As part of a joint initiative, the FHFA and HUD recently released a presentation outlining potential revisions to the current system of servicer compensation. The initiative's stated goals are 1) improving customer service for borrowers, 2) reducing financial risks for servicers, and 3) helping servicers better manage their pipeline of nonperforming loans (NPLs). In light of the asset's unusual economics, an examination of servicing compensation is long overdue. However, the initiative's proposed changes to the compensation structure for NPL servicing runs counter to the notion of reducing the mortgage market's dependence on the government.
The presentation outlines two distinct issues for discussion. It primarily analyzes a series of proposals for reducing the minimum required servicing, which currently is 25 basis points for most products.(As a brief refresher, the minimum servicing is an interest strip that must be held by the servicer. The remaining servicing is considered "excess," and falls out based on the originator's execution decision.)
The idea behind reducing required servicing is that institutions are forced to hold unnecessarily large amounts of servicing assets, since performing loans cost much less than 25 basis points to service. The servicing asset is difficult for institutions to fund and hedge. Moreover, servicing will be a financial burden for large institutions once Basel III's requirements go into effect, as it will require dollar-for-dollar capital once it exceeds 10% of an institution's Tier 1 capital.
The idea of reducing required servicing has been debated for years. It was repeatedly opposed by a number of large investors who believed that reducing servicers' "skin-in-the-game" would incentivize servicers to solicit their customers for refinancings, driving up prepayment speeds. The presentation (rightly, in my view) points out the flaws in this argument. A servicer's incentive to solicit its existing customers increases with the amount of servicing held, as servicers seek to retain increasingly important assets. (If servicers don't contact their customers, someone else will.)
The other issue, addressed in much less detail, proposes changes to compensation for NPL servicing. While servicers are overpaid to service performing loans, NPLs are typically serviced at a loss. The initiative proposes to pay servicers "market rates for performing the defined NPL servicing protocols," with such payments being "financed by the guarantor." This statement implies a major change to servicing practices. In all but a few cases, servicers currently do not receive any servicing fee for delinquent agency-eligible loans. (For non-agency loans, fees are accrued and paid out of liquidation proceeds.) Depending on how it is defined and implemented, the financing of NPL servicing costs by "the guarantor" suggests that some costs associated with servicing would be transferred to the GSEs or government agencies (and, by extension, taxpayers).
Under this interpretation, the initiative's proposed changes to NPL servicing compensation are fundamentally inconsistent with the goal of phasing out the GSEs and replacing them with "private capital," as outlined in the Obama administration's white paper. Although the presentation makes no mention of NPL servicing in the absence of a "guarantor," the proposal would dramatically improve the relative economics of servicing agency NPLs vis-a-vis non-agency loans; servicers would receive timely compensation for agency loans, while costs for non-agency loans would be recovered only upon the loans' liquidation. A cynical explanation is that the proposal represents an attempt by the FHFA (the GSEs' regulator) to ensure its future by entrenching Fannie and Freddie in the servicing process.
I support the concept of reducing the amount of required servicing. However, changes to NPL servicing compensation practices must be fashioned to avoid making the servicing industry increasingly dependent on the government's backing.
Bill Berliner is a mortgage and capital markets consultant based in Southern California. His email is firstname.lastname@example.org and his Web site is www.berlinerconsulting.net.