While the notion of Transaction Cost Analysis has been broadly adopted by the equity markets, it doesn’t have similar adoption in the OTC world. Why is this so? I will attempt to explain the various factors that are driving this extremely important topic.
TCA adoption in the equities world
As the equity markets became more fragmented (chiefly as a desire from the politicians/regulators to encourage more competition amongst exchanges), investment managers (IM) demanded more information from the execution brokers as to how & where the brokers chose to execute trades done on behalf of the IM.
Provision of a metric or series of metrics allowed the IMs to calculate their broker’s execution performance. Originally this was via factors such as VWAP or implementation shortfall.
As execution algorithms became widely adopted, other factors were added to measure the algo performance across lit, dark & internal execution venues. In addition, the regulatory focus on best execution meant that TCA could provide an important measure of best ex. You cannot control a cost unless you can measure it.
FI & FX
In the meantime, the FI & FX markets were developing their own transaction models such as ESP, RFS & RFQ.
It was quite difficult for an IM to measure execution performance in these markets as there was no equivalent of a consolidated tape with time & sales information (whilst the equity markets in the US have this, Europe is still discussing how to deliver such a facility).
As the FI & FX asset classes trade in what can be considered dark markets (from a public exchange perspective), it was up to the individual execution venues to provide this sort of information. But there was no economic rationale for them to do so.
Hence the NASD pushed the corporate bond brokerage industry in the USA to TRACE which provided the consolidated post-trade tape. Once again, no such facility exists in Europe. Also for FX, there is no system that provides a consolidated timestamped tape.
What does the research indicate?
Clearly this is a major impediment to providing any measure of TCA in the FI & FX markets. Certainly the available research that attempts to investigate these asset classes is (1) patchy & (2) has various assumptions to be able to provide some measurement of cost analysis.
However, the research does indicate strongly that the existing markets are not as efficient nor transparent as they could be. While one response is to focus on industry-wide initiatives such as TRACE & the forthcoming MiFID 2 framework, I would argue that a bank offering these asset classes can provide a measure of TCA & that the bank clients would find such a service very useful.
Single-dealer platforms vs Multi-dealer platforms
If you use the example of an IM trading either FI or FX via a bank’s single-dealer platform, the bank can augment the platform by providing a measure of execution performance via the single-dealer platform. This could be by a comparison of the traded rate with the pricing of that instrument in the wider market at the time the trade was done.
While no bank wants to expose the mark-up to a client, I would argue that the multi-dealer platforms are doing that already by putting the banks in competition with each other where the client (IM) can see the all-in rates from each bank.
The problem for the banks in this situation is that the only feedback they get is the hit rate from each MDP! Or, for the smaller banks, the percentage of rates they quoted & didn’t get the trade. Not a nice number.