“Exchange-traded funds (ETFs) replicate their base index as accurately as possible. It seems logical to me that this can lead to small deviations. However, the differences between ETFs on the same index are sometimes considerable. How is that possible?”
Until a few years ago, exchange-traded index funds (ETFs) simply replicated an index by physically buying the underlying securities, for example all 20 stocks in the SMI with the corresponding weighting. Thus the deviation from the index, the so-called tracking error, was minimal and largely identical for all ETFs on the same index. Small differences arose from the different handling of the dividend, for example from the distribution frequency and of course from whether they reinvest (reinvest) or distribute the dividend. This has since changed. Many ETFs no longer buy all securities, but only a selection, which should then be representative. In the case of small indices – for example exotic stock exchanges – this can make sense because small second-line stocks often cannot be bought at all or would lead to massive price fluctuations. Or they just replicate a similar but not identical index. Or – and this is now much more frequently the case – they buy only partially physically and underlay the rest with a derivative. Or they have built in a lever or appear to be contrary (Short ETF). This can result in considerable deviations in ETFs on one and the same index.
It is therefore no longer sufficient to buy any ETF on an index without looking at it. As with derivatives, you have to look closely at how the ETF is constructed. In our opinion, ETFs that are constructed as simply as possible (all values physically) on indices that are as liquid as possible are still the best for “normal investors”. Everything else requires a lot of effort and expertise and brings with it a mostly unwanted speculative component. You can find out how an ETF is constructed from the factsheet that is published on the website of the relevant trading venue.