Rebuttal to GSEs Are Worthless Analysis

Posted October 20, 2009, on

On Monday, the respected financial services boutique brokerage firm, KBW, published a report declaring the common and preferred shares of Fannie Mae and Freddie Mac to be “worthless.” The conclusion of the report was based on a contrived scenario where Fannie and Freddie are separated into good-bank/bad-bank entities and the future profits are diverted away from paying down the government’s ownership stake in the companies. If this scenario were to happen, it would be another huge subsidy for the big mortgage banking firms at the expense of taxpayers like you and me.

Bank Co-operative Model is a Bad Idea

Fannie and Freddie should not be restructured into bank-owned co-operatives. The brokerage report held up the FHLB System as a model for the future structure of Fannie and Freddie. The FHLB System is a poorly constructed system and is not an enviable model. The FHLB System allows the large banks to borrow money at government-subsidized rates. The banks can use this borrowed money for any kind of lending they want such as auto loans, commercial loans, boat loans or even loans to foreign countries. This system has not prevented the FHLB Banks from posting large losses during the housing crisis. Another reason the FHLB System is bad is that it presents sizable systemic risks because it has no permanent capital. Members of the FHLB system can withdraw their capital on 90 days notice. In spite of nominally higher capital levels, the fact that FHLB capital is withdrawable makes the banks less stable than Fannie and Freddie. Moving Fannie and Freddie to a less stable capital structure is a bad idea.

Good-Bank/Bad-Bank Proposal Doesn’t Make Sense

Separating Fannie and Freddie into a good-bank/bad-bank as the KBW authors propose will not work because it is not a classic good-bank/bad-bank restructuring. First, in this proposal, the bad bank can’t support itself. Second, the authors transfer ownership of the good-bank to new owners. The concept of a good-bank/bad-bank split only makes sense if both entities are viable and self-supporting and have identical owners. The good-bank is a clean entity and can be more easily valued by the stock market. The bad-bank is capitalized to be self-standing and management can workout problem assets over time to realize maximum value without worrying about the timing of accounting gains or losses. The classic practical application was Mellon Bank in the 1980’s. In this report, the authors propose setting up the bad-bank as a non-viable entity from the start, and they assume a transfer ownership of the good-bank without any compensation. I submit that you can claim any financial institution in the country is worthless under the authors’ version of a good-bank/bad-bank scenario.

Holes in the KBW/GSE Model

There are multiple problems with KBW’s model that shows Fannie and Freddie having problems paying back the Treasury.

1. Bad-Bank Focus– As discussed earlier, as long as Fannie and Freddie are not separated into good-bank/bad-bank entities, they will be able to use revenue from new business to payoff the Treasury’s ownership position. Using the revenue from the good-bank will add $38 billion to Fannie and $25 billion to Freddie.

2. Double Counting Operating Expense – KBW’s model has operating expenses double counted. This adds $14 billion in expense to Fannie and $10 billion in expense to Freddie.

3. Severity Assumption is Too High – KBW uses 60% severity in its base case compared to John Hempton’s severity assumption of 45%. (BTW, the definitive analysis of the current value of Fannie and Freddie can be found in John Hempton’s series of blog postings on Fannie and Freddie.) I trust Hempton’s assumption more than KBW’s as he builds it up by vintage year. This accounts for $29 billion at Fannie and $15 billion at Freddie.

4. Mortgage portfolio run-off is much greater than mandated by Treasury – KBW assumes a run-off of the current on-balance sheet mortgage portfolio at a rate faster than mandated by the Treasury agreement. This accounts for $18 billion in lost revenue at both companies.

5. Assumed NIM is low – KBW assumed a 1% net interest margin which is low given the steep yield curve. If we assume Fannie and Freddie can keep their current margin for a three more years then revert to a 1% margin, it adds another $19 billion of revenue to both companies.

With these changes to KBW’s model, I calculate both companies can payback the Treasury:

($ billions) Fannie Freddie

Assumed shortfall from KBW model -49 -39

No good-bank/bad-bank separation +38 +25

Not double counting expenses +14 +10

Hempton’s severity assumption +29 +15

Slower portfolio run-off +18 +18

Current NIM +19 +19

New Capital Surplus +69 +48

With any model, the results are heavily dependent on assumptions. The user of a model must be well-versed in the assumptions before relying on its output. This analysis shows that by changing (and/or correcting) a few assumptions in the KBW/GSE model that the companies have plenty extra capital to pay back the Treasury, which is the opposite conclusion of the KBW report.

Another important assumption in the KBW model, which makes the GSEs’ capital positions look weaker is the assumption that they cannot raise capital from the public markets. Once the GSEs return to profitability, they may be able to raise equity to repay the Treasury.


The GSEs aren’t worthless until Congress restructures them and shareholders no longer have valid claims. Until Congress passes a law, the GSEs are working hard to mitigate their problem loans and to provide continued credit support to the housing market. Their income statements will flip from posting losses to reporting positive net income as we pass the peak loss years on the mortgages of 2006-2008. Freddie reported a profit in the last quarter, and it will be difficult to wipe out Fannie and Freddie shareholders once they return to sustained profitability.