On Tuesday, November 30, Amazon Web Services dominated the news cycle.

In Las Vegas, the cloud giant was holding its re:Invent conference and a slew of announcements came out of that. Most significantly, Nasdaq said that it will migrate its North American markets to AWS, starting with US options. This comes on the heels of Google investing into/partnering with CME Group a couple of weeks ago. Here’s what I wrote about that deal on November 9:

The obvious big question is, Will this announcement lead to a domino effect? Amazon’s dominance in the space won’t be hurt by this, but Google becoming the core infrastructure technology provider of the largest exchange in the world is quite the feather in its cap.

As noted before, Nasdaq is vocal about moving its trading tech to the cloud. As the Journal article stated, the Singapore Exchange (SGX) and Aquis Exchange embarked on a pilot project to prove “it’s possible to run an exchange capable of handling ultrafast, high-volume trading on AWS technology.”

Does Google investing in CME lead the other Big Tech cloud providers to strike up similar deals? Or, is this the start of an arms race by the exchanges—fearing that they’re going to fall behind in the innovation space—pushing to sign contracts of their own? Also, can Google do a deal with another major exchange without pissing off CME and souring the relationship right at the jump? (In my humble opinion, I think one thing is for sure: AWS and Jeff Bezos and ICE and Jeff Sprecher are unlikely to join forces.)

Beyond Nasdaq, on Tuesday CNBC reported that Goldman Sachs “is opening up access to its trove of market data and software tools to hedge funds and asset managers in an offering designed with Amazon’s cloud division.”

And getting a bit lost in the shuffle, FactSet announced that it was joining—and providing 30 proprietary datasets to—the AWS Data Exchange, which was officially unveiled two years ago.

We will be writing about all three of these developments—but only after we’ve had a chance to talk to a swath of industry participants to better understand what these projects really mean for capital markets professionals. If you’re a subscriber of WatersTechnology then you know we prefer to write deep-dive, analytical features rather than knee-jerk news stories. Hopefully you view that as a positive and a reason to be a subscriber.

This column, though, is a space for my random thoughts and opinions—I leave the deep-dives to our smart journalists.

Speaking of which, let’s talk tape—consolidated tape. (Said in James Bond voice.)

Tape it

On Wednesday, November 23, we published a story reporting that—according to those pesky “knowledgeable sources”—Bloomberg, MarketAxess, and Tradeweb are considering joining forces to create a consortium that would deliver a regulated consolidated tape for bonds in the European Union—an EU CT, if you will.

Right now, it’s just in the talking stage as the three data giants are said to be awaiting the final legislative outcome of the review of the Markets in Financial Instruments Directive and Regulation, aka Mifid II /Mifir. 

Well, on November 25, progress was made on that front as the European Commission proposed measures to improve markets in the EU, and which addresses an EU CT.

Now, with all that said, I need to come clean about something: Jo Wright and Jo Gallagher are the experts when it comes to all things consolidated tapes—EU or US. After talking with Jo², here are some takeaways as I understand the state of things, but please do correct me if I’m wrong: anthony.malakian@infopro-digital.com.

* While the idea of an EU CT was first broached with the rollout of Mifid II, it never took off because there was no commercial incentive to build and run the thing.

From what I gather, this legislation represents something of a move away from a model based on multiple consolidators working in unison, to the emergence of one consolidator for each asset class—equities, derivatives, and fixed income. It would be a hub-and-spoke model where the consolidator sits in the middle of the wheel collecting data from the different venues for their specific instruments. That information then gets disseminated out to subscribers.

* One of the most interesting aspects of this pseudo-reboot is the CTs in each asset class would share their revenue with data contributors so as to compensate for their lost revenue. Which leads to…

* There’s a potentially lethal threat to the CT lurking—the idea of mandatory consumption. This concept was not in the proposals, but there’s been a fair amount of chatter around mandating consumption.

Mandating consumption is exactly what it sounds like: investment firms would be forced to consume the CT so that it could get up and running. The European Securities and Markets Authority (Esma) has recommended it: in a report published in late 2019, Esma said that while, ideally, a high-quality CT should attract demand in and of itself, it might need an initial boost to make back the resources spent in establishing itself, and anyway, market participants should contribute to the costs of establishing something that will benefit them. 

My thinking is that just like with making Covid vaccinations mandatory, there’s a crowd that exists that will not be pumped about the idea—but if it’s for the best of the market’s health, perhaps it needs to be done?

* The other concept that could prove to be a poison pill for market operators is the idea of “reasonable commercial basis”. This is actually quite important when it comes to market data costs in the EU.

Basically, Mifid II tells market operators and CTPs that they have to make pre- and/or post-trade market data available on a reasonable commercial basis—it has to be fair, affordable, and non-discriminatory. The problem is that “reasonable commercial basis” hasn’t been very clearly defined. Esma went with a transparency approach to RCB, expecting trading venues and systematic internalizers to publish clear and comparable information on how they charge for market data. That would theoretically enable both regulators and data users to compare venues’ pricing and revenue. 

Market participants say, however, that these venues don’t disclose all the information required, particularly when it comes to revenues from market data and how the price was set. They say the data is difficult to find and fee schedules are too complicated and unstandardized to be comparable. (Regulated markets disagree, of course, generally blaming market fragmentation for the differences in their publications).   

Esma set some guidelines to clarify the concept of RCB; the proposal will make these law so that the regulator can develop stricter technical standards on how it must work in practice. 

So I think this proposal is trying to define this concept more clearly—what is the fair price for market data? (Good luck!)

* And that brings us to the core problem surrounding market data costs: Do you charge based on what it takes to produce the data? If so, some might argue that the exchanges don’t actually produce the data—trading firms do. OR, do you base that cost on value to the consumer—which would favor the exchanges, which bring together ma and pa with the Flash Boys, pension funds, and Bobby Axelrods or Steve Cohens of the world?

Exchanges can be, like, “Excuse me…we have this incredibly valuable resource, why should we have to sell you this for pennies on the dollar when you’re going to use it to create Citadel or Renaissance?”

Conversely, those hedge fund titans create the data and the liquidity that has fueled the Dow Jones Industrial Average to cross the 35k threshold. Why are they having to buy back what they create?

* So it is as it ever was—it all comes down to the cost of market data and who should benefit the most from that information? Is there an equitable option that works for the vast majority of players involved? How can you make being a CT a viable, working product? And who will come forward to be the CTP?

It’s what the industry has been talking about for a while now, but the answer is still elusive.

* I won’t get into it much here because it’s a complex issue, but specific to a fixed income consolidated tape, there are major data issues that will need to be overcome before an EU CT in FI can become a reality. (Read Jo G’s deep-dive on the subject here.)

* There’s another long-standing question that has yet to be answered: How much will building, running, and creating a consolidated tape for equities and bonds cost the industry? The official estimates diverge wildly—for equities, €2.2 million to €7.8 million a year ($2.5 million to $8.8 million); for bonds, €1.5 million to €5.4 million ($1.7 million to $6.1 million).  

* Finally, even with all this fogginess, the greatest threat to an EU CT from the perspective of potential users is probably concessions made to established data suppliers. There’s a concern among people we’ve spoken with that as the tape goes through the notice and comment periods, they will be left with a product that doesn’t cost much less than the proprietary data they buy now. 

Just like in the US, the Infrastructure Bill and the Build Back Better bills are not as ambitious as some would like, or they are considered too bloated by others. You’ll never make everyone happy. The key is to find compromise that creates something effective and useful, and not a hollowed out stump or the creation of a camel.

The image accompanying the story is “A Woodland Road with Travelers” by Jan Brueghel the Elder, courtesy of The Met’s open-access program.