Trepp is an alternative data firm that focuses on the commercial real estate market for investors. Their data and methodology are packaged into reports that, according to their website, are targeted towards asset managers, brokers, CLO investors, and more.
This firm released a report a few weeks ago warning that many regional banks are the creditors on many CMBSs and mortgages, such as KeyBank’s 30% market exposure to the Key Center in Cleveland. In one report a professor called these assets “toxic”, while others pointed that the heavy concentration of banks in these regional markets could spell a liquidity risk if they disappeared.
There’s a lot in the report and the reporting that one can debate. That, after all, is how finance works: marketplaces of ideas battle it out, with the victor making money for being right.
In such a competition, your biggest fuel for winning is data. The second biggest is methodology.
This is partly why Trepp’s methods aren’t publicly released; that’s their secret sauce. But it also means someone with a better method and data can do better.
Thus the market for alternative data has many firms vying for position. But how are they made?
By definition, a report sold is going to be irreplicable, at least in theory. The report writers will either rely on unique data (in this case, Trepp has access to more pricing and bidding data on the CMBS market), a unique approach to public data (such as an algorithmic or statistical model combining multiple data points based on a theory of that data’s significance), or a combination of the two.
How much should a report cost? That’s a tough one. Theoretically, a report that is worth one dollar less than the delta between the profit you make on that report and what you’d make otherwise is the right price for the report. In reality, there’s never a way to know which reports will add the most value, which is why their prices will vary a lot and investors often have a loyalty with the research they use and a contempt for research they do not.
But does the research actually work? Sometimes. How do you know when you need to buy analysis and when it offers no value? To answer this question, large firms have analysts whose job is to analyze the analysis that other firms hire analysts to make. Is this all a bit weird and unnecessary? Possibly, but it’s the system we have until someone disrupts it–and no one has come even remotely close to doing that.