During last week's SFIG Residential Mortgage Finance Symposium, I moderated a panel on best practices in disclosure and reporting data related to private-label mortgage securities. We discussed many of the challenges confronting issuers, investors, rating agencies, and the industry with sharing relevant data in general and with implementing the SEC's Regulation AB II requirements in particular. Five minutes after my panel ended, my colleague Suhrud Dagli moderated a panel that discussed the applicability of blockchain technology to the securitization industry. Walking out of the symposium a short time later, I began to wonder how interesting it would have been if our two sessions had been combined.
Think about how much easier life would be if mortgage data were managed on private, permissioned blockchains.
The Status Quo of Exchanging Mortgage Data
Two decades ago, I worked in the investor reporting department of a $4 billion credit union. More accurately, I was the investor reporting department. I spent most of my time on the job doing two things:
- Processing and submitting data about our mortgage servicing portfolio to Fannie Mae, Freddie Mac, and the private investors who owned our loans
- Investigating and clearing hard and soft rejects that came back from those investors following my submissions
I spent far more time on the second task than I did on the first.
More recently, during a consulting engagement in the investor reporting department of an over-$250 billion mega-servicer, I noticed that virtually nothing of substance about investor reporting had changed. Some of the interfaces were a little cleaner—a reflection of enhanced web-based technologies that were still in their infancy when I did investor reporting for a living—but the overall process was the same. The institution employed large teams of people to do what I had done in my credit union job—albeit on a much larger scale. These people were tasked with extracting data about millions of loans from the institution's servicing system, transmitting that information to Fannie, Freddie, Ginnie, and numerous private investors, and then dealing with the immense fallout (in the form of hard and soft rejects) that invariably came when the investors received data that differed from what they were expecting.
It shouldn't have to be this difficult. What might contribute to making mortgage data exchange harder than it needs to be? At the crux of it is information asymmetry—the simple fact that mortgage servicers and mortgage investors are not dealing with the same amount of data. In most cases, hard and soft rejects would not occur if the investor knew everything that the servicer knew about the loan. But this is never the case. The relatively small handful of loan-level data fields that servicers transmit on a monthly (or exception) basis force investors to try to make sense of a chess match when they can only see half the board.
Recent efforts have been made, in both the Agency and non-Agency space, to reduce this asymmetry by expanding the data issuers and servicers share with investors. But these efforts, however noble, only chip away at the problem. Ultimately the problem of data asymmetry will only be resolved when originators, issuers, investors, and other relevant parties are all viewing the same transaction data at the same time via distributed ledger technology. The only lasting solution is blockchain.
Today's model for mortgage data exchange is based on an outdated notion of what is technologically feasible. Indeed, the fundamental assumptions underpinning today's mortgage data exchange model predate ubiquitous personal computers and the internet itself. This is evidenced in the terminology we still use when referring to mortgage data. Even though I am too young to have ever transmitted data via electronic tape (even 20 years ago, I transmitted loan data to Fannie and Freddie via a primitive internet connection), I still use the term tape colloquially (because everybody else does) to describe mortgage loan datasets.
The term tape conjures an image of a big mainframe computer somewhere that holds all the "real" data and generates copies (tapes) of some of the data it houses to share with other parties who don't have access to it. The fact that yesterday's mainframes have been replaced by cloud-based servers and tapes have been replaced by CSV or XML files has not changed our view. We still think of the servicer's database as a stand-alone "system of record" and the investor's database as a downstream application that needs to rely on, reconcile, and make sense of loan-level "tapes" generated by the system of record.
What if, instead of a single system of record residing with the servicer, every detail of every mortgage and every subsequent transaction were captured on a blockchain distributed to investors? Investor reporting as we have always known it would cease to exist. Gone would be the need to communicate to investors what happened last month, because investors would already know. They would know the instant it happened when Borrower A made a principal curtailment payment, when Borrower B's payment was rejected for insufficient funds, when Borrower C refinanced, and when Borrower D began a loss mitigation trial. Investors would not require formal reporting to know these things because all these transactions would be posted to a private, permissioned, distributed ledger that the investor could then use to update its position on the standing of the loans it owns.
Investors would still be empowered to reject questionable transactions, but this would be an extraordinarily rare occurrence. The overwhelming majority of today's hard and soft rejects result from information asymmetry—from investors not being able to see everything that's happening. Blockchain technology has the power to solve that. The immutability of transactions posted to the ledger would create an unmistakable audit trail—no more trying to unravel what happened to that loan that is 120 days delinquent with four NSFs, a partial payment in suspense, and an interest rate change somewhere in the middle. Trying to report the net effect of all that on a monthly tape is certain to blow something up. On a blockchain, it's just a sequence of transactions that everyone can decipher.
A Solution, Not a Panacea
Some challenges associated with mortgage data go beyond what can realistically be solved using blockchain. A blockchain, after all, is just a sophisticated database. It can't transform bad data into good data. Correctly and cost-effectively capturing loan-level data (including relevant borrower and property characteristics) at origination is something the industry is still figuring out how to do. Progress is being made, however, and once we're there, the march to locking that information down on a blockchain is sure to follow. The specific form this solution will take is yet to be determined, but forward-thinking fintech firms (like SmartLink Lab) are investing in them, their future is a certainty, and it can't happen soon enough.