Fabio Franceschini

This paper provides robust empirical evidence that shocks to aggregate Research and Development (R&D) have persistent effects on macroeconomic dynamics and represent a significant risk for investors, as predicted by the "long-run risk" literature. The analysis focuses on a single variable, "effective R&D", which captures the entire contribution of R&D to productivity growth, flexibly accounting for knowledge spillovers and product proliferation effects. Deviations of effective R&D from its equilibrium level can be empirically identified leveraging the error correction term in the cointegration relationship among R&D, total factor productivity, and the labor force. In US data, structural effective R&D shocks affect productivity and consumption growth rates beyond business cycle horizons and are associated with a significant risk premium in a cross section of stock and bond portfolios (around 2% annually), with cash-flow sensitivities proving a key determinant.

Empirical returns systematically depart from CAPM predictions, with deviations declining in asset betas. This paper decomposes this pattern into mismeasurement of risk and the risk premium, using a framework that accounts for leverage constraints and multiple risk factors. This spans and generalizes two previously separate explanations of the anomaly, showing how bid-ups for high-risk assets arise from funding tightness in the presence of risks beyond market exposure. Crucially, even with binding constraints, any factor model can be expressed as a single aggregate risk measure multiplying the expected market return. Funding tightness and exposure to omitted risks are then demonstrated to offset each other in explaining the beta-related departures, with their relative contributions quantifiable. GMM estimates show both channels are significant, with omitted risks accounting for a slightly larger share, and the spread generated by funding tightness at around 2% per year.

When the sustainability of assets is priced, its impact on discount rates depends not only on the asset's sustainability but intrinsically also on its risk profile. This has implications often overlooked in portfolios used to assess the sustainability premium or to hedge sustainability-related shocks. Specifically, the average returns of sustainability-sorted long-short portfolios are shaped by the risk profiles of their components, even when the portfolio is risk-neutral, which also affects the portfolio's sensitivity to shifts in sustainability concerns. Using Refinitiv ESG scores for US stocks, a weak sustainability premium is observed, whose significance differs importantly from that of the plain long-short portfolio returns.

We develop an econometric methodology to estimate risk premia conditional on states of the world defined by a threshold on an observable variable. The proposed methodology relies on asset pricing restrictions induced by the no-arbitrage assumption and employs established techniques to optimally select the threshold defining the states. The derivation of asymptotic results allows inference on the values of risk premia in different states, including testing whether they are zero in a given state. We illustrate the method using models with two and three states; the approach can be readily extended to a larger number of states. Applications will consider states defined by market returns, intermediaries' funding conditions, and expected volatility, speaking to different strands of the literature.

We propose a methodology to construct local indices of physical climate uncertainty, focusing on monthly average temperature series across 44 grid points in Italy. Uncertainty is defined, following an established econometric framework, as the predictable dispersion of temperature around its forecastable fluctuations, where forecasts account for trends, structural breaks, and more standard time dependencies. Trends in temperature levels are considered both as deterministic (allowing for structural breaks in the slope) and as stochastic (captured through 12-year moving averages), while the stationary components are predicted using a parsimonious dynamic factor model that exploits their factor structure. Conditional expectations of the squared forecasting errors are then obtained by estimating models with structural breaks in the mean and, for the deviations from the mean, univariate seasonal ARMA models. The results reveal significant geographical heterogeneity in average uncertainty and a marked increase in aggregate temperature uncertainty since the 1980s, particularly in coastal areas.