Rob Daly: OTC Algo Trading? Not Quite Yet

Ever since politicians and regulators moved to put trading of certain over-the-counter (OTC) derivatives contracts on electronic markets by way of swap execution facilities (SEFs), industry participants and observers have been asking: “When will we see algorithmic trading in these markets?” I understand this level of interest since any coverage of other electronically traded markets—equities for example—cannot go without the required editorial about the evils of high-frequency and algorithmic trading. However, the industry and the public are putting the cart before the horse here.

No Rules
First, the rules governing how SEFs operate do not yet exist and probably will not for the next 18 months or so. The US Commodity Futures and Trading Commission (CFTC) and the US Securities and Exchange Commission (SEC) kicked the Dodd–Frank can down the road by moving many of the July 2011 deadlines to the end of this year. And given that 2012 is a presidential election year, no meaningful reforms will get through Congress, which is likely to continue to starve the regulators of funding and keep them from meaningfully implementing Dodd–Frank.

If the Democrats are able to recapture the White House and the House of Representatives, then Dodd–Frank will move forward much faster. Yet, if the Republicans maintain their majority in the House and/or win the presidency, they will let the regulation die or kill it outright. Either way, it looks as though it’s going to be 2013 before the dust settles on the political issue.

From a technology perspective, the entire trading infrastructure to support algorithmic trading in the OTC space still needs to be put in place. The first issue is not who has the fastest network pipe to the market, but who has all of the necessary historical data to back-test algorithmic trading strategies before they are unleashed on the market.

When setting up the new market structure, regulators will look to create a trade repository, which would allow the creation of historical trade databases. To do this, the industry needs to resolve the legal entity identifier (LEI) issue, which is needed to trade these instruments electronically. Some data experts expect this alone to take between 18 and 24 months to deliver.

When these markets do eventually launch, it will not be who has the best trading algorithm but who has the most intelligent smart order router. I have lost count of the number of organizations that have filed (or plan to file) for SEF status, and I think we are going to see more planned OTC liquidity pools than the 40 or so dark liquidity pools we have in the US today. With that level of liquidity fragmentation, dealers are going to have to route orders to multiple venues in order to get best execution.

Weaklings
The muscle these routers are going to need will make current smart order routers look like 98-pound weaklings. The amount of market data that they will need to consume and process will be closer to the equities options and fixed-income markets rather than simple FIX-based equities messages. After consuming all of the market data from the various SEFs, there is still the issue of each SEF’s individual market structure. Through most of the discussions around SEFs, market participants expect to see a request-for-quote (RFQ) model rather than order-driven markets. There are a few upstarts out there looking to create order-driven markets, but they are solidly in the minority. Yet, smart order routers will need to address how to route orders between the typically slower moving RFQ markets and their faster order-driven counterparts.

The eventual good news is that the SEF marketplace will consolidate relatively quickly compared to the US equities markets under Regulation NMS. The bad news is that it is going to be a rocky ride in this space for at least the next three to five years.