The role of correlations in low-volatility portfolios

Robeco pioneered low-volatility investing. In theory stocks with low risk should also give low returns, but in practice the relation between risk and return turns out to be flat, or even inverted. Based on this finding we introduced our “Conservative Equities” funds which specifically target low-risk stocks.

An important question for low-risk investors is whether they should only buy stocks which have a low risk when viewed in isolation, or whether they should also include stocks which, on a stand-alone basis, have higher risk, but which help to reduce risk at the portfolio level because of their low correlations with other stocks. In this internship we want to take a deep dive into the role of correlations for low-volatility investors.

Available projects

Below are some examples of specific questions we would like to investigate:

How much weight should be given to correlations when doing minimum variance optimization? Shrinkage is a commonly applied technique here.
The beta of a stock is the product of its correlation with the market and the ratio of its volatility versus that of the market. As both components involve different degrees of estimation error, how should each of these be adjusted in order to get the best estimate of future stock betas?
Global investors need to deal with multiple currencies. Should low-risk investors evaluate stocks on their risk in local currency, or should they also take the correlation of foreign currencies with the base currency into account?
Low-volatility stocks tend to be more sensitive to interest rate changes than conventional stocks. But are some low-volatility stocks more or less sensitive to interest rates than others? And can this be used to make a better low-volatility strategy?

Literature: Blitz & van Vliet, “The Volatility Effect: Lower Risk Without Lower Return”, Journal of Portfolio Management, 2007.