Estimating future foreign asset returns in Rands via convolutions.
By Jaundré Scheltema, Client Solutions Team
Recent changes to Regulation 28 have had definite impacts on the investment landscape, with investors (both institutional and retail) now able to increase their offshore asset allocations from 25% to 30%. This opens up questions about exactly how much investors should allocate offshore and where they should invest their money.
Our aim is to provide a framework where investors can evaluate these questions while considering exchange rate views in their portfolio construction process. In doing so, we primarily consider the perspective of an investor who intends to invest offshore, but receives returns in Rands.
Assessing investment opportunities for foreign asset classes is often based on their performance, measured in local currency. This is usually done by multiplying the historic foreign returns with the relevant exchange rate, and then using these domestic returns as the basis for making future asset allocation decisions. This paper shows how using the traditional method as the basis for foreign asset allocation can be misleading.
There are additional considerations that investors may miss when making decisions about how much and where to invest offshore when the traditional method is used.
In contrast to this, the convolution method used for determining foreign asset performance in local currency allows investors to consider three distinct factors:
- The varying returns of the foreign asset.
- The currency fluctuations when converting to domestic currency.
- The dependency structure (e.g. correlation coefficient) between the foreign asset and the exchange rate in question.
Ultimately, this report aims to show that investors should make use of a convolution method to incorporate these three elements correctly when making robust decisions regarding the assessment of foreign opportunities. Our analysis will show that using the convolution method enables consistent views for an asset management firm, where the team responsible for working on foreign exchange rate distributions (e.g. for pricing derivatives or trading) can share their model with the asset allocation team, supporting a consistent house view when making asset allocation decisions.
Jaundré Scheltema talks us through the detail.
A detailed breakdown of the statistical and mathematical background to convolutions is provided here: