The accuracy of transaction cost analysis (TCA) systems is predicated on the market data they are fed. It is this data that frames a fund's trading decisions and allows a TCA provider to draw comparisons and assess performance against benchmarks.
Inconsistent or inaccurate data is created when firms report trades in different ways. Most stock exchanges developed their own identifying tags or flags in isolation. The European regulatory framework for trade reporting facilities that report OTC trades has allowed venues such as Markit BOAT to tred a similarly singular path.
But measurement requires a standard unit. And there isn't one, as far as trade reporting goes. Management, so the mantra says, requires measurement. It follows then that management of best execution is currently suboptimal.
Post-MiFID, European trades are still reported on the exchanges on which they were made. But they may take place in many other locations. Brokers might report OTC trades on exchange or they might report them to Markit BOAT. They might report both the buy and sell of a trade to Markit BOAT, thus double counting them. They might not. The data will not necessarily have identifiers that are needed for TCA purposes, for example whether it was crossed on a broker network.
When one takes into account the number of venues that are reporting trades, using the different trading flags and identifiers that they have accumulated, it is not hard to see why the picture of post-trade data for European trading is a mess.
TCA is not just an issue of performance measurement. It is also an issue of trust. A broker producing a report for its buy-side clients based on its own choice of data and metrics is essentially saying, “Trust me, you look great”. But what then is stopping a broker routing trades to a preferred venue at which it gets rebates for providing flow, or by internal crossing, saving the broker on exchange fees and maybe making the spread?
No doubt reputational risk outweighs other motivations for most brokers, but there are then two areas where standardisation could enhance the validity of TCA as a valid measure of execution performance. The first is the trade identifiers used by trading venues. Members of the Federation of European Securities Exchanges (FESE) have pledged to make a number of changes to post-trade data by the end of 2010, including the creation of cross reference capabilities across trade-type flags, access to post-trade data at reasonable cost and timeliness to assist with accurate supply of post-trade data by commercial suppliers. These measures, the members believe, will remove any post-trade reporting problems from FESE venues, leaving the brokers' OTC trade reporting as the final hurdle.
The Committee of European Securities Regulators (CESR), which is mandated by the EC to harmonise securities regulation across Europe, has introduced a working group comprised of market participants from all sides of the business to establish a common set of standards that will be used by all firms that report OTC trades. The group's recommendations will be fed into the European Commission's review of MiFID, for which CESR is acting as a technical advisor, at the end of July.
The recommendations are expected to include rules on timeliness, double booking of trades and an introduction of standardised flags to identify types of trades (for example internally crossed trades). Increased granularity and standardisation of information is the goal.
To put this in context, FESE believes that 30-40% of European equities trading occurs OTC. If a head of trading cannot rely on the information relating to 30-40% of a stock that he is trading, he cannot say for certain that he can measure his execution against the rest of the market accurately. Nor can his TCA prove it to be so. Common standards would provide certainty.
Click here here to vote in this month's poll