[We propose a decomposition for time series in components classified by levels of persistence. Employing this decomposition, we provide empirical evidence that consumption growth contains predictable components highly correlated with well-known proxies of consumption variability. These components generate a term-structure of sizable risk premia. At low frequencies we identify a component correlated with long-run productivity growth and commanding a yearly premium of approximately 2%. At high frequencies we identify a component with yearly half-life, which contributes to the equity premium for another 2%. Accounting for persistence heterogeneity, we obtain an estimate of the IES strictly above one and robust across subsamples.]

We show that machine learning methods, in particular, extreme trees and neural networks (NNs), provide strong statistical evidence in favor of bond return predictability. NN forecasts based on macroeconomic and yield information translate into economic gains that are larger than those obtained using yields alone. Interestingly, the nature of unspanned factors changes along the yield curve: stock- and labor-market-related variables are more relevant for short-term maturities, whereas output and income variables matter more for longer maturities. Finally, NN forecasts correlate with proxies for time-varying risk aversion and uncertainty, lending support to models featuring both channels.

We propose a decomposition for time series in components classified by levels of persistence. Employing this decomposition, we provide empirical evidence that consumption growth contains predictable components highly correlated with well-known proxies of consumption variability. These components generate a term-structure of sizable risk premia. At low frequencies we identify a component correlated with long-run productivity growth and commanding a yearly premium of approximately 2%. At high frequencies we identify a component with yearly half-life, which contributes to the equity premium for another 2%. Accounting for persistence heterogeneity, we obtain an estimate of the IES strictly above one and robust across subsamples.