European guidelines issued in November, the latest regulatory attempt to address what market participants say is the high cost of market data, might help to harmonize some existing practices at trading venues. What they won’t do, though, is address the root causes of these high costs, or the complexity of market data usage policies and licenses, market participants and practitioners say.

“This document is a way of trying to satisfy everybody, but really ends up satisfying nobody,” says Aquis Exchange CEO Graham Dick.

In November, the European Securities and Markets Authority (Esma) released regulatory guidance in the form of 16 draft guidelines for public consultation, aimed at helping trading venues and market participants understand how regulation applies to keeping market data costs reasonable, and pricing policies transparent. But sources say the guidelines will not have any impact on costs because they do not address the complexity of policies or the fees that trading venues are now levying for so-called “non-display” data and “derived” data.

Regulatory Flashpoint

Regulators must be used to pleasing no one when it comes to their efforts around market data costs. This is a volatile flashpoint for them—there are just so many vested interests and competing claims involved.

Consumers of market data have a list of complaints: Exchanges have always had effective monopolies and massive market power because of the exclusive data they produce as a by-product of their activities; exchanges want to charge for data based on its worth, rather than on a “reasonable commercial basis,” (or, in other words, fees that cover the cost of producing and disseminating the data, plus a reasonable margin); and the Markets in Financial Instruments Directive (Mifid and Mifid II) and its delegated regulation, Mifir, have failed to bring down high costs.

These guidelines are trying to set common standards, and to make sure that market data providers are not able to manipulate the basis on which different types of customers can access the data without good justification.

Michael Thomas, Hogan Lovells

These consumers complain that as they are increasingly taking in and using market data in new ways—routing it to execution algorithms and risk calculations, deriving their own products from it—the trading venues try to muscle in on what they see as new revenue sources by applying new fees and new policies, which have evolved to be opaque and of Byzantine complexity.

Exchanges, on the other hand, say they have businesses to run and bottom lines to protect. They have high overheads, and must invest in robust, resilient connectivity, data infrastructure, and human resources. And, they say, contrary to market participants’ claims, exchange fees have not rocketed up in the past few years. Rather, expenditure on data has increased because companies are buying more: Market fragmentation and automated trading have increased the volume at which market data is needed.

Guide to the Guidelines

Having canvassed these opinions, Esma put out a consultation almost a year ago that assessed industry views of the cost of data since Mifid II and Mifir came into force. The regulator also said at the time that it would issue guidance for exchanges on the requirements for trading venues that would bring down costs, a promise that resulted in the November guidelines.

The 16 draft guidelines clarify the Mifid II and Mifir rules that exchanges must provide data on a non-discriminatory basis, that they charge for market data on a per-user basis, that they unbundle data for consumers, that they make prices and terms and conditions transparent and easily accessible, and that they provide data free of charge 15 minutes after publication.

Michael Thomas, a partner in the financial services team at law firm Hogan Lovells, says there is quite a variation in the ways that market data can reach end users, and these draft guidelines are trying to introduce a consistency of approach, regardless of how the data is consumed, to ensure that recipients using it for the same kinds of use-cases are treated equally.

“Mifid II sought to introduce a more level playing field for all aspects of the financial services sector, including, importantly, the access to data. So the aim of these guidelines is to ensure that there are consistent standards and approaches for the market, being able to access the market data they need in order to inform their own trading activities and strategies, [and] to inform the design of products that rely on market data,” Thomas says.

For example, the guidelines say that customers buy market data to use for multiple purposes like research, index production, or portfolio management, and are sometimes required to pay multiple times for it.

“Where data use-cases are not clearly predefined, such practice renders it very difficult for customers to understand which fees are applicable to them. Furthermore, a basic fee is sometimes added by default regardless of the use made of data, inflating unnecessary [sic] the price,” the guideline consultation document says. “Esma is of the view that customers should not be required to pay multiple fees for the same data.”

Guideline 4, therefore, requires market data providers to consider where customers could belong to more than one customer category and make sure they are classified as one type only.

Or, in another example, guideline 6 clarifies that trading venues must charge on a per-user basis—not per-screen. If a trader uses, say, a Refinitiv Eikon and a Bloomberg Professional service, they should not be charged twice for viewing the same data on both.

“These guidelines are trying to set common standards, and to make sure that market data providers are not able to manipulate the basis on which different types of customers can access the data without good justification,” Thomas says.

Most of these draft guidelines are already practiced, more or less, by all trading venues in Europe—what the guidelines are trying to address is the “more or less.” Guideline 6, in the example above, is standard practice now at larger exchanges, but there might be one or two smaller ones in smaller member countries that still charge per screen, and the draft guidelines make it clear that Esma will enforce the rule to a particular standard.

“Charging per screen [rather than per user] became undefendable about five years ago,” says one market data expert who has worked at both large banks and exchanges. “Nearly all of the exchanges have stopped using that and have moved to the per-source, or the ‘bums-on-seats’ model. And they seem to be more willing to simplify and charge by the number of people using the data, mostly because they are making money out of other policies now.” 

Janet Mail, head of commercial management at market data consultancy CJC, says guideline 11, which says that market data providers should adopt regulators’ standard definitions in their market data policies and price lists, will be helpful. Some of these terms, like “customer,” “derived data,” and “non-display data,” may seem pretty basic, but their definitions are not harmonized across policies; different trading venues have slightly varying framings of what these terms mean.

Euronext, for example, includes derived data under its definition of “original created work”; Deutsche Börse incorporates it in its definition of “information.”

These standard definitions are a step in the right direction toward harmonizing the policies of the 40 or 50 different trading venues in the EU, Mail says.

“The standardization of definitions—what derived data is, what an application is—will help,” she says. “Right now, a user can’t just say, ‘Right, [we should] use the derived data,’ because they would have to check that against 40 or 50 derived data policies, each of them slightly different. What I consider to be derived data might fall into the definition of derived data for 15 exchanges, but it might fall outside of the definition for 15 other exchanges. If there was standard wording, that would start the alignment of policies.”

Pleasing No One

But, Mail says, while these guidelines are just a step in the right direction, “they won’t have much impact.” Even if there are aligned definitions between 40 different policies, there are still 40 different policies.

Dick from Aquis agrees that the draft guidelines are not enough.

It doesn’t truly address the problem, and it doesn’t simplify or make market data more transparent at all.

Graham Dick, Aquis Exchange

“There have been loud cries for transparency in market data pricing because it is considered to be opaque. Every different provider of market data has a different pricing policy, and they can run into 60, 70, 80 pages. So where there is some standardized template for how market data pricing is done, whether that will change anything, I doubt it. And we will probably have a 70- or 80-page template anyway to try and fit in what everybody wants to get in,” he says. “It doesn’t truly address the problem, and it doesn’t simplify or make market data more transparent at all, from what I can see.”

He adds: “The issue is that trading venues have implemented so many kinds of licences that it has got into this ridiculous level of detail.”

The reality is that the ways in which market data is consumed, and the uses to which it is put, have evolved, and exchanges have evolved their policies along with them. When data used to go principally to terminal screens, exchanges could charge per screen or per user—the “bums-on-seats” model. However, about 10 to 12 years ago, data usage began to shift away from display usage—prices flickering on a screen—more and more to non-display purposes, for feeding into execution algorithms or smart order routing or pre-trade risk calculations. Because no human ever looked at this data on a screen, there were fewer users overall, and fewer bums on seats to charge for.

The Federation of Stock Exchanges (FESE), in a response to Esma’s December 2019 consultation, quoted research that demonstrated how dramatic this shift has been at some firms. Goldman Sachs’ cash equities trading floor employed 600 traders at its height in 2000. Today, there are just two equity traders left.

“Complex trading algorithms, some with machine-learning capabilities, first replaced trades where the price of what was being sold was easy to determine on the market, including the stocks traded by the Goldman 600. These new data users (quant, robotic, and artificial intelligence systems) require constant investment in hardware and software by data providers in order to keep up with the new technologies used by these systems,” the FESE said.

Trading venues were losing out by failing to capture these new applications of their data. So they started adding non-display policies to their market data pricing policies.

The London Stock Exchange introduced its initial non-display policy in 2010. In 2012, the Oslo Børs began charging separate fees for clients that used its data in non-display applications, in line with a similar policy at Deutsche Börse. In 2015, the New York Stock Exchange (NYSE) scrapped a policy for certain data feeds that allowed users to net their costs and pay once for using the data in multiple display devices, while introducing new policies on non-display usage to better reflect the increase in non-display use of the data.

Somewhat more recently, trading venues have tried to capture revenue from derived data, as users increasingly turn the market data they ingest into their own products, such as indices or pricing models, by running it through their own models, combining it with other data, and otherwise manipulating it.

These new licenses have incensed market data consumers, who say that trading venues have no production costs associated with a market participant’s derived data products.

In April, US trade body the Securities Industry and Financial Markets Association (Sifma) vocalized the opinions of market participants on both sides of the Atlantic when it complained about “a new business practice whereby exchanges force market data purchasers to consent to a licensing interest in derived market data. … Unlike other, typical contract negotiations, market participants are forced to agree to such terms as a condition to purchasing the market data from an exchange that is an exclusive purveyor of that data.”

Exchanges respond that they have to make money somehow, and their data is of huge commercial value to the market. As the FESE put it, “Complete business models are based on the valuable reference price data which is provided by exchanges on a non-discriminatory basis. … Exchange data fees represent production costs for third parties’ business models, rather than a regulatory requirement, as often presented.”

Broadly speaking, EU exchanges will probably support most of Esma’s guidelines, though there are some stipulations that could produce additional administrative costs. If they demonstrate compliance with Mifid II’s reasonable commercial basis requirement, regulators may choose to ignore calls from market participants for much blunter instruments. Some dealers have called for revenue caps based on the “long-run average incremental cost” model that is sometimes used to regulate the fees of telecommunications providers with significant market power.

The Real Solution?

For some in the industry, a European consolidated tape—particularly a post-trade tape that is cross-asset, cheap, or even free, and as close to real-time as possible—is the real answer to market data costs, and the complexity of market data contracts. Esma’s consultation included questions on why a tape provider failed to emerge in Europe, despite regulatory provisions in Mifid II. But proponents are optimistic that regulators and lawmakers will address the tape next year, perhaps with changes to actual Mifid II legislation.

Jonah Platt, US head of government and regulatory policy at Citadel, says that “even if these [Esma] guidelines were fully implemented, we think that a consolidated tape of post-trade data is just fundamental to allowing market participants to be able to achieve an aggregate view of trading activity across the continent.”

However, he says, “There must be mandatory contributions to the consolidated tape provider (CTP) from each trading venue and approved publication arrangement (APA). Just as it’s not viable for market participants to go around and negotiate individual data agreements with each trading venue and APA in the EU, it would not be viable for a CTP either.”

Dick from Aquis is also an enthusiastic supporter of the idea that where Esma’s regulatory efforts fail, the consolidated tape will succeed.

A consolidated tape of market data would “create a single golden source. This is important because—principally because of all these market data contracts—you still don’t know what a particular security has been traded [and] where on a real-time, consolidated basis. The very top professionals can take and source all of that data all of the time, so they know. But everybody else doesn’t. So it would simplify the data, but also simplify the contractual distribution of the data,” Dick says.

“If you were to implement that, a large percentage of everyday users could sign a single contract with a single entity that would cover all their usage on a pan-European basis. That is very much where I would like Europe to go in the long term,” Dick says. “Where we are today is that it’s a series of additions and modifications, piling up old policies on top of new policies on top of old policies that have created these monster market data agreements.”