Paper: arXiv 2512.20460
Authors: Ilias Aarab
Abstract
A growing empirical literature suggests that equity-premium predictability is state dependent, with much of the forecasting power concentrated around recessionary periods (Henkel et al., 2011; Dangl and Halling, 2012; Devpura et al., 2018). I study U.S. stock return predictability across economic regimes and document strong evidence of time-varying expected returns across both expansionary and contractionary states. I contribute in two ways. First, I introduce a state-switching predictive regression in which the market state is defined in real time using the slope of the yield curve. Relative to the standard one-state predictive regression, the state-switching specification increases both in-sample and out-of-sample performance for the set of popular predictors considered by Welch and Goyal (2008), improving the out-of-sample performance of most predictors in economically meaningful ways. Second, I propose a new aggregate predictor, the Aligned Economic Index, constructed via partial least squares (PLS). Under the state-switching model, the Aligned Economic Index exhibits statistically and economically significant predictive power in sample and out of sample, and it outperforms widely used benchmark predictors and alternative predictor-combination methods.
Complexity vs Empirical Score
- Math Complexity: 7.0/10
- Empirical Rigor: 7.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced econometric techniques like a state-switching predictive regression with an ex-ante yield curve indicator and uses Partial Least Squares (PLS) for dimensionality reduction, indicating substantial mathematical complexity. It is grounded in robust empirical design with a long-term dataset, real-time state definition, in-sample and out-of-sample testing, and comparison to established benchmarks, demonstrating high backtest-ready rigor.
Research Flowchart
flowchart TD A["Research Goal<br>Identify state-dependent equity-premium predictability<br>and improve forecasting accuracy"] --> B["Methodology<br>State-Switching Predictive Regression"] B --> C["Input Data<br>U.S. Stock Returns<br>Yield Curve Slope<br>Welch & Goyal Predictors"] C --> D["Process 1: State Definition<br>Real-time economic regimes<br>defined by yield curve slope"] C --> E["Process 2: Predictor Construction<br>Aligned Economic Index via<br>Partial Least Squares (PLS)"] D --> F["Computational Analysis<br>In-sample & Out-of-sample testing<br>State-switching vs. Standard models"] E --> F F --> G["Key Findings/Outcomes<br>1. State-switching improves forecasting performance<br>2. Aligned Economic Index shows significant predictive power<br>3. Outperforms standard benchmarks"]
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 a real‑time state‑switching predictive regression based on the yield‑curve slope and a new aggregate predictor called the Aligned Economic Index constructed via partial least squares.
Source sentence
I introduce a state‑switching predictive regression in which the market state is defined in real time using the slope of the yield curve.
Assumes 6
- The term spread (10‑year minus 3‑month Treasury rate) is a leading indicator of future economic conditions. market
Source sentence
To define market states in real time, I construct an ex‑ante indicator based on the slope of the yield curve (the term spread), computed as the 10‑year Treasury yield minus the 3‑month Treasury bill rate.
- The expectations hypothesis holds, linking the term spread to expected future short rates. other
Source sentence
Under the expectations hypothesis, the term spread reflects the difference between the current short rate and expected future short rates over a longer horizon.
- When the slope coefficients are equal (β1 = γ1) the model reduces to the conventional one‑state predictive regression. model
Source sentence
When β1 = γ1, the model collapses to the conventional one‑state predictive regression.
- Recursive out‑of‑sample forecasts use only information available up to the forecast origin to avoid look‑ahead bias. statistical
Source sentence
This recursive procedure avoids look‑ahead bias by construction.
Show 2 more
- Using a τ between 3 and 12 months does not change the main findings. model
Source sentence
Using τ between 3 and 12 months leaves the main findings unchanged.
- NBER recession dates are ex‑post and therefore not usable for real‑time forecasting. data
Source sentence
NBER recession dates can be retrieved from the Federal Reserve Bank of St. Louis (FRED). Because these dates are determined ex post, they are not directly usable for real‑time forecasting exercises.
Data
- Yield‑curve slope (10‑year Treasury yield minus 3‑month Treasury bill rate) macro public
Source sentence
To define market states in real time, I construct an ex‑ante indicator based on the slope of the yield curve (the term spread), computed as the 10‑year Treasury yield minus the 3‑month Treasury bill rate.
- 16 predictors from Welch and Goyal (2008) plus two additional series macro public
Source sentence
I aggregate information from the 16 predictors in a more parsimonious way by constructing an aligned economic…
- NBER recession dates macro public
Source sentence
NBER recession dates can be retrieved from the Federal Reserve Bank of St. Louis (FRED).
Costs
Not applicable
Finds 4
- The state‑switching specification improves both in‑sample and out‑of‑sample performance relative to the standard one‑state predictive regression. performance
Source sentence
Relative to the standard one‑state predictive regression, the state‑switching specification increases both in‑sample and out‑of‑sample performance for the set of popular predictors considered by Welch and Goyal (2008), improving the out‑of‑sample performance of most predictors in economically meaningful ways.
- The Aligned Economic Index (EPLS) shows statistically and economically significant predictive power in‑sample and out‑of‑sample, outperforming benchmark predictors and other combination methods. predictive-accuracy
Source sentence
Under the state‑switching model, the Aligned Economic Index exhibits statistically and economically significant predictive power in sample and out of sample, and it outperforms widely used benchmark predictors and alternative predictor‑combination methods.
- In the state‑switching model, EPLS and EFC both significantly predict the equity premium across expansions, recessions, up‑states and down‑states at the 5% level or stronger. performance
Source sentence
both EPLS and EFC can significantly, on the 5% level or stronger, predict the equity premium during expansions, recessions, up states and down states.
- EPCA significantly forecasts the equity premium during recessions, up‑states and down‑states, but not during expansions (R²₍oos₎≈0). predictive-accuracy
Source sentence
During expansions however both EPCA and the benchmark model predict, on average, the equity premium with the same capacity producing an
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 40 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 (35)
- The Stock Market’s Reaction to Unemployment News: Why Bad News Is Usually Good for Stocks (2005) — link
- Stock Prices, Earnings, and Expected Dividends (1988) — link
- Predicting Excess Stock Returns Out of Sample: Can Anything Beat the Historical Average? (2008) — link
- Approximately Normal Tests for Equal Predictive Accuracy in Nested Models (2007) — link
- Presidential Address: Discount Rates (2011) — link
- Why Does Return Predictability Concentrate in Bad Times? (2017) — link
- In Search of Attention (2011)
- Predictive Regressions with Time-Varying Coefficients (2012) — link
and 27 more
- Is Stock Return Predictability Time-Varying? (2018)
- Dividend Yields and Expected Stock Returns (1988)
- Business Conditions and Expected Returns on Stocks and Bonds (1989)
- The Evolution of Corporate Dividend Policy (1997)
- Economic Implications of Bull and Bear Regimes in UK Stock and Bond Returns (2005) — link
- On the Out-of-Sample Predictability of Stock Market Returns (2006) — link
- Time-Varying Short-Horizon Predictability (2011)
- Investor Sentiment Aligned: A Powerful Predictor of Stock Returns (2015)
- Manager Sentiment and Stock Returns (2019)
- On the Predictability of Stock Returns: An Asset-Allocation Perspective (1996) — link
- The Three-Pass Regression Filter: A New Approach to Forecasting Using Many Predictors (2015)
- Predicting Returns with Financial Ratios (2004)
- Forecasting Stock Returns: A Predictor-Constrained Approach (2020) — link
- Cross-Sectional Forecasts of the Equity Premium (2006) — link
- Out-of-Sample Equity Premium Prediction: Combination Forecasts and Links to the Real Economy (2010) — link
- Market Timing over the Business Cycle (2018) — link
- Predictive Regressions (1999)
- A Comprehensive Look at the Empirical Performance of Equity Premium Prediction (2008) — link
- The Yield Curve and Predicting Recessions (2006)
- Learning about Predictability: The Effects of Parameter Uncertainty on Dynamic Asset Allocation (2001)
- Regime shifts and stock return predictability (2018) — link
- Short interest and aggregate stock returns (2016) — link
- Do Retail Trades Move Markets? (2009) — link
- Forecasting Using Principal Components From a Large Number of Predictors (2002) — link
- Are more data always better for factor analysis? (2006) — link
- Partial least-squares regression: a tutorial (1986) — link
- Principal Component Analysis (2002) — link
0 of 35 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.