EMA Risk Databook

Macro trends in correlations and volatilities

The “databook” is a free-to-download publication in the form of an Excel workbook that highlights a selection of shifts in forecast correlations and volatilities. It is updated daily and looks back as far as 1 year.

It consists of tables for each of; equity countries, currencies, equity sectors, and the top 100 stocks by market capitalisation. The countries represented are a selection of major developed and emerging countries. The currencies are a selection of major developed and emerging countries, excluding those pegged to another currency. The sectors are global equal weighted portfolios of all stocks in the relevant sectors. The top 100 stocks by market capitalisation are calculated periodically and may not accurately represent the true top 100 on any given date.

The forecasts of correlation and volatility look ahead one month and are drawn from a GARCH extrapolation of the variances and covariances of a set of maximum-likelihood systematic market factors plus a stock specific residual variance scaling factor. The system is known as “FASTVaR” as it was developed to produce short-term VaR forecasts that take account of the fat-tails in security returns as required by the EU’s UCITS regulations.

The worksheet reports show summary information where the reported figures have been made comparable. For correlations an average correlation is calculated as the mean correlation an asset has with all other members of its group; for example, the average correlation for a given country index is the simple average of its correlations with each of the other country indices. For volatility the calculation is the ratio of an instrument’s forecast volatility to the equal weighted average of its group.

Having calculated the mean correlation and the relative volatility at the date of publication, these figures are then compared with the equivalent figures as at 1 week, 1 month, 3 months, 6 months and 1 year ago. The reports show the change in the mean correlation, the z-score (change as proportion of standard deviation), and percent change in relative volatility.

In principle, in relation to a globally diversified portfolio, where an asset shows a preponderance of increasing correlation and relative volatility, (shown in red in the report) it is a sale candidate, while a tendency for reduced average correlation and relative volatility (shown in green) may be a buy.

To sign up for access to download the report please go here.

If you have already signed up and are logged in please go here to download.