Tabb’s latest statistics on US equity volumes are really raising some eyebrows, and not just because of the share of trade heading to the old bogey man the “dark pool”.
Nope. We’re talking about the share being gobbled up by banks and broker dealers direct. Something known as “internalisation”, and discussed at length on FT Alphaville before — something we’ve referred to as dark inventory.
The fact of the matter is that as much as 19 per cent of equity volume now doesn’t make any public or dark pool market at all. It is instantaneously “internalised”, a.k.a. matched by the broker dealer’s own flow before it gets a chance to be routed publicly.
And those are just the figures for equities. You’d be darn sure broker dealers are doing the same sort of thing with other asset classes too.
In March, 32.96% of US equities trading was executed away from registered exchanges.
While dark pools and internalized flow are even mysterious to the professional trader, 32.96% of equity trading being executed outside the exchange framework is enormous.
While we instinctively think of non-exchange orders as flow being diverted to dark pools, we estimate dark pools themselves make up only about 13% of US equities volume, and the balance and larger portion of this off exchange trading is executed via internalization.
Internalization is the practice of brokers matching orders internally on their trading desk – before orders are either sent to dark pools or exchanges.
In 2008, the internalized and dark pool figure was 15%. Today, nearly 33%. The million dollar question is why is such a large amount of US equity trading now making its way outside of traditional exchanges?
Now think about that last question folks. In our opinion it’s very important.
The obvious answer is that it’s all about preventing market impact. But really, says Tabb, it’s just as likely to be connected with reducing execution cost, matching rates and trading opportunity.
As we discussed in our post about JP Morgan’s chief investment office, this makes a lot of sense. The point is that banks are internalising ever more flow because it makes them more efficient at beating the competition in terms of funding. They receive a cost of funding, or rather financing, advantage. They also get to be more efficient as an institution. Rather than hedging every trade, they only have to hedge residual risk. And the manner in which they hedge that residual risk can become very creative.
But that’s not to say market impact doesn’t play a role. Even if it’s not the primary incentive for internalisation, the flash crash has proven that things can go awry. Because more and more flow is executed away from the public market, the public market becomes detached from reality. It becomes thin and unable to handle large trades.
Sometimes two separate markets begin to exist altogether, hence why the crossing phenomenon (when a stock’s bid/offers cross — the bid is higher than the offer) has become more common.
As we’ve noted before, one popular theory for why the flash crash happened is that flow became far too one-sided for internalisers to be able to handle the flow. Having breached internal limits, broker dealers that would usually internalise orders simply re-directed them to the public market — which being thin and illiquid, simply couldn’t handle the sudden added pressure (all one sided).
Which is why we do have to ask ourselves, what all this gobbling up of flow off-exchange is doing to our market structure.
And if you thought that was funky, consider the real risk. Cross-asset internalisation made possible by what banks consider phenomenal advances in correlation modelling. But we’ll bring you more on that later.
The non-role of internalisers during the flash crash – FT Alphaville
The power of the dark inventory – FT Alphaville
All eyes on broker-dealer internalisation – FT Alphaville
Stardate April 13, CIO sector, JP Morgan reporting VaR – FT Alphaville