Daily Event-Driven Index

Methodology

In line with our method to build the database, we collect data from a large number of sources. A difference to monthly data is that we need to process daily returns much more carefully, apply filters and aggregate differently.

• Data is collected daily. One of the many problems with daily data is that is not cleaned in the same manner and may contain noise.
• To remove noise, for instance, driven by dividend payments that are not properly incorporated into the return stream, we take the median return over many share classes. This removes some of the spikes, but not all of them.
• Moreover, we apply a statistical filter to remove outliers. Here, we control market movements that cause the filter to remove true market returns. An example of this is, for instance, the CHF intervention in 2015, which caused large losses for several currency managers.
• As we aggregate over share classes and most managers only show the “cheapest” share class in their official track-record, our returns tend to show a lower rate of return and potentially more volatility.
• As part of our final statistical test, we correlate the equivalent monthly returns, from daily compounded returns, with monthly returns streams that already exist in the monthly database.
• Entry and Exit fees are ignored.