Brent Lippman, chief executive of specialist financial services optimisation firm Response Analytics, says that new technology offers the potential to refine and streamline the mortgage
Brent Lippman, chief executive of specialist financial services optimisation firm Response Analytics, says that new technology offers the potential to refine and streamline the mortgage loan modification process, making it easier for potential investors to determine valuations for distressed mortgage portfolios.
Take a look at a mortgage portfolio today. What is it worth? If it’s for sale (and for reasons we’ll get to shortly, we can expect to see more and more of these portfolios change hands over the coming year), the seller may have listed its value according to mark-to-market accounting rules.
Currently, however, mark-to-market valuations for debt obligations are almost always arbitrary, given existing conditions. A better surrogate for value, and therefore price, is measured by the cash flow they can be expected to generate over time – in other words, their hold-to-maturity value.
Unfortunately, the uncertainties involved in mortgage portfolios have been so great that, until now, attempts to determine their hold-to-maturity value have been almost as arbitrary as mark-to-market pricing. How many of the loans will default? If they default, will they go into foreclosure? If so, how much of the original obligation can be recovered? If not, what kind of modification would be successful?
While both of these options have always been problematic, in the current environment they have become even more so. Foreclosures and REOs (lender-owned properties) have always been costly, but the ongoing decline in the housing market on both sides of the Atlantic, compounded by each additional bank sale, decreases the amount that can be actually be recovered.
In the US, foreclosures attempts will become even less attractive if new Federal regulations allow judges to re-write mortgage obligations in bankruptcy, as a ‘cram-down’ modification will almost certainly be more aggressive than one lenders would have negotiated on their own.
None of this makes loan modification a simple proposition. Recent US government numbers have shown that over 50 per cent of mortgages that had been modified in the past year are once again in default. Given that even the comptroller of the currency, John C. Dugan, confessed his own bewilderment at borrower behaviour, it is no wonder that many would-be purchasers of these loans pass them by, despairing of ever determining their future cash flow.
However, new technology, based on applications that have been in use in the banking industry for the past 10 years, makes it possible to reliably model borrower behaviour. While past attempts at similar predictions have been based primarily on FICO score (based on a statistical analysis of a borrower’s credit report) and income, by looking at a wider range of factors, a more nuanced analysis taking into account location, loan-to-value ratio, and other factors create far better predictive modelling.
Ideally, the models would be built using the lender’s own data, but we have found that where that information is not available (for example, when a prospective buyer is looking at a portfolio), industry databases and developed benchmark models serve almost as well initially. In time, benchmark models are improved using more borrower behaviour data.
Once built, behaviour models not only identify which loans will likely end up in default, but can also predict borrower responses to various modifications, and are then used to support an ‘optimisation’ process to determine the single best outcome for every loan. By applying optimisation to each of the loans, all the while taking into account investor covenants, government regulations, and other operational constraints, a single best strategy for the entire portfolio is recommended and can be exercised.
With this, the investor now has the means to arrive at prices that more accurately reflect true value, and then use that information to be more effective in buy/sell transactions, and to improve the servicing performance of the portfolio based on optimised loan modifications.
While the technology is of immediate value in pricing the portfolio, it also serves as an operational tool, giving guidance to the new owners – or the contracted servicers – direction on optimised treatments for the loans.
Where traditionally it would have taken an experienced loan officer an hour or more to arrive at a single modification (which may or may not have ended up with the borrower re-defaulting), the optimised workout for thousands of loans in a portfolio can be determined in about the same amount of time – dramatically improving the productivity of servicing organisations.
And by using the technology dynamically, providing data updates on both the economic climate and borrower behaviour in real time, the recommendations and predictions can be revised in an ongoing way, as needed.
With all the uncertainty about the governmental framework, buyers and sellers alike have been understandably reluctant to approach distressed mortgage portfolios. With the arrival of a new US administration much of this uncertainty may soon come to an end, and we can expect to see portfolio owners eager to sell these assets. The investors who are able to determine their worth, and who can service them rationally, will find plenty of value in this new environment.