GSE Subsidy Abuse: The History, How Conservatorship Ended It, and the Risk That Re-Privatization Could Enable Its Return

Fannie Mae and Freddie Mac (F&F), the two large government-sponsored enterprises (GSEs) that have long been the dominant source of American single-family residential mortgage credit, began facing criticism in the late 1990s for abusing the advantages given to them by Congress.  Those advantages, most notably government support to their creditworthiness, were intended to generate subsidies to help keep mortgage rates low for homeowning borrowers. Instead, they were also used to benefit F&F’s shareholders and management on a very large scale. [1] Many in the industry argued this was a clear abuse of those congressionally endowed advantages.

The criticism grew so loud that the Federal Reserve and President George W. Bush’s administration joined together in the mid-2000s to propose new legislation aimed at reigning in what I, along with many others, considered a significant instance of subsidy abuse. 

This article will describe the history and nature of the abuse that occurred, provide a rough estimate of its dollar size prior to conservatorship, and explain how actions taken during conservatorship effectively ended it. Additionally, it will show why we must guard against the return of large-scale subsidy abuse if F&F are re-privatized and exit conservatorship, which has been a much-discussed topic since President Trump was elected to a second term. 

Understanding the history of GSE subsidy abuse

Upon its privatization into a GSE in 1968, Fannie Mae was given several advantages by its congressionally-given charter.[2] The objective was to keep its funding costs as low as possible, ideally very close to what they would have been if there had been no privatization, while also removing their debt from the federal budget.[3] The intent was to minimize any increase in mortgage rates as Fannie Mae was transformed from a government agency funded by Treasury to a privately-owned entity relying upon public capital markets instead. 

Later, Freddie Mac was created and given the same advantages. Two of those advantages turned out to be quite consequential, as F&F’s management teams exploited them not just to keep mortgage rates low, as intended, but also to produce extra profits for the benefit of shareholders and management.[4]

• The first advantage: the very visible “implied guarantee” 

In their charters, F&F were each given a $2.25 billion line of credit to Treasury, an advantage not available to conventional private sector mortgage lenders. The mortgage markets recognized both the financial significance and the symbolic importance of these lines of credit. They noted the other types of special treatment F&F received compared to other private companies,[5] and additionally considered the massive disruptions that could arise if either of the two companies were to fail. As a result, they concluded that government would not allow F&F to fail, even without a formal guarantee.[6] This became known as the “implied guarantee.” Consequently, the two companies were able to secure favorable pricing for the mortgage-backed securities (MBS) they issued, as well as access funds at near-Treasury interest rates for their non-MBS (i.e., unsecured, general obligation) debt. 

This arrangement did indeed work to keep mortgage rates low, as originally intended, because MBS issued by F&F were charged an interest rate by the capital markets that reflected the implied guarantee described above.[7]

At some point, F&F realized they also could exploit the implied guarantee to fund large, discretionary investment portfolios. They issued non-MBS debt and used it to fund an investment portfolio consisting of various types of mortgage-related bonds, including their own MBS. [8] In this way, F&F could monetize the below-market funding cost that resulted from the implied guarantee without limit. This meant that the larger the size of the investment portfolio, the more they could abuse the implied guarantee. The earliest readily available data show Freddie Mac, as an example, took extensive advantage of this opportunity, with its mortgage investments portfolio equaling 47 percent of its total outstanding MBS issued.[9] By around 2005, at their height, F&F’s combined investment portfolios amounted to almost $1.6 trillion.[10]

The potency of the implied guarantee subsidy to the profits generated by F&F’s discretionary investing is evident in their financial statements for 2005. Out of their total pre-tax profit of $10.1 billion that year, just over half – or $5.4 billion[11] – is estimated to have come from the implied guarantee subsidy alone.[12] In fact, from the mid-1990s until just before they entered conservatorship, half or more of F&F’s profits are believed to have come from these investing activities.[13]

This all got so out of hand that the Federal Reserve began to view it as a problem that needed to be addressed. Under Chairman Alan Greenspan, the Federal Reserve began to publicly point out that such large investment portfolios funded by non-MBS debt were posing systemic risks.[14] The Bush administration then collaborated with the Federal Reserve to propose legislation that would limit the size of F&F’s investing activities.[15] This push failed in July 2005 due to political opposition in Congress,[16] a reflection of the immense lobbying power of the two GSEs.

The success of their lobbying efforts was not totally surprising. F&F had also used a portion of their profits, enhanced by those abused subsidies, to create a lobbying powerhouse in Washington. In 2005, for example, the combined spending of F&F ranked them among the five largest organizations in terms of lobbying expenses.[17] This is a great example of a phenomenon commonly referred to as “rent seeking behavior,” and more specifically, the “subsidy lobbying cycle,” a somewhat obscure term that nevertheless describes this situation quite well.[18]

• The second advantage: the less visible subsidy resulting from an intentionally understated capital requirement, combined with the implied guarantee

In 1992, Congress passed the Safety and Soundness Act (1992 Act), which set the minimum capital requirements for F&F. The Act required F&F to hold capital of at least 2.50 percent of assets against mortgage assets they held on their own books, and just 0.45 percent against mortgages that were securitized and guaranteed by F&F. 

The 1992 Act’s capital requirements were surprisingly low given that the international bank capital standards established in 1988, just four years earlier and known as the Basel accords, mandated that residential mortgages carry a 4 percent capital requirement.[19]

In reality, however, the amount of capital held by F&F was irrelevant to the capital markets. As noted by Chairman Greenspan in his testimony to Congress, the reliance by creditors on the implied guarantee meant that there was no market discipline[20] on F&F.  In fact, the corporate credit ratings for F&F at that time were equal to those given to Treasury securities because of the implied guarantee. So, F&F did not need to worry about keeping risks and leverage low to convince its creditors to purchase its debts at low interest rates. Instead, the capital requirements for F&F were solely an “inside the Beltway” issue, which meant they were subject to all the usual lobbying pressures.[21]

The 1992 Act reflected those lobbying pressures. First, its 2.50 percent requirement for mortgages held by the GSEs was modestly (i.e., by 3/8ths) under the 4.00 percent Basel standard, reflecting a moderate subsidy towards low mortgage rates. But the 0.45 percent requirement on mortgage loans funded by MBS issued and guaranteed by F&F – long their largest asset class – was so out of line with the Basel standards that it is hard to understand the logic behind it. With the benefit of hindsight, it was a giant policy mistake.[22]

Such undercapitalization was primarily raised by F&F’s critics as a safety-and-soundness issue, meaning whether F&F had enough capital to withstand significant market stresses without risking their own failure. However, behind the scenes, it was also considered a major subsidy issue. In 2005, the subsidy from undercapitalization was estimated to be $2.2 billion just on the investment portfolio, which was available to further benefit shareholders and management.[23] On F&F’s $2.57 trillion then-outstanding MBS, the subsidy was much larger due to the unduly low 0.45 percent capital requirement: $9.1 billion.[24] How much of the latter was passed through to homeowners, rather than shareholders and management, is impossible to determine. 

Thus, the estimated abused subsidy from the excessively understated capital requirement – assuming that 50 percent of the $9.1 billion associated with the guarantee book did not contribute to keeping mortgage rates low – was approximately $6.8 billion.[25]

Between the very visible implied guarantee subsidy and the behind-the-scenes subsidy created by undercapitalization, the total abused subsidy enjoyed by F&F in 2005, the last year before the impacts of the mortgage bubble started to damage their finances, was thus about $12.2 billion pre-tax. [26] There is a possibility of some double-counting due to the large amount of F&F MBS being held in their investment portfolios, so perhaps a rough but better estimate would be in the $10 billion range for that year.

The confirmation of such a large abusive subsidy that seemingly went to shareholders and management instead of homeowners can be easily seen. For example:

• Stockholder returns were excessive. For the four years ending 2005, Fannie Mae’s after-tax return on equity (ROE), a key measure of profitability for a financial institution, averaged 19.7 percent, while Freddie Mac’s averaged 17.9 percent. These figures are almost double the market level, which is generally around 9 to 10 percent. For a GSE, which has a market franchise granted by Congress to serve a public policy purpose, such high returns seemed rather inappropriate.

• Executive compensation was excessive. A highly publicized example was Franklin Raines, Fannie Mae’s long-time CEO, who earned $91.1 million in total compensation from 1998 to 2004,[27] or well over $10 million per year (equivalent to about $18 million per year in today’s dollars).[28] This level of compensation seemed inappropriately high for a company that was highly subsidized and operating a government-endowed franchise with only one competitor.[29]

Subsidy abuse was eliminated during conservatorship

In conservatorship, the Federal Housing Finance Agency (FHFA), the regulator and conservator of F&F, took over total operating control of the two companies from their shareholders. As part of conservatorship policy, lobbying was expressly forbidden; F&F executives could not meet with elected officials without FHFA permission, and any such meetings were also attended by FHFA personnel to ensure no lobbying occurred. Campaign contributions and similar expenditures were also prohibited. These restrictions completely changed the political dynamic around the abuse of F&F’s subsidies.

• The visible abuse of the implied guarantee subsidy to fund F&F’s large investment portfolios was immediately addressed by Treasury. When F&F entered conservatorship in September 2008, Treasury replaced the implied guarantee with a formal support agreement to help restore market confidence.[30] In that agreement, Treasury required F&F to immediately begin reducing their investment portfolios,[31] scaling them down over ten years to a final limit of $250 billion[32] each, i.e., about two-thirds less than their peak levels before conservatorship. This was later amended to reduce the amount to $225 billion each. The FHFA as conservator also ordered further reductions, and today the investment portfolio of each GSE is just under $100 billion, down by roughly seven-eighths from its peak level.[33] As a result, the issue of subsidy abuse today has been effectively addressed and resolved. 

• The excessively low capital requirements were later replaced by a proper safety-and-soundness capital requirement, eliminating the subsidy that flowed from it. When F&F entered conservatorship, the 1992 Act capital requirement (i.e., the 2.5 and 0.45 percent, respectively) was suspended by the FHFA. Nothing immediately replaced it. 

The FHFA’s initial effort to develop a proper, and not intentionally undersized, capital requirement took place in the mid-2010s. At that time, the FHFA developed the “conservatorship capital framework” (CCF), which was implemented in 2016-2017 for use inside conservatorship. It was later proposed as an official regulatory minimum capital requirement. After revisions based on public feedback,[34] the proposal called for about $180 billion of capital.[35] This was about six times as large as what the previous 1992 Act requirement – estimated at $25 to $30 billion – would have been at that time. The difference is dramatic, but not surprising given the history.

However, the CCF-based capital proposal was never finalized. Instead, a much higher requirement referred to as the Enterprise Regulatory Capital Framework (ERCF) was established in 2020 by the FHFA under a different director. It received significant criticism for being excessively large, estimated at over $300 billion, or roughly more than ten times the size of the 1992 Act calculation. 

Thus, the inadequate capital requirement of the 1992 Act has been officially replaced, and the resulting behind-the-scenes subsidy has thereby ended as well. 

So, conservatorship has successfully put an end to the GSEs’ long-standing abuse of the subsidies gained from their advantages. This has been achieved through measures taken by Treasury in its support agreement and by the FHFA through regulatory capital requirements. The subsidies still in place are now strictly used for their intended purpose: to make mortgages more affordable. No funds are directed toward excessive executive compensation,[36] excessively high shareholder returns, or sizeable lobbying and political influencing expenditures.

Pressure to reinstitute large-scale subsidy abuse could return with re-privatization

Ideally, the two reforms described above, which have put an end to the historic subsidy abuse, would become enshrined in legislation. This would make them rather difficult to change, even if the companies were re-privatized. However, the likelihood of such legislation being passed is considered extremely low. Thus, the reforms remain vulnerable to changes depending on the policy and political views of whoever is in the White House, which oversees both Treasury (responsible for the investment limit) and the FHFA (responsible for the capital requirement). 

Obviously, if and when F&F exit conservatorship, many things about them will be revisited and very possibly recast. The challenge then is not to lose the abuse-preventing reforms in the process, either partially or wholly, intentionally or unintentionally. Realistically, there are signs that memories of the past are fading, and voices are beginning to talk again about how homeownership could be helped by undoing, at least to some extent, the reforms achieved during conservatorship. 

For example, a leading Wall Street investor has proposed that the capital requirement from the ERCF be reduced by over 40 percent.[37] While this leaves the requirement far above the level set by the 1992 Act, it would be roughly 25 percent lower than even the revised CCF proposal described earlier.[38] It thus represents a potential slippery political slope towards the deliberately low capital requirements of the past that were designed to “help homeownership.”[39] As the discussion above has demonstrated, it would be a significant policy mistake to once again set F&F’s capital requirement on anything other than maintaining safety-and-soundness, even if it would help homeownership. Given what happened to F&F during the Great Financial Crisis, it would be highly irresponsible not to have them fully capitalized, as no one wants another government rescue to be needed. 

The possibility of allowing a large discretionary investment portfolio is also being discussed again, particularly in terms of how it might offset aspects of a conservatorship exit that some fear might push guarantee fees higher. While this change would not have the side effect of rendering F&F too weak to withstand future market pressures,[40] a major concern remains: it is unclear how to ensure the resulting subsidy would actually end up primarily benefiting homeowners instead of shareholders and management, as it did pre-2008.[41]  In short, today’s tight investment limits would seem to be best left in place for the long term.[42] 

The Trump administration, which is the actual decision-maker in re-privatization planning for the next three-plus years, has, to date, made only a few comments about how the process would exactly work, and none about how it intends to prevent the return of subsidy abuse. But there will likely be the very same historic lobbying pressures to “help homeownership” by undoing to some degree the reforms that ended subsidy abuse. This means reformers will need to stay vigilant against any such backsliding – whether intentional or not – as part of a plan for conservatorship exit, regardless of whether it’s under the Trump administration or a future one.

 

Footnotes

[1] At the same time, in contrast, F&F were widely admired and praised for their securitization-based business model, which they developed in the 1970s and 1980s. Many credited this model with making the “American” mortgage – characterized by a 30-year term, fixed interest rate, full self-amortization, and the ability to prepay at any time for any reason without penalties – sustainably available to the broad mass of U.S. homeowners at relatively low cost.

[2] The laws that created the privatized Fannie Mae, and later Freddie Mac, are known as their “charters.” These charters outline the obligations, restrictions, and benefits that apply to the companies, ensuring they serve their intended public policy goals as set by Congress despite being privately owned.

[3] The objective of the Johnson administration in privatizing Fannie Mae in 1968 was to relieve the “guns and butter” pressure on the federal budget. (“Guns” refers to Vietnam War spending, “butter” refers to Great Society social spending.) Thus, they wanted Fannie Mae to enjoy as low a funding cost as possible while still qualifying as “off budget,” i.e., not part of the federal balance sheet. 

[4] It’s still an open question whether some portion of the extra profits was redirected into additional subsidy to keep mortgage rates low. Based upon my experience as the former CEO of Freddie Mac, where I saw the internal mechanisms that assigned profit to different activities, my view is that few, if any, of the extra profits went to additionally subsidize mortgage interest rates. 

[5]  For example: (1) GSE debt is exempt from certain U.S. Security and Exchange Commission (SEC) registration and reporting requirements that apply to private sector companies but not government debt issuers; and (2) the Federal Reserve is able to purchase GSE debt for its open market operations, which otherwise are restricted to government debt. These examples show how GSE debt was, and still is, treated broadly as government debt rather than private sector debt by a variety of government entities. 

[6] The market’s assumption that it could rely on the implied guarantee was tested by the mortgage bubble bursting in 2007 and 2008. The assumption was fully validated when in 2008 the government placed the two companies into conservatorship on terms that held their debtholders harmless. 

[7] F&F then added to this rate their guarantee fee, and the total of the two is what they charged primary market lenders for mortgages. This operational mechanism ensured that the low MBS interest rates that benefitted from the implied guarantee were passed through fully to primary market lenders. It was assumed that competition among such lenders would result in almost all of these benefits being passed on to homeowners.

[8] Purchasing back their own MBS also had the effect of undoing the securitization, irresponsibly placing large interest rate and liquidity risks onto F&F’s balance sheet – the risks that securitization was designed to transfer to investors. Later, the Federal Reserve would begin to view this as a systemic risk, as such risks had already wiped out the long-standing savings and loans business model during the 1980s. These purchases may have also slightly reduced the MBS interest rate as a byproduct. 

[9] See Freddie Mac’s 2002 Annual Report, page 27, for year 2000 historic data (https://www.freddiemac.com/investors/financials/pdf/10k_022703.pdf). The MBS outstanding was $838 billion and the mortgage investments portfolio (of which over 80 percent were securities investments) was $392 billion. 

[10] To give a sense of scale, that near $1.6 trillion was about twice the size of the Federal Reserve’s balance sheet at that time.

[11] When I became CEO of Freddie Mac, I had an analysis done of investment portfolio profits. It showed that virtually all the profit came from the cheapness of the non-MBS debt issued to fund it, with little if any coming from investing prowess. 

[12] The Congressional Budget Office, in a 2024 report on the Federal Home Loan Banks, or FHLBs, (see https://www.cbo.gov/system/2024-03/59712-FHLB.pdf ), disclosed its estimate of the implied guarantee subsidy as being worth, as the center point of a range, 0.40 percent on borrowings. Assuming 85 percent of the nearly $1.6 trillion in outstanding investment portfolio assets were discretionary, this translates into a $5.4 billion subsidy that year. Note that F&F prior to conservatorship did not pay anything for their government support, making this subsidy essentially a gift from the taxpayer.

[13] For 1993 to 1995, see https://www.fanniemae.com/media/26996/display, page 34. This continued to be true until just before conservatorship. 

[14] A lengthy but very  thorough explanation of the Federal Reserve’s view can be found in testimony submitted by Chairman Greenspan in 2004 before a Senate Committee. See:  https://www.federalreserve.gov/boarddocs/testimony/2004/20040224/default.htm.

[15] See a Congressional Research Service article from 2005 on this very topic, https://www.policyarchive.org/handle/10207/4255#:~:text=Federal%20Reserve%20Chairman%20Alan%20Greenspan,ever%20have%20to%20be%20made.

[16] See https://georgewbush-whitehouse.archives.gov/news/releases/2008/10/20081009-10.html, and also a United Press article https://www.upi.com/Business_News/2005/07/28/Dems-rip-new-Fannie-Mae-regulatory-measure/20231122581039/.   

[17] See https://www.opensecrets.org/federal-lobbying/top-spenders?cycle=2005.

[18] There is a large amount of literature describing why it is so hard for subsidies, once established, to be reduced or eliminated.

[19] The Basel accords established an 8 percent capital requirement (of which a portion did not need to be common equity) for assets defined as having a 100 percent risk weight. Residential mortgages were classified as having a 50 percent risk weight, which resulted in a 4 percent capital requirement for those assets.

[20] “Market discipline” refers to a process in which a company’s creditors – such as the banks that loan it money and the investors who purchase its bonds – set their terms based on their own perceptions of the company’s creditworthiness. For example, if they believe the company is riskier, they may demand higher interest rates. This evaluation happens independently of what regulators may say about it. 

[21] One source of such lobbying pressure came from industry groups, such as homebuilders and realtors, who wanted low mortgage rates to enhance their own business prospects. Another source was advocates for low mortgage rates to help expand homeownership as broadly as possible. It is conventional wisdom that, prior to 2008, these two sources of advocacy together were very influential in housing finance policymaking. 

[22] The 0.45 percent capital requirement was also possibly related to the fact that F&F used an accounting method at that time that treated the MBS they issued and guaranteed as an off-balance sheet item. Nevertheless, because F&F guaranteed the credit risk of the underlying mortgage loans to MBS investors, F&F still bore the same credit risk as if they owned the mortgage loans outright. This off-balance sheet treatment was later deemed by the SEC to be incorrect and required years of restatements of financial reports. At that time, the CEOs of both F&F lost their jobs over what is generally referred to as their “accounting scandal.”

[23] This is calculated as a balance of $1.45 trillion in the combined investment portfolio; a capital requirement of 1.50 percent lower than necessary (which is derived from the 4.00 percent Basel requirements minus the 2.50 percent requirement set by the 1992 Act); and an assumed a pre-tax cost of capital of 15 percent, (i.e., what shareholders would expect to earn) minus the cost of borrowing the funds that would replace the unrequired capital at 5 percent (which is a net cost of capital of 10 percent).

[24] This is calculated as a balance of $2.57 trillion in outstanding MBS owned by third parties that is too low by 3.55 percent (the Basel 4.00 percent minus the 0.45 percent established by the 1992 Act). A net return of 10 percent, as defined immediately above, is assumed to be the net cost of capital. 

[25] This equals the $2.2 billion plus half of the $9.1 billion, for a total of $6.8 billion.

[26] This is equal to the $5.4 billion estimated subsidy on the implied guarantee plus the $6.8 billion estimated immediately above due to an excessively low capital requirement, for a total of $12.2 billion.

[27] See a report from the Office of Federal Housing Enterprise Oversight (OFHEO):  https://www.fhfa.gov/news/news-release/ofheo-files-notice-of-charges-against-former-fannie-mae-executive-franklin-rains-timothy-howard-and.  

[28] Based on the Consumer Price Index (CPI) from 2002 versus 2024, which was 79 percent higher.

[29] The boards of F&F, as well as the boards of the 11 FHLBs, determined their top executives’ compensation based on what they believe comparable firms pay. However, the comparisons appear to be skewed upwards, as they primarily focus on asset size to define a comparable firm. All GSEs – F&F and the FHLBs – are large in assets but they do not have a comparable scale in terms of other very relevant measures: the number of employees, product lines, or locations. For example, firms that are similar in assets to F&F would typically have at least 10 times more employees and dozens of major product lines, rather than just two or three.

[30] This took the form of a “net worth keepwell” agreement by which Treasury agreed to put in equity, in the form of senior preferred shares, to bring the net worth of Fannie Mae or Freddie Mac back up to zero if either ever became negative. This was immediately effective in restoring the market confidence in F&F that had been lost. 

[31] Recall that Treasury at that time was part of the W. Bush administration, which had attempted three years earlier, along with the Federal Reserve, to limit the size of F&F’s investment portfolio. Thus, the administration got a “second bite at the apple” and unsurprisingly seized the opportunity via the support agreement F&F desperately needed at that time to restore market confidence.

[32] The $250 billion limit was meant to allow the legitimate use of the investment portfolio to support F&F’s securitization and guarantee business. A 10-year timeframe was set to prevent F&F from having to sell mortgage securities at distressed prices in an already fragile market, which would have only driven mortgage rates higher.  

[33] This is probably less than needed to support the securitization and guarantee business, but it has proven workable given the FHFA program where F&F are required to sell off modified and defaulted loans to investors instead of keeping them in their portfolios. 

[34] The biggest criticism was the absence of a countercyclical capital buffer. In response to this, an additional $40 billion was included in the revision to address the issue. 

[35] See FHFA’s Federal Register entry proposing the capital rule, after revision, in Table 5. https://www.govinfo.gov/content/pkg/FR-2018-07-17/pdf/2018-14255.pdf

[36]  The FHFA, as conservator, has strict limits on executive compensation, and by law, F&F’s CEOs can only make $600,000 per year. 

[37] Bill Ackman, the famous and recently outspoken investor, has proposed reducing the capital requirement from 4.25 percent (which he calculates is the level that the complex ERCF formulae currently requires) to 2.5 percent, which is 41 percent lower. 

[38] That proposal called for $180 billion using F&F’s 2017 financial statements. Adjusting for F&F’s larger size today, it would be about $255 billion, or about 3.3 percent of assets. The 2.50 percent called for by Mr. Ackman is thus about one-quarter lower.

[39] Such a low capital requirement would also, of course, make it possible for F&F to achieve “full” capitalization sooner, which would make it easier to exit from conservatorship earlier, allowing investors in their equity securities to make potentially large gains more quickly. 

[40] As noted previously, if F&F were allowed to purchase their own MBS for their investment portfolio again, it would bring back the risks of interest rate and liquidity mismatch risk onto their balance sheets. 

[41] The implied guarantee is often talked about as if it were a classic free lunch. However, it isn’t one – near-Treasury securities by definition compete to attract the investors who also purchase Treasury securities. Thus, F&F issuing non-MBS debt that carries the implied guarantee causes Treasury rates to be higher than would otherwise be the case. An estimate of this impact can be found at: https://www.aei.org/wp-content/uploads/2025/04/Estimating-the-Treasury-Savings-by-Eliminating-Agency-MBS-Final.pdf?x85095. The specific estimated impact seems questionable to me, but the assumption that there is some impact is sound.

[42] I would, however, convert them to be a percentage of assets, rather than a fixed amount, to accommodate F&F growth or shrinkage over the long term.