The week in finance – network effects drive data oligarchies

Why is the market for financial data concentrated and lucrative?? Over the last couple of weeks, I’ve looked at the FCA’s interim report on competition issues in financial data and why demand for data is growing.? A couple of interlinked questions remain – why is the market for data relatively concentrated and why hasn’t all that demand created new, innovative, disruptive providers?

My basic view is that the answer is it’s a mixture of network effects and the requirements of the big buyers of data.? This week, we look at? what financial data markets look like.

Who buys data?

A lot of people.? Companies typically don’t publish a list of clients, for all kinds of reasons – commercial sensitivity and privacy are two pretty good ones.? However, LSEG and S&P Global both can’t help themselves coming out with some boasting statistics about clients.? LSEG has “>45,000 customers”, according to its 2022 Annual Report.? This seems a lot, but S&P Global has “~100k+” of them post the IHS merger.?

It makes sense for S&P to have more customers than LSEG, if for no other reason than that it is much broader. ?I show how S&P’s H1 23 revenues broke down in the chart below.

For instance, in H1 23 it earned $0.7bn of revenue from its “Mobility” business, which provides data on cars, auto dealerships and so on.? LSEG has no such business.? S&P’s Platts business (now in “Commodity Insights”) has a lot more oil related content than LSEG does.? The ratings business will have a lot of overlap with other financial data users, but the overlap won’t be perfect.? Engineering Solutions is small, but is likely to have virtually no overlap with LSEG’s business, as it is “a leading provider of engineering standards and related technical knowledge”, whereas LSEG isn’t.

A similar breakdown for LSEG is shown below.?

S&P doesn’t have any of the LSE’s trading and post trade businesses.? Very crudely, LSE is active in around 45% of S&P’s activities, while S&P covers about 60% of LSEG’s footprint. ??However, the trade/post trade area has a huge amount of overlap with LSEG’s data businesses, in a way that S&P’s energy and auto businesses just don’t.

Are there power users?

100,000 small businesses would not make THAT powerful a revenue pool.? If S&P served 100,000 artisanal coffee shops and LSEG 55,000 corner grocers, would anyone care??

Luckily, we have more boasting stats.? LSEG serves “48 of the world’s top 50 corporates”.? S&P sees their 96% and raises them “100% Of Global Fortune 100 and 80% of Fortune Global 500”.? So, a brief summary would be “the big data businesses serve almost all of the biggest companies in the world, and have many tens of thousands of clients”.

Walmart is top of the Fortune list, followed by Saudi Aramco, State Grid (you all knew that this was a “Chinese-government-owned power company, which meets about 80% of China’s electricity needs”, but I didn’t), Amazon and China National Petroleum.? You can see why these would all use S&P data (there’s a lot of oil/power in there, for a start) but I’d be very surprised indeed if any of these were a top 10 client for LSEG.? Their largest clients will, I reckon, come from the big dealer banks (JP Morgan, Bank of America, Citi, Goldman, Morgan Stanley etc), asset managers (BlackRock on down), asset owners (such as the various sovereign wealth funds), the systematic trading community (XTX, Citadel, Tower, Jump, Virtu, Flow) and so on.?

MSCI, another large, highly profitable data business, had “over 6,600” clients at the end of 2022, according to its Annual Report. ?BlackRock, the largest, represented 10.3%, with 95% of this coming from fees for index products based on MSCI products.?? 48 of the top 50 Asset Managers globally use MSCI ESG Research Products, we’re told.?

It’s not a surprise that, in terms of customers, S&P>LSEG>MSCI.? This is exactly what you’d predict based on breadth of the companies’ businesses.? However, 96% of the world’s largest asset managers being an MSCI ESG client shows how strong their market position is.? As with credit ratings, this is a very different claim to saying 96% use only MSCI data (they won’t – Sustainalytics, FTSE-Russell , ISS and others will be there too) but it’s still a strong claim.

These are oligopolies

The market structure world contains some natural monopolies.? Derivative clearing is one.? It makes overwhelming economic sense for a particular product to have one entity clearing it.? This is because at the moment the benefits of netting are orders of magnitudes more valuable than some fee savings.? I thought that MiFID II’s approach to derivatives was a sane one.? It aimed to regulate for competitive execution markets without trying to encourage competition in clearing.? In the end, this didn’t happen, but that’s another story.

Data is not like clearing.? If you decide to focus on MSCI’s index data and S&P’s credit ratings, you probably don’t see any financial disadvantage.? Maybe, just maybe, S&P might? bundle up indices with ratings and you might be able to save a bit that way, but it’s not the sort of transformational impact which netting efficiency has on derivatives.

Indeed, if you look at what people actually do, they typically use more than one provider in a particular area.? The FCA makes this point about both credit ratings and indices.? For credit ratings, people who are active in debt markets may choose more than one provider because they want complete coverage, and because some mandates also specify more than one credit rating.? With indices, again coverage may be an issue, and in any case asset owners dole out mandates referencing all kinds of different indices and this more-or-less forces the hands of the asset management community.

But there’s a limit.? Clients don’t seem to want to multiply service providers beyond a certain amount.? It looks, empirically, that that number is two to three.

Three providers to rule them

In my write-up on the FCA’s interim report, I flagged statements like “Three providers [of index data] account for a large majority of the UK revenues generated every year since 2017” (1.10) and “Approximately 92% of revenue from UK credit ratings activities are from the ‘Big Three’ CRAs, Moody’s Investors Service, S&P Global Ratings, and Fitch Ratings, a position that has remained largely stable over the past five years” (1.14).?

Another example of the “rule of three (max)” is equity markets.? Typically, in Europe you find that in each market there is a primary exchange with a market share of 50% or more (LSE, Euronext etc), Cboe with a material share and a few others (Turquoise and Aquis amongst others).? The chart below shows some data here.

So, data seems to be a set of oligopolies (equity trading is another).? Next week, we’ll look at why this has happened.