Have the Government-Sponsored Enterprises Fully Repaid the Treasury?
Larry D. Wall
March 2014, Atlanta Fed
Have U.S. taxpayers been fully compensated for their bailout of the government-sponsored enterprises (GSEs) Fannie Mae and Freddie Mac? The Treasury is reported to have argued that “the value of Treasury’s commitment to the GSEs was “incalculably large,’” with the implication that it could never be repaid. Richard Epstein, the Peter and Kirsten Bedford Senior Fellow at the Hoover Institution, responded that “the level of the Treasury commitment was not ‘incalculably large’: it was $188 billion, all of which will shortly be repaid.” The significance of Epstein’s argument is that if Treasury has been fully compensated for its bailout of Fannie and Freddie, a case can be made that the future profits of the two GSEs should go to their private shareholders. [Author's note: Epstein responded to this post that he was referring to the situation in 2012, around the time of the third amendment to the Federal Housing Finance Agency's conservatorship agreement. In contrast, my analysis relates to the entire post-conservatorship period. However, many commentators have noted that the GSEs would soon pay dividends equal to the $188 billion without such a context, leaving the implication that taxpayers had been fully compensated. See, for example, here, here, and here.]
As an accounting matter, one could argue that Epstein is correct; the dividends equal the amount of Treasury funds provided to the GSEs. And as a legal matter, the issue may ultimately be resolved by the federal courts. However, as an economic matter, the value of the government’s contribution clearly exceeds $188 billion once the risk borne by taxpayers is taken into account.
In this Notes from the Vault I examine the value of the taxpayers’ contribution to Fannie and Freddie from an economic perspective. My analysis of these contributions is divided into three parts: (1) the GSEs’ profitability prior to the 2008 conservatorship agreement (bailout), (2) the value of the taxpayer promise at the time of the bailout, and (3) support of new investments since they were placed in conservatorship.
The GSEs prior to conservatorship
The Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mae) have long been closely linked to the federal government (see W. Scott Frame and Lawrence J. White, 2005). Fannie Mae began as a government corporation chartered during the Great Depression. It was privatized in 1968, but continued to receive a variety of benefits from the federal government. Freddie Mac was chartered in 1970 and converted into a publicly traded company in 1989. It also received a variety of federal benefits. These charter benefits included a line of credit with the Treasury, exemption from state and local taxes, and the ability to issue “government securities” for the purposes of the Securities Exchange Act of 1934. These benefits, combined with prior federal support of distressed GSEs, gave rise to market perceptions of an implied federal guarantee.1 These perceptions increased the value of GSE guarantees and lowered their funding costs.
In effect, Fannie and Freddie were in a partnership with the federal government—with GSE shareholders bearing part of the risk and receiving all of the profits. In return, U.S. taxpayers bore all of the risk of loss in excess of the GSEs’ very low capital requirements and received, at most, a small portion of the profits taxed away through the imposition of minimum lending requirements to low- and moderate-income households (the affordable housing goals).2 Thus, it’s little wonder that in 1999, then Fannie Mae CEO Franklin Raines is reported to have stated “[w]e manage our political risk with the same intensity that we manage our credit and interest rate risks.”3
Embedded losses at time of conservatorship
The decline in housing prices in 2007 and 2008 resulted in increasing levels of financial stress on Fannie and Freddie (see W. Scott Frame, 2009). On September 7, 2008, the Federal Housing Finance Agency placed the two GSEs into conservatorship. Along with this action, the Treasury entered into senior preferred stock agreements with Fannie and Freddie to insure that the value of their liabilities did not exceed the value of their assets. The Treasury initially committed up to $100 billion to each of the GSEs to maintain non-negative net worth. In return, the senior preferred stock accrued dividends at a rate of 10 percent per year, which increased to 12 percent if the dividends were not paid in cash.
That stock agreement with Treasury was subsequently modified three times. The third amendments with Fannie and Freddie in August 2012 effectively removed the limit on Treasury’s promise to maintain positive net worth—that is, the Treasury would now cover the full value of whatever losses were incurred by each GSE. It also replaced the requirement of regular dividend payments with one that starting in 2013, each entity pay “the amount, if any, by which the Net Worth Amount at the end of the immediately preceding quarter exceeds zero.” In other words, essentially all of Fannie and Freddie’s profits are to be paid to Treasury.
Fannie and Freddie suffered $230 billion in losses in their businesses from 2008 to 2011, according to my calculations using the Federal Housing Finance Agency’s first-quarter 2013 Conservator’s Report. These losses far exceeded the two GSEs’ book capital of $78 billion at the time of conservatorship, which would have resulted in a combined negative net worth of $152 billion absent the senior preferred stock purchases by Treasury, according to the Conservator’s Report. In addition, Treasury earned $36 billion in dividends under the agreement effective until the beginning of 2012. In order to cover the losses in excess of capital and dividend payments, the GSEs drew $188 billion from Treasury, the figure cited by Epstein. Starting in 2012, both Fannie and Freddie have had positive earnings, which have been paid to the Treasury under the revised agreement. As asserted by Epstein, the GSEs’ dividend payments from 2008 through the end of the first quarter of 2014 will exceed their respective draws.
So have Treasury and taxpayers been fully compensated for the value of their contribution to the GSEs? There are three issues to consider when answering whether $188 billion is sufficient repayment.
First, the use of the $188 billion in senior preferred stock purchases as the measure of Treasury’s support deprives Treasury of any return on its investment from 2012 through 2014. The dividends paid to Treasury during 2008−11 were paid for by increasing the draw (Treasury buying more senior preferred stock) and, hence, included in the $188 billion “repaid” to the Treasury. However, the calculations cited by Epstein effectively apply 100 percent of the GSEs’ payments in 2012 through 2014 to reducing the balance of Treasury’s senior preferred holdings. If instead we assume that Treasury should have continued to earn a 10 percent dividend on its total holdings, then the GSEs would have owed substantially more dividends. Although a calculation of exactly how much more dividends is beyond the scope of this commentary, back-of-the-envelope calculations show the amount that should have been recorded exceeds $18 billion in 2012 alone.4 Further dividends would also have been due to the Treasury in 2013 and 2014.
Second, the calculations only compensate Treasury and the taxpayers for the losses that they bore and include no payments for the implicit capital they provided to the GSEs. That is, Treasury not only supplied funds to the GSEs but promised to bear any future losses, ultimately agreeing to provide unlimited funds to cover whatever losses were incurred in the future. Fannie and Freddie and their supporters certainly should recognize that committing to cover losses is valuable even if the losses do not arise—one of the GSEs’ primary business activities involves the provision of just such a guarantee on the mortgage-backed securities they issue.
The value of the Treasury contribution in promising to cover any additional losses on the GSEs’ pre-conservatorship loans could be measured as the fees demanded to cover such risks (as with Agency MBS or credit default swaps), or it could be measured as the expected return on the implicit capital that Treasury provided to the GSEs. Ideally, one would want to consider several different approaches to estimating the value and then take some sort of average, a task far beyond the scope of this post. A simpler task would be to get some sense of the perceived extent of the commitment during the crisis, for instance, by looking at the Treasury’s estimates of its exposure at that time. The U.S. government’s 2009 Financial Report (pages 81−82) notes that the Treasury had purchased $95.6 billion as of September 30, 2009. However, the Treasury estimated that its remaining liability under the senior preferred stock agreement ranged from a “best case” scenario of $76.9 billion to an “extreme case” of $206.7 billion, noting that “no value within the range is a better estimate than any other amount.” Thus, the Treasury estimated that a reasonable upper bound on its total preferred stock purchases was about $300 billion.
A possible counterargument to the above analysis is that Treasury imposed an excessively high dividend rate of 10 percent for the senior preferred stock. In order to provide a rough approximation of what might be reasonable, I looked at economic data from the St. Louis Fed on the yield on high-risk bonds as proxied by the Bank of America Merrill Lynch U.S. High Yield CCC or Below Effective Yield (see chart). According to Standard & Poor’s, a rating of CCC indicates the bond is still paying interest but is “currently vulnerable and dependent on favorable business, financial and economic conditions to meet financial commitments.” That seems like a generous description of the condition of Fannie and Freddie at the time they were put into conservatorship.
What we can observe in the chart is that the 10 percent rate set in the senior preferred stock agreement is rather favorable to the GSEs over this period. The bond rate was over 15 percent at the time the GSEs were being put into conservatorship and the rate was being set on the senior preferred stock. The rate subsequently spiked at around 40 percent, and dropped below 10 percent only for brief periods prior to the latter part of 2013. [Author's note: Epstein points out that the GSEs also granted a warrant for the purchase of 79.9 percent of their common shares for a minimal price to the Treasury. However, at the time these warrants were granted, the GSEs were in financial distress and could not have survived without assistance. Thus, it is not clear these warrants had much value when granted.]
Thus, looking at the value of Treasury’s contribution relative to the GSEs’ pre-conservatorship business, we find that Treasury’s $188 billion senior preferred stock holdings understate the value of its contribution by attributing no dividends to Treasury in 2012−14, by ignoring the value of Treasury’s promise to absorb all of the GSEs’ losses, and by using a senior preferred stock dividend rate that likely provided inadequate compensation for the risk Treasury bore as a creditor.
Postcrisis: government bore all of the risk of the new businesses
As we have seen, Fannie Mae’s and Freddie Mac’s pre-conservatorship portfolios generated major losses. Yet the GSEs have recently been able to generate large profits that allowed them to pay dividends equal to a substantial part of their draws from the Treasury. How did they do this? The primary reason is that a large fraction of their current portfolios consists of profitable business generated since the beginning of 2009. Indeed, Fannie Mae (page 2) reports that 77 percent of its book of single-family mortgage guarantees has been acquired since the start of 2009.
Moreover, the government’s role in supporting the GSEs is not like debtor-in-possession financing of inventory for a manufacturer where the manufacturer’s value-added comes from processing raw materials into final goods. The primary value-added of the GSEs is their promise of timely payment of principal and interest on the mortgage securities that they guarantee; and the sole reason that this promise has value in the post-conservatorship world is the implicit capital provided by the Treasury. As Alex J. Pollock states:
“Fannie and Freddie’s most important business is guaranteeing mortgage-backed securities (MBS). Even with the government’s $187 billion investment in them, their net worth has been zero, without it, hugely negative. What is the value of a guaranty from a guarantor with hugely negative capital? Zero. It is solely the fact that the government guarantees Fannie and Freddie’s obligations that gives this business any revenue or profit at all.”
Thus, as an economic matter, the post-conservatorship dividends paid to the Treasury did not arise from risks being borne by private shareholders. Rather, as pointed out by Felix Salmon, the payments came from—and can reasonably be considered a return on—the risks being borne solely by the Treasury (and taxpayers).
The claim that the taxpayers and Treasury have been fully repaid for their support of Fannie Mae and Freddie Mac is based on an accounting calculation that does not withstand economic analysis. The claim that Treasury’s commitment has been fully repaid attributes no dividend payments to Treasury starting in 2012, attributes no value to the government guarantee to absorb whatever losses arose in the pre-conservatorship book of business, and arguably reflects Treasury setting too low of a dividend rate on its senior preferred stock. Moreover, the profits that are being used to pay the dividends did not arise from the contributions of private shareholders but rather entirely reflect risks borne by the Treasury and taxpayers. Thus, the Treasury claim that the value of the aid was “incalculable” is an exaggeration; the value surely can be fixed within reasonable bounds. However, the implication of this claim, that the GSEs cannot repay the economic value on behalf of their common shareholders, is nevertheless accurate.
What should happen next to Fannie Mae, Freddie Mac, and government involvement in residential real estate finance is the subject of an ongoing debate, with a new bill introduced in the Senate earlier this month. I have considerable sympathy for the argument that Fannie and Freddie should not be kept in their current status. On the other hand, W. Scott Frame, Lawrence J. White, and I surveyed the major “reform” proposals and concluded that these proposals will result in taxpayers again being subject to very large risks in a future residential real estate downturn.
Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Kris Gerardi, Scott Frame, Paula Tkac, and Larry White for helpful comments and Brian Robertson for research assistance. The views expressed here are the author’s and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please e-mail email@example.com.
2 The GSEs returned part of the subsidy in the form of lower interest rates to mortgage borrowers. However, economic theory predicts that part of an input subsidy will be used to lower price and increase quantity demanded by a profit-maximizing firm, as I have previously explained. Thus, evidence that the GSEs lowered mortgage rates is not proof that they sacrificed profits to do so.
3 Nevertheless, whenever the topic of an implicit guarantee was raised in public policy debates, the two GSEs adamantly denied the existence of any such subsidy. Indeed, W. Scott Frame and I were treated to a lecture by a high-level Fannie Mae officer who asserted that no such subsidy existed and demanded that we rewrite a draft of our 2002 paper to take out any claims of a subsidy.
4 First, assume that all of the GSEs’ 2012 payments could have gone to repurchasing Treasury’s senior preferred stock at the start of the year, reducing the balance to $168 billion. At a dividend rate of 10 percent, the GSEs would then have owed $16.8 billion in dividends to Treasury. But this leaves only $2.2 billion of the $19 billion payment available to repurchase senior preferred shares. Thus, we can safely conclude that the balance could not have been reduced below $185.8 billion and hence that at a 10 percent rate, the required dividend would have exceed $18 billion.
W. Scott Frame and Lawrence J. White, 2005. “Fussing and Fuming over Fannie and Freddie: How Much Smoke, How Much Fire?” Journal of Economic Perspectives 19, 2 (Spring): 159–184.