Partisan Bias in Professional Macroeconomic Forecasts
Benjamin S. Kay (Federal Reserve), et al.
June 2025
Using a novel dataset linking professional forecasters in the Wall Street Journal Economic Forecasting Survey to their political affiliations, we document a partisan bias in GDP growth forecasts. Republican-affiliated forecasters project 0.3-0.4 percentage points higher growth when Republicans hold the presidency, relative to Democratic-affiliated forecasters. Forecast accuracy shows a similar partisan pattern: Republican-affiliated forecasters are less accurate under Republican presidents, indicating that partisan optimism impairs predictive performance. This bias appears uniquely in GDP forecasts and does not extend to inflation, unemployment, or interest rates. We explain these findings with a model where forecasters combine noisy signals with politically-influenced priors: because GDP data are relatively more uncertain, priors carry more weight, letting ideology shape growth projections while leaving easier-to-forecast variables unaffected. Noisy information therefore amplifies, rather than substitutes for, heterogeneous political priors, implying that expectation models should account for both information rigidities and belief heterogeneity. Finally, we show that Republican forecasters become more optimistic when tax cuts are salient in public discourse, suggesting that partisan differences reflect divergent beliefs about the economic effects of fiscal policy.
Predicting Major Stock Price Declines with Fundamentals: A Machine Learning Approach
Richard Wang (St. John Fisher University)
June 2025
This paper develops machine learning models to forecast major medium-and long-term stock price declines using firm-level accounting fundamentals and stock returns. Major stock declines are defined as a loss of 50% within one year or 75% within two years. Using US stocks from 1970 to 2024, the study evaluates eight machine learning algorithms-Logistic Regression, Random Forest, XGBoost, AdaBoost, Multi-layer Perceptron (MLP), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Gaussian Naïve Bayes-using 42 financial ratios and stock returns as predictors. Model performance is assessed systematically, with XGBoost achieving the highest predictive accuracy, closely followed by Random Forest and AdaBoost. Further analysis shows that these leading algorithms can predict significantly negative stock returns up to five years in advance, with robust differences in future returns between predicted major-loss and non-major-loss groups across all investment horizons. This study contributes to the literature by demonstrating that advanced machine learning techniques, when combined with fundamentals data, can effectively forecast substantial stock price declines at the individual firm level over medium-and long-term horizons.
Predicting the Predictable: Decomposing and Forecasting Stock Returns in a Data-rich Environment
Tingting Cheng (Nankai University), et al.
June 2025
We propose a new method for forecasting stock market returns in a data-rich environment: the factor-augmented sum-of-the-parts (FA-SOP) approach. Rather than predicting returns directly, FA-SOP decomposes them into three components-dividend-price ratio, earnings growth, and price-earnings multiple growth (gm)-and models each separately. We emphasize that gm is a more promising target for forecasting due to its stronger connection with macroeconomic conditions and greater variability over time. FA-SOP forecasts gm using latent macroeconomic factors extracted from high-dimensional data, capturing the underlying state of the economy while avoiding overfitting. Applied to S&P 500 returns from 1960 to 2022, FA-SOP outperforms predictive regressions, factor-augmented regressions, and traditional decomposition approaches, yielding robust out-of-sample gains in both statistical and economic terms. Simulations based on a present-value model further show that FA-SOP’s advantage stems from its ability to track the true data-generating process more closely. Our results highlight the value of decomposing returns and focusing on components that are more predictably linked to economic fundamentals.
Factoring in the Low-Volatility Factor
Amar Soebhag (Erasmus University Rotterdam), et al.
June 2025
Low-volatility stocks have historically delivered higher risk-adjusted returns than their high-volatility peers. Despite extensive evidence and widespread adoption in the investment industry, the so-called low-volatility factor is absent from standard asset pricing models. This paradox is attributable to asymmetry in factor legs and real-life investment frictions. A low-volatility factor substantially improves performance of factor models once accounting for these dimensions in various in-sample and out-of-sample exercises, across different low-risk measures and across methodological choices. We advocate integrating the low-volatility factor into asset pricing models, accounting for the asymmetry and frictions.
Revisiting Factor Momentum: A One-month Lag Perspective
Mikael Rönkkö (University of Eastern Finland) and Joonas Holmi
July 2025
Recent studies have questioned the relevance of factor momentum by showing that its profitability is driven by a static tilt toward factors with positive historical means and that only a minority of individual factors exhibit significant momentum. This paper shows that replacing the traditional one-year formation window with a one-month window yields significant alpha after controlling for tilt toward positive-mean factors and doubles the number of factors with significant momentum from roughly 20% to 40%. Furthermore, we show that the positive autocorrelation between the one-month formation window and the subsequent month’s return is twice as high as in the traditional one-year formation window. In the modern era of electronic trading, this autocorrelation is nearly 14 times higher. Our findings highlight that the robustness and profitability of factor momentum strategies depend critically on the formation window length.
News Sentiment Embeddings for Stock Price Forecasting
Ayaan Qayyum (Rutgers University)
June 2025
This paper will discuss how headline data can be used to predict stock prices. The stock price in question is the SPDR S&P 500 ETF Trust, also known as SPY that tracks the performance of the largest 500 publicly traded corporations in the United States. A key focus is to use news headlines from the Wall Street Journal (WSJ) to predict the movement of stock prices on a daily timescale with OpenAI-based text embedding models used to create vector encodings of each headline with principal component analysis (PCA) to exact the key features. The challenge of this work is to capture the time-dependent and time-independent, nuanced impacts of news on stock prices while handling potential lag effects and market noise. Financial and economic data were collected to improve model performance; such sources include the U.S. Dollar Index (DXY) and Treasury Interest Yields. Over 390 machine-learning inference models were trained. The preliminary results show that headline data embeddings greatly benefit stock price prediction by at least 40% compared to training and optimizing a machine learning system without headline data embeddings.
VIX Decomposition, Tail Risk Premia, and the Cross-Section of Stock Returns
Victor Chow (West Virginia University), et al.
November 2024
This paper estimates equity tail risk premia (TRP) by decomposing the squared VIX into four components: tail risk premium (TRP), realized tail (RT), variance risk premium (VRP), and realized variance (RV). Empirically, about one-third of VIX variation is driven by TRP. TRP and RT, alongside VRP, are key predictors of future equity portfolio returns. Applying this framework to individual stocks, a tail risk factor (PMN) is constructed using portfolios sorted by TRP. Fama-MacBeth regressions show PMN enhances explanatory power beyond traditional factors, including market, size, and value, capturing crucial cross-sectional return variations, especially when returns deviate from mean-variance assumptions.
Learn To Use R For Portfolio Analysis
Quantitative Investment Portfolio Analytics In R:
An Introduction To R For Modeling Portfolio Risk and Return
By James Picerno
