., Titman, S., & Wei, K. C. J. (2003) · Real Estate Economics · doi:10.1111/1540-6229.00073 · Cited by 518
Autoencoder asset pricing models
Gu, S., Kelly, B., & Xiu, D. (2021) · Journal of Econometrics · doi:10.1016/j.jeconom.2020.07.009 · Cited by 481
(1991) · American Economic Review · Cited by 453
Forecasting stock indices: a comparison of classification and level estimation models
Leung, M. T., Daouk, H., & Chen, A.-S. (2000) · International Journal of Forecasting · doi:10.1016/s0169-2070(99)00048-5 · Cited by 428
The Cross-section of Expected Stock Returns
Lewellen, J. (2015) · Critical Finance Review · doi:10.1561/104.00000024 · Cited by 407
Empirical Asset Pricing via Machine Learning
Gu, S., Kelly, B., & Xiu, D. (2018) · National Bureau of Economic Research · doi:10.3386/w25398 · Cited by 378
Risk Everywhere: Modeling and Managing Volatility
Bollerslev, T., Hood, B., Huss, J., & Pedersen, L. H. (2018) · The Review of Financial Studies · doi:10.1093/rfs/hhy041 · Cited by 359
Asset Pricing with Omitted Factors
Giglio, S., & Xiu, D. (2021) · Journal of Political Economy · doi:10.1086/714090 · Cited by 355
International Tests of a Five-Factor Asset Pricing ModelPreprint
Fama, E. F., & French, K. R. (2015) · SSRN Electronic Journal · doi:10.2139/ssrn.2622782 · Cited by 312
Average Returns, B/M, and Share Issues
FAMA, E. F., & FRENCH, K. R. (2008) · The Journal of Finance · doi:10.1111/j.1540-6261.2008.01418.x · Cited by 310
The probability of backtest overfitting
Bailey, D. H., Borwein, J. M., de Prado, M. L., & Zhu, Q. J. (2016) · The Journal of Computational Finance · doi:10.21314/jcf.2016.322 · Cited by 305
OpitzDavid, & MaclinRichard (1999) · doi:10.5555/3013545.3013549 · Cited by 293
Koijen, R. S. J., Moskowitz, T. J., Pedersen, L. H., & Vrugt, E. B. (2018) · Journal of Financial Economics · doi:10.1016/j.jfineco.2017.11.002 · Cited by 286
No Arbitrage and Arbitrage Pricing: A New Approach
BANSAL, R., & VISWANATHAN, S. (1993) · The Journal of Finance · doi:10.1111/j.1540-6261.1993.tb04753.x · Cited by 263 · cited 1×
“Bansal et al. (1993), Bansal and Viswanathan (1993), Chapman (1997), Ditt-mar (2002), and Asgharian and Karlsson (2008), among others, explore various nonlinear pricing kernels and show that such specifications outperform linear counterparts.”
BackgroundBailey, D. H., Borwein, J. M., López de Prado, M., & Zhu, Q. J. (2014) · Notices of the American Mathematical Society · doi:10.1090/noti1105 · Cited by 221 · cited 1×
“If information from the out-of-sample period is used to fit the models, consciously or not (Schorfheide and Wolpin 2012), this might lead to overstated out-of-sample predictive performance (Bailey et al. 2014, 2017; Harvey and Liu 2014, 2015; Harvey et al. 2016).”
MethodAnomalies and False Rejections
Chordia, T., Goyal, A., & Saretto, A. (2017) · doi:10.2139/ssrn.3017677 · Cited by 213 · cited 4×
“In line with Drobetz et al. (2019), we use a set of twenty-two firm-level predictors that are commonly used in the asset pricing literature (Chordia et al. 2017; Harvey et al. 2016; Lewellen 2015; McLean and Pontiff 2016).”
InfluentialAnomalies and False Rejections
Chordia, T., Goyal, A., & Saretto, A. (2020) · The Review of Financial Studies · doi:10.1093/rfs/hhaa018 · Cited by 208
Machine Learning for Stock Selection
Rasekhschaffe, K. C., & Jones, R. C. (2019) · Financial Analysts Journal · doi:10.1080/0015198x.2019.1596678 · Cited by 204
Data-Snooping Biases in Tests of Financial Asset Pricing Models
Lo, A., & MacKinlay, A. C. (1989) · National Bureau of Economic Research · doi:10.3386/w3001 · Cited by 189
A New Approach to International Arbitrage Pricing
BANSAL, R., HSIEH, D. A., & VISWANATHAN, S. (1993) · The Journal of Finance · doi:10.1111/j.1540-6261.1993.tb05126.x · Cited by 182 · cited 1×
“Bansal et al. (1993), Bansal and Viswanathan (1993), Chapman (1997), Ditt-mar (2002), and Asgharian and Karlsson (2008), among others, explore various nonlinear pricing kernels and show that such specifications outperform linear counterparts.”
BackgroundWest, K. D. (2006) · Handbook of Economic Forecasting · doi:10.1016/s1574-0706(05)01003-7 · Cited by 182
International Stock Return Predictability: What is the Role of the United States?
Rapach, D., Strauss, J., & Zhou, G. (2012) · SSRN Electronic Journal · doi:10.2139/ssrn.1508484 · Cited by 165
Average Returns, B/M, and Share IssuesPreprint
Fama, E. F., & French, K. R. (2007) · SSRN Electronic Journal · doi:10.2139/ssrn.945546 · Cited by 162
Tree-Based Conditional Portfolio Sorts: The Relation between Past and Future Stock ReturnsPreprint
Moritz, B., & Zimmermann, T. (2016) · SSRN Electronic Journal · doi:10.2139/ssrn.2740751 · Cited by 156 · cited 1×
“Both Leung et al. (2021) and Moritz and Zimmermann (2016) apply ensemble versions that avoid overfitting to study how common equity factors and lagged stock returns predict future stock returns.”
BackgroundDissecting Characteristics Nonparametrically
Freyberger, J., Neuhierl, A., & Weber, M. (2017) · National Bureau of Economic Research · doi:10.3386/w23227 · Cited by 142
Approximating the Asset Pricing Kernel
Chapman, D. A. (1997) · The Journal of Finance · doi:10.1111/j.1540-6261.1997.tb01114.x · Cited by 126 · cited 1×
“Bansal et al. (1993), Bansal and Viswanathan (1993), Chapman (1997), Ditt-mar (2002), and Asgharian and Karlsson (2008), among others, explore various nonlinear pricing kernels and show that such specifications outperform linear counterparts.”
BackgroundHarvey, C. R., & Liu, Y. (2015) · The Journal of Portfolio Management · doi:10.3905/jpm.2015.42.1.013 · Cited by 114 · cited 1×
“If information from the out-of-sample period is used to fit the models, consciously or not (Schorfheide and Wolpin 2012), this might lead to overstated out-of-sample predictive performance (Bailey et al. 2014, 2017; Harvey and Liu 2014, 2015; Harvey et al. 2016).”
Deep Learning in Asset Pricing
Chen, L., Pelger, M., & Zhu, J. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3350138 · Cited by 107
Harvey, C. R., & Liu, Y. (2014) · The Journal of Portfolio Management · doi:10.3905/jpm.2014.40.5.108 · Cited by 101 · cited 1×
“If information from the out-of-sample period is used to fit the models, consciously or not (Schorfheide and Wolpin 2012), this might lead to overstated out-of-sample predictive performance (Bailey et al. 2014, 2017; Harvey and Liu 2014, 2015; Harvey et al. 2016).”
MethodThe Cross Section of Expected Stock ReturnsPreprint
Lewellen, J. W. (2014) · SSRN Electronic Journal · doi:10.2139/ssrn.2511246 · Cited by 81
Active Portfolio Management: Quantitative Theory and Applications
Grinold, R. C., & Kahn, R. N. (1995) · Medical Entomology and Zoology · Cited by 68
Deep Learning for Finance: Deep PortfoliosPreprint
Heaton, J. B., & Polson, N. (2016) · SSRN Electronic Journal · doi:10.2139/ssrn.2838013 · Cited by 68
West, K. D., Elliott, G., Granger, C., & Timmermann, A. (2006) · Cited by 66
On the Use of Holdout Samples for Model Selection
Schorfheide, F., & Wolpin, K. I. (2012) · American Economic Review · doi:10.1257/aer.102.3.477 · Cited by 63 · cited 2×
“If information from the out-of-sample period is used to fit the models, consciously or not (Schorfheide and Wolpin 2012), this might lead to overstated out-of-sample predictive performance (Bailey et al. 2014, 2017; Harvey and Liu 2014, 2015; Harvey et al. 2016).”
Autoencoder Asset Pricing ModelsPreprint
Gu, S., Kelly, B. T., & Xiu, D. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3335536 · Cited by 58
Characteristics Are Covariances: A Unified Model of Risk and Return
Kelly, B., Pruitt, S., & Su, Y. (2018) · National Bureau of Economic Research · doi:10.3386/w24540 · Cited by 48
Practical Neural Network Recipies in C++
(1993) · Elsevier eBooks · doi:10.1016/c2009-0-22399-3 · Cited by 40
Huerta, R., Elkan, C., & Corbacho, F. (2013) · Algorithmic Finance · doi:10.3233/af-13016 · Cited by 38 · cited 1×
“Similarly, Huerta et al. (2013) train an SVM to classify stocks into future out-and underperformers.”
Stock Selection via Nonlinear Multi-Factor Models
Levin, A. E. (1995) · neural information processing systems · Cited by 37
Approximating the Asset Pricing Kernel
Chapman, D. A. (1997) · The Journal of Finance · doi:10.2307/2329440 · Cited by 34
The Promises and Pitfalls of Machine Learning for Predicting Stock Returns
Leung, E., Lohre, H., Mischlich, D., Shea, Y., & Stroh, M. (2021) · The Journal of Financial Data Science · doi:10.3905/jfds.2021.1.062 · Cited by 33
The Promises and Pitfalls of Machine Learning for Predicting Stock Returns
Leung, E., Lohre, H., Mischlich, D., Shea, Y., & Stroh, M. (2021) · The Journal of Financial Data Science · doi:10.2139/ssrn.3546725 · Cited by 28
Deep Learning and the Cross-Section of Expected ReturnsPreprint
Messmer, M. (2017) · SSRN Electronic Journal · doi:10.2139/ssrn.3081555 · Cited by 27 · cited 2×
“For example, Levin (1995) and Messmer (2017) examine the predictive power of firm characteristics on subsequent stock returns, while Heaton et al. (2016) assess the applicability of deep learning approaches for multiple regression and classification problems in finance.”
BackgroundMethodTree-based machine learning approaches for equity market predictions
Wolff, D., & Neugebauer, U. (2019) · Journal of Asset Management · doi:10.1057/s41260-019-00125-5 · Cited by 18
Stock Returns and the Cross-Section of Characteristics: A Tree-Based ApproachPreprint
Coqueret, G., & Guida, T. (2018) · SSRN Electronic Journal · doi:10.2139/ssrn.3169773 · Cited by 14 · cited 1×
“Incorporating tree-based architectures, Coqueret and Guida (2018) use single regression trees to estimate the association of firm characteristics and subsequent stock returns.”
MethodEvaluating a non-linear asset pricing model on international data
Asgharian, H., & Karlsson, S. (2007) · International Review of Financial Analysis · doi:10.1016/j.irfa.2007.04.002 · Cited by 9 · cited 1×
“Bansal et al. (1993), Bansal and Viswanathan (1993), Chapman (1997), Ditt-mar (2002), and Asgharian and Karlsson (2008), among others, explore various nonlinear pricing kernels and show that such specifications outperform linear counterparts.”
Predictability and the cross section of expected returns: evidence from the European stock market
Drobetz, W., Haller, R., Jasperneite, C., & Otto, T. (2019) · Journal of Asset Management · doi:10.1057/s41260-019-00138-0 · Cited by 8 · cited 23×
“Haugen and Baker (1996), Lewellen (2015), and Drobetz et al. (2019) examine composite estimates of expected returns obtained from cross-sectional FM regressions.”
BackgroundMethodResultInfluentialSelf-citationPredictability and the Cross-Section of Expected Returns: Evidence from the European Stock MarketPreprint
Drobetz, W., Haller, R., Jasperneite, C., et al. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3436051 · Cited by 2
Self-citationTree-Based Machine Learning Approaches for Equity Market Predictions
Wolff, D., & Neugebauer, U. (2018) · doi:10.2139/ssrn.3265107
A New Approach to International Arbitrage Pricing
Bansal, R., Hsieh, D. A., & Viswanathan, S. (1993) · The Journal of Finance · doi:10.2307/2329065
Research works printed but not linked: 2
- Chordia, T., A. Goyal, and A. Saretto. 2017. p-Hacking: Evidence from Two Million Trading Strategies. Swiss Finance Institute Research Paper No. 17–37: 1–53.
- Cochrane, J.H. 2005. Asset Pricing. Revised. Princeton (NJ), US: Princeton University Press.
Cited by
Machine learning and the cross-section of emerging market stock returns
Hanauer, M. X., & Kalsbach, T. (2023) · Emerging Markets Review · doi:10.1016/j.ememar.2023.101022 · Cited by 76
Machine Learning Goes Global: Cross-Sectional Return Predictability in International Stock Markets
Cakici, N., Fieberg, C., Metko, D., & Zaremba, A. (2022) · Journal of Economic Dynamics and Control · doi:10.1016/j.jedc.2023.104725 · Cited by 63
Forecasting Market Crashes via Machine Learning: Evidence from European Stock Markets
Dichtl, H., Drobetz, W., & Otto, T. (2021) · Journal of Financial Stability · doi:10.1016/j.jfs.2022.101099 · Cited by 48
Self-citationStock return prediction with multiple measures using neural network models
Wang, C. E. (2024) · Financial Innovation · doi:10.1186/s40854-023-00608-w · Cited by 46
“Gu et al. (2020), Leippold et al. (2022), and Drobetz and Otto (2021) replace all missing values with the cross-sectional median of each firm-characteristic variable for that month.”
BackgroundMachine learning techniques for cross-sectional equity returns’ prediction
Fieberg, C., Metko, D., Poddig, T., & Loy, T. (2022) · OR Spectrum · doi:10.1007/s00291-022-00693-w · Cited by 47
“This is achieved by applying their research design to other stock markets (see, e.g., Tobek and Hronec 2021; Drobetz and Otto 2021; Leippold et al. 2022; Liu et al. 2022; Rubesam 2022; Lalwani and Meshram 2022, for international applications and applications to individual regions or countries) or…”
Method
Metadata
- Authors
- Drobetz, W., & Otto, T.
- Affiliations
- Universität Hamburg (DE)
- Corresponding author
- Otto, Tizian
- ORCID
- Drobetz, Wolfgang
- Venue
- Journal of Asset Management
- Published
- 2021-11-12
- Volume
- 22
- Issue
- 7
- Pages
- 507-538
- ISSN
- 1470-8272 (ISSN-L)
- 1479-179X
- OpenAlex
- W3213046477
- Cited by
- 61
- References
- 81
- Citations by year
- 2026: 18 · 2025: 7 · 2024: 10 · 2023: 15 · 2022: 8 · 2021: 3
- FWCI
- 7.46
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Journal of Asset Management (publishedVersion, cc-by)
- SSRN Electronic Journal (acceptedVersion)
- Econstor (Econstor) (submittedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- Topics
- Social Sciences › Economics, Econometrics and Finance › Finance › Financial Markets and Investment Strategies
- Social Sciences › Decision Sciences › Management Science and Operations Research › Stock Market Forecasting Methods
- Social Sciences › Decision Sciences › Management Science and Operations Research › Risk and Portfolio Optimization
- Keywords
- asset pricing, stock return prediction, machine learning, European stock markets, artificial neural networks, support vector machine, nonlinear effects, price trends, valuation ratios, overfitting, high dimensionality, transaction costs
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (prompt extract-v3).
A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.