A study released by Deutsche Bank seeks to challenge the assumption that having more liquidity providers in an aggregator inevitably leads to better execution.
Aggregators are popular in the FX market, enabling trading firms to routinely put multiple liquidity providers in competition and then transact with the one offering the best price. Being able to consolidate liquidity, in the form of bid and offer prices and amounts, from various sources into a single, consolidated order book is particularly valuable in an OTC market with no centralised exchange.
“But in a market where the terms of trade are privately negotiated and the liquidity provided is bespoke to the trader, deciding on a suitable aggregation setup is not a trivial task,” according to the report, titled “Execution in an Aggregator”.
One of the key questions when setting up an aggregator is: how many liquidity providers the trading firm should include within it?
There is no consensus answer for this question, but Roel Oomen, co-head of spot FX at Deutsche Bank and the author of the report, says that some firms have operated on the assumption that having more liquidity options will lead to better prices when executing.
Once a firm has an aggregator in place, it becomes relatively easy to onboard liquidity providers, and the cost for doing so is fairly low. This, combined with the perception that access to more liquidity increases competition and leads to tighter spreads, is why some firms have continued adding more and more liquidity providers into their aggregators, says Oomen.
However, in his research, Oomen claims that the act of aggregation induces adverse selection whereby the liquidity provider that secures the trade suffers from the “winner’s curse”.
“The trader will observe tighter or even negative spreads in the aggregator, while the LP will find that the post-deal price movement is more likely to go in the trader’s favour than in his. To defend against this, the LP can enforce the last look trade acceptance criteria and adjust its tolerance to adverse selection. I show that it is this interplay that determines transaction costs, i.e., the number of liquidity providers, the nominal spread, and the trade acceptance criteria translate into an effective spread which represents the true cost of execution in an aggregator,” states Oomen.
“One of the main observations here is that the premise that adding more liquidity providers to your aggregator is always going to tighten the spread and increase liquidity access is flawed because it ignores the subsequent re-pricing of liquidity on behalf of the LPs. With more LPs competing for a client’s flow, the adverse selection effects or so-called “Winner’s curse” increases and this will be reflected back into the liquidity offered into the aggregator.”
He continues: “So the end result of adding more liquidity providers into your aggregator can in fact be a deterioration of liquidity access and increased transaction cost. It doesn’t need to be that way, but it’s certainly not a given that adding more liquidity providers into your aggregator is a positive regardless, it can be detrimental to the execution objectives of the client.”
In addition to the number of liquidity providers that trading firms put into an aggregator, Oomen says that it is also important to consider what type of liquidity providers are being included.
“The main distinction to draw here is the business model that the LP follows – i.e., are they an internaliser or an externaliser?” he notes.
Although, superficially, the liquidity of an externaliser and an internaliser can be hard to differentiate as they might be showing the same spread, that doesn’t mean that their liquidity is the same, according to Oomen.
“The execution costs, and a client’s ability to achieve their execution objectives, are in large part driven by the type of liquidity provider that they’re putting in the aggregator. While it is not possible to make a blanket statement that it is always better to trade with an internaliser than an externaliser or vice versa, an important finding is that when you mix the two together then you’re likely to end up with what is called in game theory a prisoner’s dilemma, and what that means in practice is that the internalisers may start to behave like externalisers. Clients need to be aware that adding one LP into their pool can change the behavior and liquidity of all the other LPs. This is quite a subtle effect that is easily overlooked, but can have first order impact on transaction costs and liquidity access,” he says.
In an aggregator where all of the liquidity providers are internalisers, the pricing should remain relatively stable in normal market conditions. But if an externaliser is added to the aggregator, then their trading model dictates that they will instantaneously hedge all of the risk that they have in the open market.
This creates market impact because it takes liquidity out of the open market, and this makers it harder for the internalisers to hold onto the risk that they are warehousing because the externaliser is causing the market to move in a manner that’s against their commercial interest. To prevent exposure to this market move, the internaliser may try to get out of the risk quicker than anyone else in that aggregation pool, and then there could be a scenario where everyone is trying to race out of risk in an attempt to prevent the market impact that the other competitors in the aggregator might cause by externalisation.
“This is what’s called a prisoner’s dilemma, because in the end everyone is in a worse situation, but no one individually is incentivised to change their actions,” says Oomen.