The Unfinished Business of GSE Systemic Risk: How Two of Its Three Major Sources Have Been Contained (Part 1 of 2)

 

Introduction

The Great Financial Crisis (GFC) was the most significant financial collapse and economic downturn since the Great Depression. It led regulators, central bankers, and the financial industry itself to think much more deeply about the causes of financial instability and systemic risk[1] that played such a significant role in the crisis. However, the long list of resulting changes implemented afterwards — which included legislation and many revised and new regulations — largely neglected Fannie Mae and Freddie Mac (F&F), the two large government-sponsored enterprises (GSEs) that had been at the heart of the GFC.[2]

That neglect made sense because the Obama administration intended that the GSEs would be “wound down” and replaced by something to be developed by Congress. But that plan never materialized. By about 2017, with Congress not focusing on a replacement and also in part due to F&F’s much-improved operating performance and risk profile, it had become a prevailing view among policymakers that the companies would continue to anchor the U.S. housing finance system for the foreseeable future. This made it urgent for an eventual exit from conservatorship to happen only after the key weaknesses[3] in their pre-conservatorship business model[4] were fixed. Unfortunately, by that time, financial regulators had moved on to other issues. This created a gap in a comprehensive post-GFC strategy to reduce systemic risk in the U.S. financial system.[5]

In Part 1 of this two-part series, I examine several of those weaknesses, rooted in the structure Congress created for the GSEs, and how this inadvertently led to three specific sources of potential financial instability. The article then reviews how two of the three identified sources of financial instability — (1) the excessive concentration of mortgage interest rate and liquidity risk, and (2) significant undercapitalization — have effectively been contained. This was mainly achieved through actions taken during conservatorship, a significant policy success that is rarely discussed in the industry or government.[6]

The third major risk that could lead to financial instability for F&F — the excessive concentration of mortgage credit risk — remains far from resolved despite advances made between 2013 and 2019, and is in fact heading in the wrong direction. This is the bad news that will be discussed in Part 2.

Instability as the unintended consequence of the GSE charters

Before the Great Depression, the U.S. government was little involved in housing finance.  However, starting in 1932, it began to play a significant role, and by the late 1930s, it arguably became its dominant force and has maintained that position since. 

To implement this intervention, the federal government mostly created, both directly and indirectly, a series of specialized financial institutions focused on residential mortgage lending. Such institutions included savings and loans, which were the mainstay of residential mortgage lending in the immediate post-war decades; F&F, which took over this role through the 1980s; the Federal Home Loan Banks (FHLBs); the Federal Housing Administration (FHA); the Federal Savings and Loan Insurance Corporation (FSLIC), a counterpart to the more widely known FDIC; and others.[7] In essence, Congress created a parallel banking system dedicated to residential housing finance. The purpose in doing so was to allow homeownership policy to be implemented in a focused way, and in particular to effectively channel subsidies to “help homeownership,” a phrase still frequently heard in Washington. Unfortunately, one type of subsidy took the form of relatively light capital requirements on the various specialized mortgage institutions created in order to make mortgages less expensive.

This policy approach, unfortunately, left the issue of stability for these financial institutions insufficiently considered or addressed. Organizations with concentrated risk — in this case, in mortgage assets — are subject to higher levels of instability. In contrast, diversification of mortgage assets across the broader banking and financial system would have been considerably more stable. In addition, while a light regulatory capital requirement may have made mortgages more affordable, it also left those specialized mortgage institutions unduly vulnerable, as large losses could rapidly escalate into a crisis.[8]

In addition, after World War II, the standard mortgage in the U.S. became what is today known internationally as the “American mortgage.” Such a mortgage has three major features:

• A long-term repayment period, initially set at 15 years but later standardized at 30 years during the 1960s.
• An interest rate that is fixed for the same time period, allowing equal monthly payments that fully pay off the loan balance by maturity, known as full self-amortization.  
• Borrowers can make prepayments at any time for any reason without penalty.

This is an extremely borrower-friendly structure,[9] and grew out of a government agency that did not need to worry about liquidity (since its funding came from Treasury as needed) or interest rate risk (since any losses would be absorbed by the federal budget).[10]

However, when this same structure became standard for private sector lenders after World War II, including those created directly by the government (like F&F after their privatization) or indirectly (like the savings and loans associations), the resulting liquidity and interest rate risk became quite problematic. Simply put, at that time, no lender funded by deposits or the debt markets had access to funds to match the liquidity or interest rate risk profile of a balance sheet dominated by American-style mortgages.[11] Instead, those institutions would be short-funded, creating a tremendous concentration of interest rate and liquidity risk, a source of potential major financial instability.

To summarize, F&F had a propensity towards major financial instability because their legislative charters, designed to enable them to implement housing policy in a focused manner and channel subsidies efficiently, contained three major flaws. Those flaws were: (1) an undue concentration of interest rate and liquidity risks associated with the American-style mortgage, (2) an undue concentration of mortgage credit risk, and (3) inadequate capital requirements given the risks they took.[12] This situation is a classic example of unintended consequences. 

Concentrated GSE interest rate and liquidity risk becomes contained

Chronologically, concentrated rate and liquidity risk were the first source of GSE financial instability to be contained. Unfortunately, there was a significant episode of backsliding that required interventions years later during conservatorship to fully re-establish the risk being contained. 

In 1970, Ginnie Mae completed the first “pass-through” mortgage securitization,[13] allowing investors who purchased mortgage-backed securities (MBS) to bear all the interest rate and liquidity risks generated by a pool of American-style mortgages.[14] Those particular investors, however, did not want credit risk mixed in, and so Ginnie Mae provided a guarantee against credit loss.[15] In the interest of having a more stable balance sheet and also steadier earnings, Freddie Mac adopted such pass-through MBS as its standard funding method beginning in 1971, with Fannie Mae following in 1981. Their guarantee that investors would have no credit losses, since it was backed by the “implied guarantee”[16] of the government, was considered rock solid, assuring investors that they would not face any credit losses, as desired. This government-backed pass-through MBS business model has been the only successful method of responsibly and sustainably funding on a large scale the long-term and variable maturity nature of the typical American-style mortgage without incurring excessive interest rate or liquidity risks. Thus, by passing those risks to the investors in the MBS, one of the three sources of potential financial instability for the GSEs was successfully contained.[17]

Unfortunately, there was later backsliding to partially reverse that success. F&F built up large investment portfolios consisting of mortgage-related securities, which were funded short; these portfolios also included significant amounts of their own issued MBS, thus taking back onto their balance sheets some of the interest rate and liquidity risks that had previously been passed through to investors. This was all done on a massive scale,[18] with these portfolios becoming, by the late 1990s, roughly half as large as all the loans on F&F’s balance sheets.[19]

As markets lost confidence in F&F in 2008, the government intervened to restore that confidence by placing them into conservatorship and, importantly, by also instituting a formal support agreement to replace the historic implied guarantee. As part of that process, the government insisted, from the first day of conservatorship, that those investment portfolios be reduced in an orderly fashion by about two-thirds; the FHFA later ordered them reduced even further. Today, their value has decreased by about seven-eighths from their peak of nearly $1.6 trillion pre-conservatorship, currently totaling just over $200 billion collectively for F&F.[20]

So, as long as F&F (1) use their “pass-through” securitization and guarantee business model, and (2) have tight limits imposed by the government on their investment portfolios, their tendency towards financial instability caused by the concentration of the American mortgage’s interest rate and liquidity risk should remain successfully contained.[21]

GSE capital requirements: From intentionally low to fully safe and sound

Going into the mortgage bubble years in the early 2000s, the capital requirements for F&F were based on the Safety and Soundness Act of 1992 (the 1992 Act). This law set a capital requirement of 2.5 percent for mortgages that F&F held on their balance sheets without securitization while the requirement was only 0.45 percent for mortgages that had been securitized and guaranteed.

As background, just four years earlier, in 1988, the first international bank capital standards were established, known today as Basel I. This set a capital requirement for home mortgages of 4 percent. In comparison, the 1992 Act’s requirements were moderately low — about 38 percent lower than the Basel I standard— for non-securitized mortgages. This approach was consistent with the longstanding practice of having somewhat low capital requirements apply to mortgage-specialty institutions in order to “help homeownership.”[22]

However, the 0.45 percent requirement on the larger portion of mortgages that had been securitized and guaranteed was so low — almost 90 percent lower than the Basel I standard — that it went beyond any reasonable notion of a targeted reduction in capital requirements to help homeownership. In retrospect, this was a major policy mistake out of line with any realistic risk economics and left the GSEs woefully unprepared to absorb market stresses.[23] This was a major reason they lost market confidence in 2008 and were subsequently placed into conservatorship. 

At that time, the FHFA suspended the outdated requirements of the 1992 Act, considering them irrelevant in those early years of conservatorship.[24] And there things sat for a while. However, having a fully developed capital system is actually necessary for a financial intermediary to run efficiently, a topic which I will discuss in Part 2.  In response, Freddie Mac developed its own modernized capital system in 2013 for use during conservatorship.[25] In 2017, the FHFA developed and promulgated a very similar system, known as the Conservatorship Capital Framework (CCF), which both GSEs were required to use in their internal decision-making, replacing their homegrown systems. The CCF was then proposed in 2018 as a formal, regulatory capital framework. It called then for $180 billion in capital, approximately six times as large as the intentionally low requirement of the 2012 Act at that time.[26]

This formal regulatory proposal was never completed. Instead, in 2020, a different approach called the Enterprise Regulatory Capital Framework (ERCF) was approved by the FHFA as the official regulatory requirement. Today, it calls for over $340 billion in capital.[27] Thus, the amount of capital required is clearly no longer understated to help homeownership but instead has been set based on a prudential regulator’s belief of what is fully safe and sound — a standard that really should always have been in place.[28]

In 2019, F&F’s net worth was not much above zero, as the Obama administration had intentionally de-capitalized them in line with its “wind down” policy. However, in 2019, the first Trump administration switched policy and had F&F begin to retain their earnings to build capital, so that today they have a net worth of $173 billion.[29] While this net worth is now quite substantial, it is still well below what is needed today to meet the minimum specified by either the ERCF ($343 billion) or the smaller CCF (estimated to be $250 billion). Of course, each quarter, more earnings are being retained, which puts F&F on a well-defined trajectory to have a fully safe and sound level of capital. This will then contain and indeed eliminate F&F’s intended undercapitalization as a second source of potential financial instability. 

Conclusion

Part 1 of my article explains how the government’s design of F&F unintentionally led to three major sources of potential financial instability within the two companies. Government has thus some unfinished business to attend to when it comes to systemic risk and the GSEs. It is crucial that the government addresses the three identified issues to prevent a loss of confidence similar to what was experienced during the GFC from happening again.

As discussed above, two of those three sources of instability are now effectively being contained (or are well down that path) thanks in large part to key actions taken during conservatorship. The two contained sources of instability are (1) the excessive concentration of the hard-to-handle interest rate and liquidity risk associated with the American-style mortgages that dominate F&F’s balance sheets, and (2) the intended undercapitalization of the two companies as reflected by the 1992 Act which, while only partially completed to date, is on a defined path for each additional quarter of earnings retention to produce incrementally greater safety and soundness until the full regulatory requirement is met.

This leaves the third source of financial instability — F&F’s undue concentration of mortgage credit risk — to be addressed in Part 2. This risk was on a path to substantial containment that started in 2013, but due to a combination of intended and unintended consequences, began to head in the wrong direction around 2020. Action is needed to return to containing the highly concentrated mortgage credit risk at F&F, i.e., to get back on the path it had been on prior to 2020, in order to address and contain all three sources of major financial instability inherent in the GSEs’ design. This would then complete a comprehensive post-GFC program to reduce the systemic risk in the U.S. financial system. 

Footnotes

[1] In this paper, “financial instability” refers to significant actual or potential losses by a large financial institution (or a group of highly interconnected institutions) that the market considers too great for its capital base, leading to a loss of market confidence in the institution. “Systemic risk” is, in turn, defined by how this loss of market confidence spreads, almost like a chain reaction, to additional financial institutions, creating the risk of major financial distress leading to a significant economic downturn. 

[2] F&F accounted for roughly 40 percent of all single-family first mortgages outstanding in the U.S. at that time. In September 2028, as the market was experiencing a significant loss of confidence in large financial institutions, especially those related to mortgages, they were placed into conservatorship and given stronger taxpayer support to prevent their failure.

[3] Those weaknesses included things such as (1) being allowed to carry mortgage investment portfolios without limit, (2) having very low capital requirements, and (3) not paying a fee for taxpayer support to their creditworthiness. 

[4] Their business model was largely defined by the congressional charters that gave them specific public policy objectives and obligations while still allowing them to be owned by shareholders. The charters also limited their operations to the secondary mortgage market.  

[5] The components of that comprehensive plan include, among other things, legislation requiring stress testing of very large financial institutions under the supervision of the Federal Reserve, and regulations increasing bank capital requirements, especially on the largest banks. 

[6] Interestingly, that success was achieved without any formal master plan to reduce or eliminate the threat posed by the GSEs to the systemic risk of the country’s financial system, but instead mainly as an adjunct to other, related challenges being tackled by the FHFA and the two companies, mostly during conservatorship. However, since some changes, such as investment limits, have not been implemented in legislation or regulation and rely on conservatorship remaining in effect, it is unclear how permanent some of that success can be.

[7] Some others, established during the Depression, did not survive into the post-war period. 

[8] In fact, the first two major financial crises after World War II were both centered on mortgages — the S&L Crisis of 1989 and the Great Financial Crisis of 2008.

[9] In more recent years, the borrower-friendliness of the American mortgage has added additional features. First, it is broadly available with very low down-payments. For example, the GSEs require only a 5 percent down payment, while the FHA requires just 3.5 percent, even without the borrower qualifying for a special program. Additionally, borrowers can lock in their interest rates for up to three months before closing on the mortgage.

[10] This mortgage structure began with the Home Owners Loan Corporation (HOLC), a government agency created in 1933. HOLC was designed to refinance troubled mortgages, which it did by modifying those loans. One key aspect of these modifications was ensuring the lowest possible monthly payment while also ensuring that borrowers would not face an increase in those payments in the future. This approach led to the creation of the first American-style mortgage, which had a 15-year term, considered lengthy for that era. The intention behind this design was to be favorable to borrowers.

[11] In terms of liquidity, lenders like the S&Ls or the GSEs (after they were privatized, but before they adopted securitization) were short-funded. This meant they relied on continually refinancing their debts or maintaining the level of deposits with short or open-ended maturities (such as passbook savings accounts, popular in that era) in order to avoid running out of cash, as their residential mortgage assets typically had considerably longer maturities. Additionally, they struggled to lock in a constant interest rate spread, as their net interest income became unstable due to fluctuations in the average maturity of their loans, which varied with prepayment rates — these typically increased when interest rates were low or decreased when those rates were high. Both of these challenges — in liquidity and interest rates — posed significant risks to specialized mortgage lenders that concentrated on American-style mortgages. 

[12] It’s not controversial to conclude that these three major flaws can lead to financial instability at F&F. The history of the GFC, including its many subsequent examinations, and of the S&L Crisis of 1989 (caused primarily by interest rate and liquidity mismatches) support this view. Additionally, past episodes of bank stresses, like the high concentration of bank failures in Texas in the 1980s due to their credit risk over-concentrated in Texas commercial mortgages, further illustrate these potential sources of instability. A possible fourth source of risk for the GSEs could stem from their multifamily mortgage operations, but these assets are just not large enough to overwhelm the GSEs’ capital. For example, at Freddie Mac, multifamily loans account for only 13 percent of the total loans financed, compared to 87 percent for single-family mortgages.

[13] Ginnie Mae was created in 1968 to act as the mortgage securitization arm for the FHA and the mortgage lending units of the U.S. Department of Veterans Affairs and the U.S. Department of Agriculture.

[14] Investors took on this risk, as the phrase “pass-through” indicates, by being paid back not on a preset schedule but as principal and interest payments were received from the underlying borrowers (except in the case of defaulted borrowers, where F&F make investors whole). Thus, large increases or decreases in interest rates, producing prepayments occurring earlier or later than usual, did not significantly impact the profit of the issuer of the pass-through MBS, including F&F. Instead, the MBS investors absorbed this risk. 

[15] There are many fixed-income investors who do not want credit risk and focus solely on interest rate risk.  Traditionally, these investors purchased Treasury securities of a particular maturity. To them, the MBS issued by the GSEs were an alternative, albeit complex, to Treasuries, typically providing a somewhat higher interest rate. This investor base today supports about 70 percent of the $13 trillion U.S. mortgage market, which is quite substantial. By comparison, the market for investors willing to take on this type of interest rate risk intermixed with credit risk is believed to be much smaller. 

[16] The implied guarantee was the assumption by bond investors that the U.S. government would prevent them from losing money on debt instruments issued by F&F. This belief was based on several factors, including a $2.25 billion line of credit from each of F&F to Treasury that had been established by their charters. This assumption, of course, proved to be correct as those investors suffered no losses in 2008.

[17] If this risk seems obscure or minor, it is definitely not. That was proved by the S&L Crisis of 1989. The main asset of S&Ls, i.e., the typical American mortgage, was funded by a pool of deposits (1) with maturities measured in months, not years, or (2) with an open-ended maturity that could be withdrawn on short notice. This was successful as a strategy as long as interest rates were low and stable. However, as inflation during the 1970s and early 1980s skyrocketed, interest rates followed — mortgage rates peaked at 18.6 percent in October 1981. The result was that the thrifts suffered a giant loss of deposits as their customers sought higher yields, and the depositors who stayed had to be paid an interest rate higher than what the S&L’s fixed-rate mortgage loans earned on average. This led to losses that, despite various government strategies to address the distress — which sometimes even made the situation worse — resulted in the collapse of the industry in 1989. About one-third of the over 3,000 S&Ls failed. To protect depositors, the government spent over $130 billion at the time, a sum equivalent to more than $300 billion today. 

[18] F&F were motivated to invest on a large scale because they were able to access funds by issuing unsecured debt issued at below-market rates, thanks in part to the implied government guarantee. In fact, the profits from their investments were fundamentally wholly due to the below-market cost of this funding rather than any particular investing prowess on the part of F&F.

[19] The peak total size of those investment portfolios, i.e., near $1.6 trillion, was at that time much larger than the balance sheet of the Federal Reserve System. It can also be argued that these investment portfolios were as much a cause of the losses at F&F in 2008 as the underlying issues in the mortgage securitization and guarantee business. 

[20] Early in the 2000s, the Federal Reserve had noticed these substantial, short-funded portfolios and considered them too much of a potential source of financial instability. It therefore began advocating a limit on the size of GSE investment portfolios, which were uncapped. Together with the presidential administration of George W. Bush, the agency proposed legislation to implement such a cap but was defeated in Congress in 2005 by the lobbying strength of F&F. Three years later, the Bush administration saw an opportunity to impose a limit via conservatorship and took it.

[21] Interestingly, today there are some well-publicized proposals from primary market lenders to relax the tight restrictions on the investment portfolios of F&F, looking for them to be ordered to purchase large amounts of their own MBS (one proposal specifies $600 billion, i.e., $300 billion each) to help push down MBS interest rates to make mortgages lower cost. This would be a major reversal of government policy about the riskiness of the GSEs, putting large interest rate and liquidity risks right back on them after many years of unwinding just such activities.  

[22] Such low capital requirements were true for the GSEs, for the Federal Home Loan Banks, and for federally chartered savings and loans, as examples.

[23] The low requirement of 0.45 percent was likely due to confusion surrounding the accounting practices then used by the GSEs. Until the mid-2000s, F&F treated mortgages they sold into pass-through MBS and then guaranteed as off-balance sheet liabilities, i.e., only listed in the footnotes of their financial statements. Regardless of the accounting, F&F had the full credit risk of those mortgages — something apparently not fully appreciated in 1992. The SEC ruled in late 2004 that the off-balance sheet accounting was wrong, and since then the MBS liability has been fully included on their balance sheets. 

[24] The marketplace did not look to the capitalization of F&F to judge their creditworthiness during conservatorship but to Treasury’s support of them via an explicit mechanism contained in a legal agreement. It was not as good as a full faith and credit guarantee by the government but was much stronger than the implied guarantee that had been assumed before conservatorship. It successfully and quickly restored market confidence in F&F and has been in place ever since. The agreement’s formal name is the Preferred Stock Purchase Agreement (PSPA).

[25] This was developed by Freddie Mac at my specific direction early in my tenure as CEO. 

[26] During the regulation proposal process, a counter-cyclical buffer was added based on feedback from the public. The $180 billion requirement included that buffer. 

[27] It has been criticized as being both too high and too complex. As of September 30, 2025, the ERCF required $343 billion in capital. By comparison, if the FHFA went back to the 2018 CCF-based proposal, today it would require roughly $250 billion in capital, as F&F’s balance sheets have grown significantly since 2018. 

[28] In my view, even the smaller 2018 proposal — which would require about $250 billion today — would also be fully safe and sound.

[29] It is important to note that the official 2023 and 2024 stress test results for F&F, as designed by the FHFA but using Federal Reserve-created assumptions, showed that, even with a modeled downturn in home prices exceeding 30 percent (that’s higher than the drop experienced during the GFC), they would essentially have not lost money over the nine-quarter test horizon. This discrepancy between the stress test results and the regulatory capital requirements is striking, yet the FHFA has never really explained the seeming inconsistency of the two approaches.