The Gateway to Algorithmic and Automated Trading

Why can't we be friends? MMT and the quest for data standards

Published in Automated Trader Magazine Issue 32 Q1 2014

It's not every day that rival data vendors, venues and other companies gather round the table and agree to work together towards a common goal. But the rapidly progressing Market Model Typology (MMT) initiative shows that this kind of cooperation can happen, and the benefits are expected to be enormous.

Marc Berthoud, executive director, SIX

Marc Berthoud, executive director, SIX

"In terms of reducing the costs of data consumption, the mid-term perspective is a huge reduction in the expenses."

There are more than 60 equity market operators and some 90-odd data feeds in Europe.

"Imagine you've got all the major data vendors having to capture all those feeds, understand the logic of the feeds and try to normalise the output in order to have a coherent output for the end user," said Marc Berthoud, executive director, SIX. "The point is, for those 90 feeds there is no chance for the data vendors to have a fully converging educated guess of the data normalisation process."

MMT promises to change all that.

Instead of having to produce individual best guesses on the normalisation process, vendors would know that there was a standard in place for all trade messages. Trades of a similar nature would get flagged in a similar way and the data vendors could consume all the data in an automated way. No local reinterpretation. No guessing. No fuss.

"So in terms of reducing the costs of data consumption, the mid-term perspective is a huge reduction in the expenses," Berthoud said. In absence of something like MMT, vendors typically would have to employ squads of specification experts and developers to understand the logic of various feeds.

Beyond making life easier for vendors -- and by extension the clients of those vendors -- MMT has a number of other attractive features. Perhaps chief among them is how it provides an example of how the industry can work together, offering valuable lessons at a time when many in the financial community want to pursue industry-led solutions rather than finding themselves subject to regulatory cudgels.

As industry collaborations go, MMT is more the exception than the rule.

"I think it's fair to say it's been relatively rare," said Andrew Allwright, head of regulatory strategy for trading at Thomson Reuters, when asked whether there were many other instances like MMT. "It probably reflects the fact that there are common interests in terms of delivering a standard set of trade identifiers. Delivering a standard set of trade identifiers is something that really isn't against anybody's interest and is in everybody's interest."

Both Allwright and Berthoud have been directly involved in the initiative.

Andrew Allwright, head of regulatory strategy, Thomson Reuters

Andrew Allwright, head of regulatory strategy, Thomson Reuters

"Delivering a standard set of trade identifiers is something that really isn't against anybody's interest and is in everybody's interest."

MMT in action

"The idea of MMT is in fact very simple," Berthoud said.

The people spearheading the effort started by looking at data at a very granular level.

"The very first start was to agree on a common data model, finding logic, and very importantly, one common definition, making sure that in the end everyone understood exactly the same things about the semantics of the data elements," the SIX executive said.

That ultimately led to a huge spreadsheet for trade flag mapping. All the vendors and market operators agreed to what goes in those Excel cells so that apples can be compared with apples rather than oranges. As an example, in one FAQ document concerning MMT, it spells out exactly what the Bucharest feed does when there is an auction.

Or, to take another example, if you're a trader or a portfolio manager wanting to look at VWAP on a European blue chip on a variety of venues and lit order books, it can be highly difficult to compare all the data. But as a result of MMT, you would have the chance to apply MMT code filters on data from vendors and then you would know the data would be comparable.

Timestamp processes can present another minefield.

Berthoud said that in a trade message worst case scenario there could be an execution timestamp, a dissemination timestamp, a data processing timestamp by the data vendor and finally a single timestamp on Bloomberg, Thomson Reuters or SIX Financial terminal. "But you don't know which one it is," he said.

That can mean differences of microseconds or milliseconds. In OTC trading, it can get much worse.

"Typically an OTC trade is executed over the telephone, you will have one timestamp which is the end of the phone call and then someone will enter manually or electronically trade details in a trade matching engine. And then ultimately maybe it will be trade reported and published and might have three completely different timestamps," Berthoud said.

Prior to MMT, individual firms or vendors have had to interpret the data feeds from the various different exchanges and take one of two courses. They could just carry the underlying trade condition codes and leave it to the end user to worry about mismatches. Or they could try to normalise those codes into an internally devised model. But even if they did the latter, the lack of clear open standards meant that there was room for differentiation.

Allwright noted that you could have differing interpretations from different vendors or other firms on something as basic as a calculation of volume weighted average price. "In some cases you might be including certain types of trade or excluding trades, depending upon decisions that you have individually made as a firm in terms of what those codes meant," he said.

Berthoud said MMT is now in good shape. "We've got the consensus on the methodology,

on the data hierarchy," he said. Trade flag mapping Version 2.0, at the time of writing, was on its way and it contains even more granularity than the previously completed Version1.0.

Georg Gross, head of content, market data and services, Deutsche Börse

Georg Gross, head of content, market data and services, Deutsche Börse

"MMT provides a rulebook on how to do this and this will enhance efficient data consolidation."

The road to MMT and beyond

To get to this stage, however, required considerable effort.

The origins of MMT came from the start of the MiFID II review process. The Committee of European Securities Regulators (CESR) and the European Securities and Markets Authority (ESMA) had a working group on post-trade transparency. As part of that, CESR in October 2010 produced a document called Technical Advice Recommendations.

"It was in fact the very first document making recommendation on how to flag different types of trades in order again to bring more standardisation because there is a huge jungle of different trade flags for lots of good and bad reasons," Berthoud said. Local regulations, historical reasons, legacy issues all played a part in keeping this jungle thick.

That led to the Federation of European Securities Exchanges (FESE) inviting data vendors such as Bloomberg, Thomson Reuters, SIX Financial and Fidessa to get involved, along with FIX Protocol Limited (FPL) as a standards organisation. Berthoud said that for the early brainstorming, working under the umbrella of FESE was good.

MMT took on board the original CESR recommendations and added others that had not been there before, he said.

MMT's advocates say the effort now has enough momentum that its success is seen as inevitable. New versions of various feeds are now being released in MMT-compliant formats and over time that is expected to gather more momentum until MMT becomes the clear market standard. For instance, SIX will introduce MMT tags in its trade messaging from Q2 of this year.

Deutsche Börse is another venue that is planning to incorporate MMT. The exchange's head of content, market data + services, Georg Gross, noted how exchanges, MTFs and data vendors all collaborated on the flagging to achieve a true industry standard.

"We will integrate the MMT into our own data feed," Gross said. He said the big advantage is for data consolidators, reducing costs as well as potential for errors when consolidating.

"MMT provides a rulebook on how to do this and this will enhance efficient data consolidation," Gross said. He did not rule out that exchanges might consider getting involved in data consolidation themselves in the future.

This flexible approach means the industry can work at its own pace.

"The big difference between a regulatory and an industry initiative is the industry initiative is cheaper because you've got flexibility in terms of when to implement it. So logically, this is also part of the idea, that most market operators and data vendors will leverage their standard release plan to embed the implementation of MMT in it," Berthoud said.

One of the reasons for the project's success has been that the people behind it avoided 'scope creep'. MMT is ambitious enough and they didn't allow its aims to get too big or wide-ranging. "If you tried to address all the issues simultaneously you will fail because we never have the resources," Berthoud said.

Another factor that made a difference was how the people involved with MMT organised themselves and identified focus areas where there were lots of inconsistencies in the way data was handled. The industry stepped up to the plate by devoting resources and that meant that an MMT Technical Committee could be formed. Berthoud said "all the brains of data normalisation in Europe" were in the same working group and were all happy to work together.

In February of this year, ownership of the technical aspects of MMT was shifted to FPL, which is now going by the name of FIX Trading Community.

"We are not a regulator, and FPL is not a regulator, meaning everyone picks up what they need from FIX and implements the standard," Berthoud said. Instead of having a regulatory big bang, when a regulator orders certain activity by a certain date, the transition can occur at a more convenient pace.

The transfer of responsibility to FIX was a big step.

"FIX brings two things. Firstly, it ensures that the MMT standard will be open and free to use by the entire industry. Secondly, extending membership of the MMT steering organisation, especially on the buy and sell side," said Jim Kaye, co-chair of the FIX Trading Community Global Steering Community.

But there is much to be done. A key goal is to get to what Kaye referred to as the critical mass point, where enough firms are committed to delivering the required software.

"There is some level of work, mostly technological, involved within some firms which might cause a delay. But in order for this to really gain traction, there will be an element of coordination required between markets or brokers. Once we've reached this critical mass point, the data becomes useful, and that means a reasonable number of firms committing to and delivering the required software," he said.

Kaye added that the work has involved not just standards for MMT coding, but also for how to transport the data, handle amendments, cancellations and other issues. "Thus far this has mostly been a theoretical exercise so the challenge is going to be seeing how this works in practice as the industry starts implementing this, and responding to any issues that come out of that process," he said.

Allwright of Thomson Reuters said FIX's involvement could have a galvanising effect. "I think increasingly, given the strong FIX community around electronic trading, what we would expect is that that will increase the calls from end-users of data on exchanges and MTFs to incorporate MMT then in their basic feeds," he said.

Eventually, MMT will be a boon to efforts to create a consolidated tape in Europe, something that has been notoriously difficult to achieve.

"I think the next key aspect will be the definition by ESMA on its implementing measures around data consolidation and the consolidated tape," Allwright said. "I would think that there will be a combination of the commercial rollout of incorporation of MMT and underlying data feeds."

Berthoud said MMT will be supportive of whatever comes out of MiFID II in terms of the requirements for consolidated tape providers. That is because MMT will help address the issue of poor quality data flowing into the consolidated tape providers (CTPs).

"If you've got bad data flowing in with multiple inconsistencies, the CTP logic won't be able to perform any data cleansing, and if you've got 80 expensive feeds flowing in it's impossible to have a cheap output or someone needs to subsidise the output," he said.

"If all the data providers who are CTPs are relying on MMT definitions and MMT tagging you will massively improve the data input into the CTP."

Jim Kaye, co-chair, FIX Trading Community Global Steering Community.

Jim Kaye, co-chair, FIX Trading Community Global Steering Community.

"Common coding with well-defined usage rules means that post-trade data will become cleaner and easier to read, and contain more useful information."

Win win

While there appear to be no losers as a result of the MMT initiative, the biggest winners look like they're the data vendors, which will have the double-bonus of improved quality and lower costs. The accuracy of their aggregated data products is bound to go up and they will have a much easier task producing them.

Allwright said Thomson Reuters had not tried to put a figure on the savings. "We haven't entered into this with the kind of view of: 'What's in it for Thomson Reuters? If MMT is successful, we're going to save x million dollars per year, or we're going to make x million dollars a year in extra revenue.'," he said.

Rather, the vendor's clients will get a better product. "Ultimately, what this kind of standardisation means is it makes the data cleaner, better to use for our customers and we'd be crazy not to support it."

Berthoud also said it was not possible to put precise numbers on the savings. There are too many factors at play in terms of pricing of information products. "So from our perspective it's not possible to quantify," he said. "The only thing we know is the data consumption, data normalisation work will get cheaper."

Kaye of FIX Trading Community said ultimately the biggest benefit will be clarity. "Common coding with well-defined usage rules means that post-trade data will become cleaner and easier to read, and contain more useful information,"
he said.