Errikos Melissinos

Real Term Premia Explained by Stochastic Volatility

Real term premia of risk-free bonds exhibit substantial variation over time. This paper shows that stochastic volatility can give rise to such behaviour, with the required amount of volatility in consumption or wealth being similar to the amount of volatility in the stock market. This explanation is consistent with both the intermediary asset pricing approach and with supply and demand driven explanations, in both of which a small group of investors play a special role in the pricing of securities. Furthermore, under standard preferences and without time-varying risk aversion such stochastic volatility is also necessary to explain real term premia. The paper analyses models with both time-separable and recursive preferences. The latter model variations are solved via a novel perturbation method with respect to the parameter for intertemporal elasticity of substitution.

A Perturbation Solution Method for Models with Recursive Utility

I illustrate a novel method for pricing assets within recursive utility models in continuous time, that has first been used in Melissinos (2023). My method builds on the analytic solution of Tsai and Wachter (2018). While their solution is valid for a value of the intertemporal elasticity of substitution equal to 1, I provide the full perturbation series in terms of the IES, which gives rise to a global perturbation approximation in terms of the state variable. This allows the pricing of assets for a much larger range of values for the IES, which are economically meaningful. I comment on the convergence properties of the perturbation series, and I show that the method provides a straightforward and reliable approach to asset pricing. I employ my method to derive prices of long-term bonds, the price consumption ratio and the instantaneous return of the consumption perpetuity.

Measuring On-the-job Learning Rates in Multidimensional Skills

In this paper, we introduce a framework with multidimensional skills, in which we estimate how fast skills accumulate due to on-the-job experience. We model an individual’s wage as a weighted sum of her productivities in different skills. We call this skill-specific productivity expertise. Since expertise is not directly observable, we proxy this variable with skill-specific experience, which depends on the years of labor market experience across different occupations and the importance of the corresponding skill in those occupations. We compute skill-specific experience using the data on occupational skill requirements from O*NET. We then estimate the wage equation using skill-specific experience to evaluate the speed of expertise accumulation (learning rate) in different skills. We find that expertise in different skills grows with skill-specific experience and that different skills exhibit different learning rates.

A Monte Carlo Solution Method for Heterogeneous-agent Macro-finance Models