Any time now, the Administration is expected to present its plan for reform of Fannie Mae and Freddie Mac. And by all accounts, one goal of that plan will be to shrink the footprint of the government-sponsored enterprises (GSEs) and the predominant role those two firms currently play in mortgage finance. That goal faces a potential roadblock, however. If the share of the mortgage market held by the GSEs is to shrink, either the mortgage market must also shrink by a corresponding amount — an outcome no one wants — or the private sector, and banks in particular, must fill the void. But currently, a combination of Federal Reserve and FHFA regulations impose dramatically higher capital requirements on banks than on GSEs for the holding and servicing of mortgages, and make it uneconomic for banks to be engaged in those activities or make markets in any mortgage-backed security not guaranteed by the GSEs.
The most recent data on mortgage originations indicate that more than half of the mortgage loans guaranteed by the GSEs were originated by non-banks. According to the Federal Reserve’s Flow of Funds Accounts, approximately 45 percent of all mortgage debt is guaranteed by Fannie Mae and Freddie Mac. The GSEs are necessary because non-banks do not have the capital to hold these loans on their balance sheets. Furthermore, the current market increasingly consists of non-banks servicing those mortgages for the GSEs. Those servicers carry far less capital than their bank servicer counterparts but can persist because they are unregulated for safety and soundness and because the GSEs do not impose significant capital or liquidity requirements on them. So, to the extent that the Federal Housing Finance Agency (FHFA) required the GSEs to increase their capital and shield themselves to servicer default (which would require servicers to hold more capital), the GSE-dominated part of the market would inevitably shrink and grow more expensive.
For the bank share of the market to grow and serve to moderate the impact of GSE reform on the pricing of mortgage loans, however, the FHFA cannot do it alone; the Federal Reserve must do its part.
To illustrate why, consider the two stress testing regimes that now produce the binding capital requirement for the GSEs (through the FHFA’s stress test) and large U.S. banks (through the Federal Reserve’s Comprehensive Capital Analysis and Review stress test). Both tests use effectively the same severely adverse scenario — that is, the same hypothetical rise in unemployment and drop in housing prices. Yet for the average mortgage loan in the 2018 stress tests, the Federal Reserve estimated a 3.2 percent loss rate for the set of banks that hold mortgages with similar characteristics to the GSEs, which is eight times higher than the loss rate projected by the FHFA (0.4 percent). That means that for comparable loans experiencing the same stress, the Federal Reserve requires a bank to hold more than twice the capital than the FHFA requires of Fannie Mae and Freddie Mac.
Two factors explain much of this difference in stressed loss rates, and two additional criteria further increase the stringency of bank capital requirements relative to GSE capital requirements.
First, the FHFA considers private mortgage insurance in determining the loss given default of mortgages held on the books of the GSEs. The Federal Reserve ignores such insurance in calculating the loss given default of mortgages held by banks.
Second, the FHFA encourages the GSEs to use loss mitigation programs and avoid foreclosure by explicitly accounting for the loss-reducing impact of those programs in its stress test. The Federal Reserve gives banks more limited credit for their (quite robust) loss mitigation programs via modeled pre-default cure rates. As a result, the default of a mortgage with very similar characteristics results in a lower loss under the FHFA stress test than under the Federal Reserve’s.
Third, and moving beyond loss rates, another component of the Federal Reserve’s stress test is the risk-weight assigned to mortgage and other assets before the stress occurs — that is, the starting point for how risky they are. For this purpose, the Federal Reserve uses the standardized risk weights developed by the Basel Committee, which is set at 50 percent — basically half the risk of an unsecured loan — regardless of the loan-to-value ratio, credit score of the borrower, and other risk attributes. The FHFA’s proposed risk-based capital requirements are more tailored to the characteristics of mortgage loans, yielding a lower (and almost certainly more accurate) average risk-weight. Putting these components together, the holding of a prime or relatively low-risk mortgage is quite profitable for Fannie Mae or Freddie Mac, but hugely unprofitable for a large bank.
Fourth, there is also a significant disconnect between the Federal Reserve and the FHFA regarding the risk-weighting of mortgage servicing assets, where the Fed has already acknowledged that its regulatory treatment is too harsh. It is also why we have seen banks abandoning the mortgage servicing business.
In most of these cases, we could debate whether the view of the FHFA or the Federal Reserve is correct, or whether the use of GSE and bank internal models (reviewed by the regulators) would be a better approach than either. Let’s leave that for another day. What seems undeniable is that for any semblance of rational mortgage market reform to occur, those two agencies — both part of the U.S. Government — need to meet and decide on a common approach, whatever that may be.
A meeting of the minds between the FHFA and the Federal Reserve on the capital treatment for mortgage and servicing assets is necessary for banks to hold and service more mortgages. However, further action — here, by the Fed alone — is also necessary to resuscitate the securitization market and allow other private-sector actors to put capital to work. Since 2008, less than 1 percent of new mortgage originations have been securitized as private-label mortgage-backed securities (MBS). Although prime jumbo origination volume has increased since 2008, only about 3 percent was securitized in 2017.
Here again, the assumptions embedded in the Fed’s stress tests drive the outcome, by establishing unrealistic loss assumptions that dramatically favor MBS guaranteed by the GSEs over private-label MBS. For securitizations held for investment, the Fed relies on data from the past financial crisis to project the credit losses instead of collecting the required data needed to model the stress losses of more recent securitizations, which include relatively low-risk prime jumbo loans instead of subprime and Alt-A mortgage loans. Furthermore, for securitizations held for trading, the Fed projects the spread of MBS guaranteed by Fannie Mae to rise 45 basis points under the instantaneous global market shock component but projects the spread of AAA private-label residential MBS to rise 1940 basis points — almost 45 times wider. That difference, which seems inexplicable (and which the Federal Reserve has never explained) is effectively a mandate to banks not to hold private-label MBS. It also clearly illustrates the power the Federal Reserve exercises through the assumptions that it embeds in its stress test.
Put it all together, and you have one agency of the government poised to push mortgage risk out of GSEs, with another regulator pushing back. Maybe it’s time for them to have a conversation.
Disclaimer: The views expressed in this post are those of the author(s) and do not necessarily reflect the position of the Bank Policy Institute or its membership, and are not intended to be, and should not be construed as, legal advice of any kind.