Title: BUMVU estimators
Authors: Aleksey Kolokolov - Manchester Business School (United Kingdom) [presenting]
Roberto Reno - University of Verona (Italy)
Patrick Zoi - Bank of Italy (Italy)
Abstract: The theory of Uniformly Minimum Variance Unbiased Estimators is extended to the special class of Block estimators. We provide necessary and sufficient conditions for a block estimator to have uniformly minimum variance. We show the relevance of this theory uncovering new results in two classical statistical problems: estimation of the error term variance in homoskedastic nonparametric regressions, and estimation of volatility functionals of semimartingales observed at a high frequency.