House under construction

Housing Perspectives

Research, trends, and perspective from the Harvard Joint Center for Housing Studies

The New Government Mortgage Lending Limit Approaches $1 Million: Does This Call for Reassessment and Revision?

On November 30, 2021, the Federal Housing Finance Agency (FHFA), the regulator and conservator of Freddie Mac and Fannie Mae, the two government-sponsored enterprises (GSEs), announced new limits on the dollar amount of a single-family mortgage the GSEs can purchase in 2022, known as the conforming loan limit. Prior to 2008, this was a single limit that applied nationwide; legislation that year, though, added in a complex set of varying limits for “high cost” areas of the country, but with an upper bound of 150 percent of the general baseline amount. For 2022, the baseline limit was set at $647,200, a substantial 18 percent increase over 2021, reflecting (as specified in the 2008 legislation) the similarly large nationwide average increase in house sale prices. The upper limit for the high-cost areas was thereby set at 150 percent of this number, or $970,800.

The FHFA news release announcing the new limits and its associated FAQ provide a good, detailed description of the limits.

The baseline limit and the high-cost area ceiling are the headline numbers, attracting public attention each year when they are announced. Given how sizeable the 18 percent increase was and how the high-cost area ceiling came in just below the attention-grabbing $1 million level, the interest generated by the most recent announcement was understandably greater than usual.

This raises a fundamental policy question: should the GSEs, which finance mortgages on favorable terms due to significant subsidies from the taxpayer, be doing so to benefit families who are wealthy enough to carry a million-dollar mortgage? Is that an appropriate use of taxpayer support? Also, as other housing subsidies like the mortgage interest deduction generally lead to higher investment in – and consequently a higher price on – larger homes, the new conforming loan limits arguably further exacerbate this distortion, possibly even leading to reduced investment in lower-priced, more affordable homes, for which there is a well-known shortage. In fact, when the FHFA announced the changes, Acting Director Sandra Thompson issued a statement that the agency was already examining the limits for the latter reason. (In a similar vein, on January 5, the FHFA announced that interest rates on most GSE high-balance loans will be increased, which will somewhat reduce the subsidies going to wealthier borrowers.)

Indeed, I believe it would be a good thing for the newly-announced limits to trigger a broad reassessment—and possible revision—of the loan limit system at the GSEs and also at the two other government mortgage agencies, the Federal Housing Administration (FHA) and the Department of Veterans Affairs (VA). To help inform this debate, there are five important questions:

1. While $1 million is an eye-catching number, the baseline and high-cost area ceiling calculations simply reflect extremely high house price growth in recent years, which seems to be exactly how the loan limit system should work. Wouldn’t any reduction of the limits thus be a takeaway from the American homeowner?

When considering solely the conforming loan limits, this is a valid viewpoint. However, those limits do not exist in isolation, and the totality of factors driving how much the government supports and subsidizes home mortgages has evolved to push up the percentage of new mortgages that are financed by the four government agencies to high levels in recent years. Specifically, the four-agency cumulative market share averaged almost 55 percent from 2001 to 2003, which is seen as a relatively normal period not distorted by the bubble that was just beginning. More recently, the share has averaged about 70 percent from 2014 to 2019, when the dislocations of the bubble bursting were mostly in the past, and so this higher percentage seemed to represent a “new normal.” This large market share increase stemmed primarily from the collapse of the private label (i.e. not government supported) securitization market, which became particularly uncompetitive after its weaknesses were revealed in 2007-08, and secondarily by less dedication of bank balance sheets to first mortgages. Interestingly, this market share loss was replaced almost wholly by the FHA and VA, and not the GSEs, in part because the loan limits of the former increased relative to the latter after the bubble burst. (In the pandemic, the market share of the government agencies has increased even more, to about 75 percent, with the GSEs gaining share at the expense of banks and even FHA and VA; this has not yet proven to be more than temporary, however.) That 70 percent range is an extraordinarily high level and I do not believe was ever contemplated when the 2008 legislation revised the conforming loan limit system. Therefore, it can reasonably be argued that a reduction of the limits is an appropriate and easily-implemented policy lever to force the government agency market share partially or wholly back down to the 50 percent range, leading to less taxpayer subsidies for higher-priced homes. As this would restore the government agency share closer to its historic pre-bubble norm, it is not really a fundamental takeaway from the American homeowner.

2. While a reassessment is an interesting idea, since the loan limits used by the GSEs (as well as the FHA and VA) are embodied in legislation, can it lead to an actual revision without Congress passing new legislation, which is an unlikely outcome at this time?

This is a widely held view. However, during 2013, when the FHFA was led by Acting Director Edward DeMarco, he obtained an opinion from the FHFA legal staff that the agency, acting as the conservator of the two GSES, could set the loan limits lower than the formula called for in the legislation, but never higher. While this was never acted upon, it makes clear that a reassessment could produce a revision without congressional action. However, given that the FHFA lost its regulatory independence last June via a Supreme Court ruling, this now means that the Biden administration is calling the shots at the GSEs, as it already does at FHA and VA. Therefore, the administration would need to be supportive of such a reduction and it’s not clear that it would be. When it comes to reducing FHA or VA limits, unfortunately, legislation is likely required.

3. Given the politicization of America’s housing finance system, would any proposed revision of the GSE loan limits survive lobbying by influential economic and ideological interest groups, or would it be a waste of time and effort?

I have heard senior FHFA officials over the years express their frustration with the process of getting public feedback on any proposal, as it is typically dominated by the “usual suspects” expressing long-held, predictable views. I would mostly expect that to be true in this case: (i) conservative think tanks would strongly support a reduction in the GSE limits in line with their longstanding view that any reduction in the GSE footprint is a good idea; (ii) the mortgage banking industry (both banks and non-banks) would fight any reduction in the limits as it would lead to less profit for their members, although there may be dissenting views from some bank lenders, and (iii) the private label securitization (PLS) industry would be in favor of a reduction as it would lead to more volume passing through its hands and thus more profits. But, the viewpoint of liberal housing advocates, who are particularly influential with the Biden administration, is harder to predict. On one hand, they should be supportive of lower limits so the GSEs would subsidize the upper-middle class less (i.e. reducing “welfare for the well-off”), which would in turn allow the companies to focus their energy and subsidies more on low- and moderate-income (LMI) borrowers. On the other hand, the GSEs cross-subsidize the pricing of their mortgage purchases as higher-balance loans produce part of a subsidy pool that then is used to reduce the interest rate on mortgages to LMI borrowers; as such, liberal housing advocates might not support a limit reduction because it would shrink that pool of available subsidies. So, how those advocates (who speak with many voices, not just one) would face this dilemma is unpredictable; perhaps one might come up with a creative way to reduce the loan limits while protecting the bulk of the cross-subsidies. This is where the value of a public reassessment could be, and which might also be determinative of the outcome.

4. What about the FHA? Should its loan limits be reassessed as well?

FHA loan limits only partially mimic those of the GSEs. Basically, the FHA sets a loan limit for each specific geographical area (mostly counties) at 115 percent of the local median sale price of homes. There is a floor on how low the limit can be, which is equal to 65 percent of the GSE baseline loan limit (now $420,680) and a ceiling set at the GSE high-cost area limit (now $970,800). So, the FHA at its core is clearly focused on its target market of LMI borrowers, which is appropriate, and does not appear to need a reassessment. However, the maximum (for the limited number of geographies that qualify) is still almost $1 million, which seems especially questionable given that the raison d’être of the FHA is to focus on first-time homebuyers and families with more marginal creditworthiness. This calls for a reassessment as to whether such high-balance borrowers deserve the large taxpayer subsidy (which is even larger than that enjoyed by the GSEs) contained within all FHA mortgage financing activities.

I also note that the GSEs and FHA have long-standing special statutory provisions relating to Alaska, Hawaii, Guam, and the Virgin Islands, including an FHA limit of almost $1.5 million. These certainly should be included in any reassessment.

5. Similarly, should there be a change in how the VA limits the dollar size of the mortgages it insures?

The VA operates much like the FHA in how it finances mortgages, except that some features are more generous to veteran-borrowers (e.g. allowing zero downpayment, versus the FHA’s 3.5 percent minimum) as a way for Congress to deliver fringe benefits for their military service. That certainly seems reasonable but, interestingly, at the beginning of 2020, by law all limits on the size of mortgages insured by the VA were removed. This is curious and difficult to understand. For the small number of veterans who become wealthy enough to afford such large mortgages (potentially millions of dollars in size), is it really good public policy for the taxpayer to subsidize them without limit, even while respecting their military service? I know a few veterans who are wealthy and I cannot see any of them supporting such a subsidy for themselves. And it would undoubtedly be a scandal if it became publicly known that a very wealthy veteran received, for instance, a taxpayer-subsidized $5 million, zero-downpayment mortgage through the VA. Thus, a reassessment should include putting a dollar limit on the size of VA mortgages, and it could even be a generous one to reflect that it is designed to be a fringe benefit for veterans.

My conclusion is that the high-cost area limit approaching $1 million (and undoubtedly handily breaching it next year) should be taken as an opportunity for a reassessment and revision of the loan limits by all four government mortgage agencies. However, dramatic or unusual changes are unlikely to be implemented given the political influence of the interest groups involved. Instead, I see there being a narrow window for a few modest but well-chosen revisions within the existing framework: (i) for the GSEs, either reduce the limits modestly (e.g. 20 percent) or institute a cap (at a number under $1 million), while protecting the majority of the cross-subsidies that go to LMI borrowers; (ii) for the FHA, have a similar reduction or cap; (iii) for the GSEs and FHA, reduce some of the special treatment given to Alaska, Hawaii, Guam, and the Virgin Islands; and (iv) for the VA, cap the now-unlimited amount allowed, but at a generous level to reflect that its borrowers are veterans.