The riddle of differing fund flows and AUM
Looking at market statistics from different providers such as data vendors, associations, central banks, and others, one realizes that none of the providers state the same number for a fund’s flows or assets under management for a specific date.
Even though this may sound a bit odd, it is normal and the nature of the beast.
Since all data vendors, associations, and others have a different basis for the data they provide, flow numbers will be different from one provider to another. Data vendors calculate flows based on the funds in their database, while associations use the data on flows and assets under management they receive from their members. These data may include mandates and special-purpose vehicles such as pension funds, which are not mutual funds at all.
In contrast, central banks use the holdings data from their associated banks to evaluate the holdings of mutual funds. Statistics calculated for the same topic and for the same market can vary widely, since the underlying data can be totally different.
Good examples of the differences in databases for a market segment are the several reports available on the European exchange-traded fund (ETFs) market. While the Thomson Reuters Lipper report focuses only on ETFs, i.e., products that are funds and regulated as such, other reports focus on the whole market of exchange-traded products (ETPs), which means those reports also include structured notes such as exchange-traded commodities (ETCs).
Another factor that always leads to differences in numbers is the currency in which the report is calculated, since some providers use euros, while others use the U.S. dollar for the denomination of fund flows and assets under management.
But even if two providers of fund flows reports use the same data to calculate the flows for a given region, they may end up with totally different results. The employed methodology for the calculation of the flows might be different and would by definition lead to different results.
In addition, all results are dependent on correct and timely data input from the fund promoters, since any inaccurate numbers in the database impact the quality of the statistics. Even though a data vendor might have quality checks in place for the incoming data, it may not find all the corrupted data.
Even though quality checks do help to get the numbers right, some data may be missing and have to be estimated with an algorithm. This also explains why flows and assets under management data change over time, since it takes a while for all the fund promoters to deliver correct data.
All in all, it can be said that the most recent fund flows and assets under management statistics published shortly after the end of month should be seen more as a guide to evaluate market trends than as a scientific result.
Anybody who uses these kinds of statistics should make a decision about which statistics suit their needs best and then stay with those statistics. This does not mean that one should not question whether the displayed data are right, but one should realize that there always will be differences in flows data for any given month.
Detlef Glow, head of EMEA Research at Thomson Reuters Lipper.
The views expressed are the views of the author, not necessarily those of Thomson Reuters Lipper.