We study the horizon dimension of cross-sectional return predictability using a model where characteristics contain both persistent and transitory components. We test the implications of this model for the average returns of popular characteristic-based trading strategies at short versus long horizons after portfolio formation. Our evidence supports the claim that the relative compensation for persistent and transitory components varies across characteristics, in both magnitude and sign. Benchmark factor models cannot explain the returns of portfolios sorted on characteristics where either the persistent or transitory component is dominant. Finally, we discuss implications for the long-term discount rates of firms.

Fiscal policy matters for bond risk premia. Empirically, government spending level and uncertainty predict bond excess returns, as well as term structure level and slope movements. Shocks to government spending level and uncertainty are also priced in the cross-section of bond and stock portfolios. Theoretically, government spending level shocks raise inflation when marginal utility is high, thus generating positive inflation risk premia (term structure level effect). Uncertainty shocks steepen the yield curve (slope effect), producing positive term premia. These effects are consistent with evidence from a structural vector autoregression. Asset pricing tests using model simulated data corroborate our empirical findings.

We classify asset pricing anomalies into those exacerbating mispricing (build-up anomalies) and those resolving it (resolution anomalies). We estimate the dynamics of price wedges for well-known anomaly portfolios and map them to firm-level mispricings. We find that several prominent anomalies like momentum and profitability further dislocate prices. Multi-factor models designed to eliminate one-month alphas still produce large price wedges. Our estimates yield a novel decomposition of Tobin’s q, revealing that q’s mispricing component has substantial explanatory power for firm investment. Overall, our results suggest that financial intermediaries chasing build-up anomalies negatively affect price efficiency and associated real capital allocation.

The component of the volatility of total factor productivity (TFP) that is orthogonal to the dividend price ratio is shown to have long-run predictive ability for excess market returns. This finding implies that TFP volatility should also predict real cash flows and/or real interest rates: it is found to mainly predict real cash flows through inflation. A model with endogenous growth, Epstein-Zin preferences and price rigidities reconciles both TFP volatility-driven long-run predictability and its real implications. Within the model, we justify the similar (to that of TFP volatility) predictive ability of a low-frequency notion of market volatility as well as the cross-sectional pricing of TFP volatility risk in alternative asset classes.

We represent risk factors as sums of orthogonal components capturing fluctuations with cycles of different length. The representation leads to novel spectral factor models in which systematic risk is allowed—without being forced—to vary across frequencies. Frequency-specific systematic risk is captured by a notion of spectral beta. We show that traditional factor models restrict the spectral betas to be constant across frequencies. The restriction can hide horizon-specific pricing effects that spectral factor models are designed to reveal. We illustrate how the methods may lead to economically meaningful dimensionality reduction in the factor space.