Diego Ronchetti (University of Groningen) on "Consistent estimation of optimized functions for the analysis of portfolio strategies"

I propose a nonparametric econometric method for the consistent estimation of measures of unhedgeable risks and the values for the structural parameters that cap them at reference levels. Each measure is a solution of a dynamic stochastic optimization problem. The estimation method is a kernel-based estimation technique to match (non-linear) functionals of the probability density function of the state variables with reference levels. I illustrate the method in the estimation of the minimal endowments and costs for trade execution that bound the degree of incompleteness of a financial market featuring unspanned stochastic volatility of the asset returns. I describe the asymptotic properties of the estimators for a large time series, and how to study their finite sample properties in a Markovian setting through a resampling method.