Shrinkage Estimation for Large Portfolios under Estimation Risk
-
Series
-
Speaker(s)Ming Yuan (Columbia University, United States)
-
FieldEconometrics, Data Science and Econometrics
-
LocationErasmus University Rotterdam, Campus Woudestein, ET-14
Rotterdam -
Date and time
October 09, 2025
12:00 - 13:00
Abstract
We study a general form of shrinkage estimates for the portfolio weights and apply them to maximize the out-of-sample Sharpe ratio of a large portfolio under parameter uncertainty. Our approach simultaneously shrinks the expected returns, covariance matrix, and the portfolio weights; and contains almost all existing shrinkage estimates as special cases. We provide analytical equations for finding the optimal shrinkage parameters and show that the resulting portfolio rule can converge to the corresponding optimal Sharpe ratio asymptotically. We apply our method to a number of data sets and find that it significantly improves existing approaches.