If Jerome Sabah had his way, any time a client changed their trading patterns, he would get a tap on the shoulder, so he could call them and find out what they need. Data, he says, could give him that tap.
“Let’s say an investor trades something on a regular basis and I define a series of something relevant – trading times, amounts or instruments,” says Sabah, global head of rates, credit and forex sales for financial institutions at Societe Generale. “Any time that series diverges by 50%, I’d have a pop-up telling me something had changed.
“Data is becoming part of who we are and the way we have to be organised,” he says, “so we need to be able to read data in an efficient way.”
Many dealers, Societe Generale among them, had hoped Europe’s year-old trading and transparency regime – the second Markets in Financial Instruments Directive – would be a big step towards a world alive with data; a font of crystalline pre- and post-trade information, packaged and ready for consumption.
The reality is different. Instead of a rich vein of tradeable information, Mifid II has mostly produced a hash of jumbled, mostly unusable data.
At Societe Generale, Sabah says time and resources have been sunk into trying to glean some advantage from the Mifid numbers. “But in the end, the quality of the data and the takeaways aren’t there,” he says.
To be sure, banks have found one use for the Mifid II data set as it stands: they can get a better sense of their own market share. And there are at least three other ways they’d like to use the data: to check prices, analyse liquidity and turbocharge algorithmic pricing engines. But the data would have to improve first.
“We’re at a stage now where we’re evaluating potential use-cases for the data,” says a managing director of market structure at one US dealer. “The fact we’re still evaluating gives you a hint of the practical challenges associated with the data’s usability and usefulness.”
Too little of a good thing
While Mifid II introduced new data that didn’t exist before, dealers say there still isn’t enough of it to be of much use.
The regulation is replete with exemptions, deferrals and waivers. For instance, if an instrument is not deemed to be traded on a trading venue (TOTV), no pre- or post-trade reporting is required.
But even if a transaction occurs on a trading venue, there are ways to get out of reporting. If a trade is not considered liquid, for example, or is larger than a given size (in industry jargon, the size specific to instrument, or SSTI), then no pre-trade reporting is required. In addition, post-trade reporting can be deferred for up to four weeks, depending on national supervisors.
As it stands, very few instruments are deemed to be liquid under Mifid: in the bond market, for instance, only 439 bonds were considered liquid by the European Securities and Markets Authority (Esma) in its update last month. In the swap market, just five euro swap and two US dollar swap tenors made the cut.
According to MarketAxess, even by 2021 – Mifid II’s fourth year of implementation – only 38% of corporate bonds would be considered liquid for reporting purposes. In the EU, around 16,961 bonds were issued in 2017, according to Bloomberg data.
The result is a glaring lack of data on many bonds and derivatives.
“Esma’s current liquidity calculations deem almost every non-equity instrument to be illiquid, so there isn’t much data that’s required to be published in the first place,” says the head of government policy at a US hedge fund.
The size, or SSTI, thresholds are also seen as low: for a 10-year euro interest rate swap, the cutoff is €55 million ($62 million); for corporate bonds, it’s €1.5 million. Reporting deferrals apply to trades larger than those amounts.
A mirror of market share
Given the rampant pre-trade transparency waivers, only post-trade Mifid II data can be considered useful within the derivatives market – the biggest use of which being that banks can now better work out what their market share is.
“If you had asked any bank up until the end of 2017 what their market share in foreign exchange derivatives was, they wouldn’t have known because the information wasn’t public – they’d know what volume they were trading with clients, but they wouldn’t know the overall size of the market,” says Frédéric Jeanperrin, managing director and sponsor for Mifid II for global markets at Societe Generale.
“Today, if you’ve got the willingness, the tools and the time, you can retrieve the transactions that have been done and then you can try to approximate what your market share is,” he says.
Of course, market-share benchmarking services already exist, with firms such as Coalition Greenwich Associates and Tricumen selling reports to dealers across a range of asset classes.
We don’t see Mifid II data as an ideal substitute for benchmarking services since they tend to offer a much richer analysis of the composition of our client-franchiseManaging director of a US investment bank
While there is a risk that Mifid II data could make these firms obsolete, the managing director of a US investment bank says he considers that unlikely.
“We don’t see Mifid II data as an ideal substitute for benchmarking services since they tend to offer a much richer analysis of the composition of our client-franchise,” he says. “For example, deriving market share from publicly reported data only gives a single-dimensional analysis, whereas a survey like Coalition is more focused on wallet-share and also identifies opportunities for a firm to grow its franchise.
“So, it’s unlikely we would stop using benchmarking services purely because we could derive our market share ourselves,” he says.
Coalition and Greenwich declined to comment. Seb Walker, a partner at Tricumen, says the data could actually help develop its business. For example, it could use reporting data to predict the revenues of different banks in a given quarter.
“We’ve done some work there and have something which is pretty accurate when you start to backtest it,” says Walker.
Looking forward, wistfully
Given the very thin assemblage of Mifid data, market participants seem resigned to making the best of what’s on offer, and planning for another day. One future use of post-trade data is pricing; a market-maker could use Mifid II data to see what transactions have been done at what prices, giving them an idea of how the market is moving. Were more data available, traders could even feed such Mifid II data into pricing algorithms.
“I can’t give too many details, but I’m aware that our house and others are doing that to some extent on the data that is available,” says a managing director focusing on data at one global dealer.
Banks have already automated many trading and hedging decisions, and more data would allow them to expand those efforts. “I would, however, caution that most of this progress lies outside the realms of what is achievable with Mifid II data,” he says.
While the market structure source at the US dealer believes Mifid II data could one day be used for pricing algorithms, he, too, sees it as a distant prospect.
“Feeding this data into pricing algorithms is several years down the line,” he says. “The data would have to be a lot cleaner and more comprehensive and would have to add incremental transparency on top of those non-Mifid II data sources we already feed into our algorithms.”
Besides dealers, asset managers could use it to better understand pricing dynamics, says the head of government policy at the US hedge fund.
“Without an understanding of the other ongoing trading activity in the market, it’s difficult to benchmark whether you’re receiving a fair price,” he says.
Mifid II post-trade data could also be used to measure liquidity, although deferrals make that difficult.
“It’s possible this data could be used as an input for evaluating market liquidity,” says the US dealer’s market structure source, “but its usefulness will always be tempered, since by the time you get any data it’s inherently after the fact and doesn’t tell you what liquidity was like at the time you made the price.”
Generally speaking, the Mifid data has been a bit of a wash, he says.
“Most market professionals in fixed-income markets would say they already had a reasonably strong view of liquidity and where the market is, and that the Mifid data has not really added much on top of that,” he says.