Market Concentration Cash vs. Future
Introduction - HHI
As of today, there are 144 trading participants on Xetra and 247 on Eurex. Not all of them are equally active, obviously. Some trade very little, some only trade a subset of the listed instruments, others have a disproportionate market share across the entire universe.
The Herfindahl-Hirschman Index (HHI) is a popular metric for measuring market concentration. In short, a value of 0 corresponds to an infinite number of comparable participants whereas a value of 1 would correspond to a single participant with 100% market share.
Fig. 1 below shows the time series of the daily HHI for trading in the DAX future (FDAX) on Eurex and the DAX cash equity constituents on Xetra. The top panel only considers the aggressive volume, the center panel only the liquidity provider, and the bottom panel only looks at auctions (only relevant for Xetra here).
Observations
Xetra
The HHI is surprisingly stable for the cash equities over the entire 6-year period and on a day-to-day basis. Similarly, there is little difference between aggressive and passive trading. The auctions show a somewhat higher degree of market concentration. Likely, this is because a lot of the "usual suspects" are not active in the auctions. All in all, on an aggregate (over the individual stocks) level, the data is pretty boring - in a good way. The most noteworthy occurrence is the short-lived modest peak in March 2020 when Covid-volatility was highest.
领英推荐
Eurex
The data for Eurex is much more "feature-rich". The aggressive HHI has phases where the value is comparable to that of the underlying cash equities. But there are also extended periods where it is very volatile with peaks reaching 4X the "quite value". The highest market concentration is seen at the height of the Covid-induced volatility. There could be a few reasons for this. First, participants suspending trading activities because market state is not represented by existing quantitative models. Secondly, reduced liquidity means that fewer aggressors can get filled. Interestingly, the passive HHI reaches a low point at this time. My guess is that some previously dominant liquidity providers pulled out of the market or significantly scaled back their activities. So, while the total number of participants has decreased, the resulting more homogenous distribution of market shares lowered the HHI.
The aggressive HHI does not exhibit an overall trend over time. In fact, its current value is comparable to its value in late 2017. The passive HHI, on the other hand, shows a persistent uptrend, doubling over the same period. Not sure what to make of this. If the latency race was pushing the market towards a winner-takes-all mode, then one would expect to see the same trend (and more pronounced) for aggressive trading.
The FDAX had a tick size change in late December 2020 (0.5 to 1). This probably explains the simultaneous dip of the aggressive and passive HHI starting around this time.
Technical note: The Xetra & Eurex HHI files provide even more detail. They include the HHI split not only by aggressive/passive/auction but also by buy/sell and participant category; plus, the data is on an instrument-by-instrument level. On the other hand, they do not include figures for the DAX40 in aggregate, for example. So, the data shown in blue in Fig. 1 is based on proprietary data.
PS (an anecdote): Many years ago a colleague noted some weird observation he made in the data. I quickly found a narrative to explain his finding. Just as I had finished, he noted that he had misspoken and had meant the opposite. Awkward. Worse (and containing a deeper message): I immediately came up with a narrative explaining the revised observation.
Why am I telling this? Because for the Xetra/Eurex HHI dichotomy could have easily come up with a narrative justifying any outcome (Xetra HHI ? Eurex HHI).