Faces are red at Refinitiv as numbers prove far from final
In Bolton town centre, just around the corner from Primark, you will find the offices of Hilton Smythe, a small business broker. If you’re in the mood for a coffee, coffee, or takeaway, “we’ve got exactly what you’re looking for,” its website says. The last deal he handled was the £1.3million sale of a cattery in Haywards Heath.
Not quite Goldman Sachs. Still, according to Refinitiv, some smaller consulting firms you’ve probably never heard of made huge sums during last year’s corporate deal boom.
In Refinitiv’s ranking of UK transaction advisory fees, the top spot went to Goldman. But the finalist was more unexpected. It was K3 Capital Group, an AIM-listed professional services firm that Refinitiv said had fees of $265 million through early December, beating JPMorgan.
In fifth was Preston-based Altius Group, and in seventh with fees of $138 million – ahead of Citigroup, Morgan Stanley and Barclays – was Hilton Smythe. These three small companies combined earned $613 million in fees for advice on mergers and acquisitions in the UK last year according to Refinitiv. This represents 17% of all UK transaction fees and 1% of the total worldwide.
As you may have already understood, this is nonsense. Of course, K3 didn’t earn more fees than JPMorgan. Refinitiv’s estimates for K3, Altius and Hilton Smythe are completely wrong. All declined to comment, but the numbers are at least a factor of 10.
READ Companies brace for another bumper year for mergers and acquisitions
For one of the world’s leading financial data companies, acquired by the London Stock Exchange Group for $27 billion last year, that would be embarrassing enough. But this is not a unique case. Refinitiv has been publishing wildly inaccurate fee estimates for these companies for at least a year.
For example, in its European Investment Banking Review for 2020, it claimed that Altius and K3 together had 4.7% of all European M&A fees.
It has made similar errors in estimating fees for small advisers in other countries and these errors were presumably included in the 2021 aggregate figures for investment banking fees widely quoted in the media. They would also affect the evolution of the combined market share of Wall Street banks in Europe, a subject that is attracting considerable interest.
How the hell did this happen? After alerting Refinitiv just before Christmas, the company was slow to figure out what was wrong. Here is his response.
In most cases, M&A fees are not disclosed, so Refinitiv and competitors such as Dealogic have models to estimate fees based on transaction values and other factors. Where no value is disclosed, Refinitiv assigns a default fee of $1 million. It works very well for JPMorgan. But more than a year ago, Refinitiv began estimating fees for K3, which advises dozens of small transactions with undisclosed values. Each of these transactions was assigned the default fee of $1 million.
Refinitiv’s next mistake was not verifying its estimates with the companies. In K3’s case, he could have simply Googled the company’s annual report, which shows that its M&A business had revenue of £16m last year. Not $265 million. Curiously, none of Refinitiv’s customers picked up the errors either, perhaps because their data is used less in banks than Dealogic’s.
What makes matters worse is that Refinitiv has made similar mistakes before. Four years ago, I pointed out that he systematically underestimated merger and acquisition fees collected by the big four accountants. Now part of Thomson Reuters, he discovered that the numbers were indeed wrong and he hadn’t bothered to check.
READ LSE-Refinitiv: a data-driven deal that pays off for the City
What safeguards have been put in place to prevent the same thing from happening again? Why didn’t someone with a little knowledge of the M&A industry look at the latest rankings and immediately realize they were wrong? Why hasn’t Refinitiv issued corrections and apologized to banks, law firms and other advisors who use its data for marketing and internal benchmarking purposes?
I asked the company these questions, but all they said in the file was this: “Our modeling is robust and continuously validated. The methodologies for counting transactions – volumes, values and fees – differ from company to company. Therefore, they cannot be used for comparable comparisons. Our approach is to estimate transaction fees below a certain threshold, as the majority of transactions do not disclose values. We believe this is the best way to increase transparency, providing our clients and market participants with broad coverage of trading activity in global markets. As 2021 was an exceptional and unique year for closing deals, it generated stakeholder feedback that will allow us to adjust and update our models accordingly.
It would be a mistake to put this out of proportion. Errors affect a tiny portion of Refinitiv’s business. But it’s a high-profile profile. Mistakes certainly won’t help Refinitiv’s efforts to challenge Dealogic on pricing data.
And if I were David Schwimmer, the general manager of LSEG fighting to convince the city that Refinitiv is a good buy, I wouldn’t be happier.
To contact the author of this story with comments or news, email David Wighton