統計学輪講(第20回)

日時      2011年12月06日(火)    15時50分~16時40分
場所      経済学部新棟3階第3教室
講演者    井上 彰 (経済M1)
演題      Improved estimation of the covariance matrix of stock returns
with an application to portfolio selection (literature review)

概要
When solving a stock portfolio optimization problem, using a sample
variance-covariance matrix has been traditional. The difficulty, however,
is that we cannot always take the inverse of it. The estimator discussed in
the thesis is the weighted average of a sample variance covariance matrix
and that of the single-index model suggested by Sharpe (1963). The thesis
shows the calculation and asymptotic property of the optimal weight and
that the estimator above with the optimal weight reduces a risk of a
portfolio.