Erdmann Housing Tracker · Housing & Cities
TIER 5 Tue, 11 Nov 2025 23:43:40 +0000
FHFA chair Bill Pulte has recently been lighting up news feeds with announcements of changes at Fannie Mae and Freddie Mac. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ | | ---|---|--- | | | Forwarded this email? Subscribe here for more --- --- # Credit Access at Fannie & Freddie | | Kevin Erdmann --- | Nov 11 --- | --- --- | | | --- | | --- | | --- | | --- | | READ IN APP --- FHFA chair Bill Pulte has recently been lighting up news feeds with announcements of changes at Fannie Mae and Freddie Mac. The agencies are being given more freedom to consider different credit score sources, and the 620 minimum credit score is being removed. Upgrade to paid I am happy to see discussion of credit standards, and to see a tendency toward loosening standards. That is definitely one shift that would improve housing markets. However, the tightening in lending standards reaches all the way up to scores in the mid-700s, so my inclination is to assume that an administration that sees the 620 minimum as the important binding constraint doesn't really have a good idea of what the problem is. Nobody would be happier to see more mortgages being originated to borrowers with scores under 760, but I will reserve that happiness for when change actually happens. In the meantime, Pulte faces a public that is still in the embrace of the moral panic that caused the 2008 financial crisis, and so this post is my reaction to the reaction. The public reaction to Pulte's announcements seem to be overwhelmingly negative. (See replies in this Twitter thread.) Among the reactions, I saw a citation to a 2021 paper, titled, "Moral Hazard during the Housing Boom: Evidence from Private Mortgage Insurance". Here's the abstract of the paper: > Was the mortgage boom fueled by optimism around house prices, or did misaligned incentives in the mortgage industry also play a role? In this paper, we provide novel evidence of a role for misaligned incentives. We document that private mortgage insurance (PMI) companies dramatically expanded insurance issuance on high-risk mortgages purchased by Fannie Mae and Freddie Mac at the tail-end of the housing boom, without changing pricing and despite knowledge of heightened housing risk. The PMI expansion facilitated an unprecedented increase in Fannie Mae's and Freddie Mac's purchases of high-risk mortgages, extending the mortgage boom into 2007 and precipitating their collapse as well as that of the PMI industry. We argue that this unraveling reflects a general moral hazard problem associated with insurance providers coupled with misaligned incentives in the government-backed portion of the mortgage market. That abstract is an example of the pre-determined conclusions about 2008. "Experts disagreed about whether the mortgage boom was fueled by saucer people or reverse vampires." The literature is filled with these "process of elimination" conclusions that leave out the most important alternative explanations. The mortgage boom was fueled by mortgage insurance trends at Fannie and Freddie _in 2007_?! The first sentence of the paper commits the fallacy I mentioned in another recent post. "The collapse of the housing market in 2008, which resulted in over six million foreclosures, as well as the failure of the Government-Sponsored Enterprises (GSEs) Fannie Mae and Freddie Mac, initiated a heated debate over the causes of the boom and bust." The "boom and bust" is a single event. The arrival of a bust proved its own inevitability. That _some form of_ unsustainable excess was responsible for the bust was axiomatic. All that was left was to argue which form of excess it was. More fatalism follows. "In October 2006, forecasters at Moody's wrote, 'The U.S. housing market downturn is in full swing,' predicting the first national decline in nominal house prices since the Great Depression and a 'crash' in house prices in nearly 20 metro areas in 2007." The fatalism is subtle. There is a lot happening in the background here. Most importantly, as of late 2006, is that the Fed reaction to this changing sentiment was to encourage rather than stop it! Universally, all forecasts of doom in that period are treated as prescient and all forecasts of stability are treated as foolish. The larger problem is that they were treated that way when they were issued, before the outcome played out. #### Mortgage Insurance Risk in 2007 I need to be careful here. The issues here are subtle. The authors present evidence that there was moral hazard among mortgage insurers working with Fannie and Freddie. One good piece of evidence for that: > We show that PMI growth in 2007 was sharply lower and close to zero above the conforming loan limit (CLL) where the GSEs do not operate. This finding is consistent with private investors doubting the ability of PMI to pay out in a downturn, whereas the GSEs, which had little incentive to consider counterparty risk, remained willing to do business with the PMIs. We further show that PMI companies were willing to take risk in the non-GSE market above the CLL, as they reduced underwriting standards similarly in both markets during the 2000s, and their policies experienced similarly high delinquency rates. In essence, the subprime lending boom could have expired when private market participants stopped issuing piggyback loans and supporting private-label securitizations, but the collaboration of the PMI firms and the GSEs allowed exceedingly risky lending to continue through 2007. They note, "The timing is crucial: Mortgage insurers accelerated risk-taking exactly when price forecasts were negative and other private participants retrenched in response to a number of red flags in the housing market. For instance, in October 2006, Moody's described the national market outlook as 'increasingly dark pessimism'." There is no doubt that their observations of moral hazard are accurate. There is no doubt that the large number of loans made in 2007 at Fannie and Freddie created losses for them. But, note that all of this is served on a platter of fatalism. Without this moral hazard, the crash could have happened sooner. The timing _is_ important. These insurance products were issued well into the housing downturn. They objectively had nothing, and could have had nothing, to do with rising home prices or construction activity. At best, they slowed the collapse. And, reading between the lines, slowing the collapse is axiomatically treated as the problem. "Overall, …PMI fi rms disproportionately expanded issuance to consumers, products, and markets that were ex-ante riskier based on widely observable measures." Yes. There was moral hazard. The risks were underpriced. And, when the universal demand for collapse was finally implemented, they took losses. Should there be reforms or regulations to prevent moral hazard? Yes. Did moral hazard contribute to losses? Yes. Were those losses inevitable? No. Was the moral hazard causally related to a housing bubble or its reversal? Absolutely not. The primary reason for the increase in mortgage insurance at the agencies was that the subprime market had already collapsed by 2007. There had been a large number of "piggyback" loans where borrowers take out a second mortgage. Those were mostly originated in the private securitization market. When that market dried up, some borrowers turned to the agencies, where they would refinance into a single mortgage with a loan to value ratio above 80%, which required mortgage insurance. One _fundamental purpose_ of Fannie and Freddie is to smooth the cycle - to step aside when markets are hot and step in when markets are in disarray. They had stepped aside when markets were hot. They were stepping in when markets were in disarray. The unspoken but overwhelming subtext of this paper and so many others is that 2007 was different. We were owed a collapse, and any activities which stabilized the market were inappropriate. Figure 1 is from the Urban Institute. The amount of market share the agencies were willing to give up during the boom was Herculean. There is so much motivated reasoning blaming them for the crisis. We created countercyclical institutions, and they gave us this. My goodness, that is commendable. We don't deserve them. | | ---|---|--- Figure 1 The moral hazard issue was real, but it was _beside the point_. And yet, these sorts of papers dominate the literature. Nobody would dare write a paper about those dismal expectations in 2006 and wonder if policy might have aimed for something brighter. #### A Note on Timing I agree with these authors. Timing is important. Figure 2 is from the paper. This shows delinquency rates of privately securitized mortgages (blue dashed) and agency mortgages with insurance (red solid). Clearly, delinquency rates increased on the cohorts originated in 2007. | | ---|---|--- Figure 2 But, this chart subtly sneaks in the fatalism. These rates are the delinquencies experienced within 36 months of origination. The delinquencies could have happened as late as 2010. Charting them this way subtly treats things that happened in 2010 as if they were set in place in 2007. Figure 3 is from a 2009 report from Fannie Mae. You can see that the poor results were largely a result of future market outcomes. Mortgages originated in 2004 performed better than loans from 2000 and 2001 until after the 2008 collapse. There is a similar pattern with 2005. Defaults turn up in year 3 and after - after the market collapse, and after mortgage access had been deeply curtailed. Defaults turn up in 2006 and 2007 in year 2, again a bit faster in the 2007 cohort than in the 2006 cohort. | | ---|---|--- Figure 3 What would default rates have been if the market hadn't crashed? If the crash was fated, that's not a question we can ask. And that choice - to ask or not ask that question - is much more important than the facts on the ground about risk taking in 2007. What if we had _stabilized the market_ and reformed moral hazard issues at the insurers? The top 2 panels of Figure 4 are from the Urban Institute. Estimated default risk for the whole market and for the agencies. I have aligned those charts with high and low tier home prices in Las Vegas. Total default risk had risen during the private subprime boom. By 2007, that market had crashed. Fannie and Freddie increased their default exposure while that happened, but it didn't come close to creating stable conditions. Very quickly, by the middle of 2007, when credit risk at the agencies peaked, credit risk across the entire market had fallen to new lows. By 2008, the agencies were tightening too. As these authors note, when Fannie and Freddie and their insurers were taking on more risk, the riskiest markets were in decline. Loans with less equity were largely due to declining prices. And, surely those markets had room to correct. But _**why shouldn 't we have aimed for that correction to be slower and calmer**_? These authors explicitly call out Fannie and Freddie in 2007 for slowing the decline, as if steeper decline was preferable. | | ---|---|--- Figure 4 Furthermore, over the period of time that these loans were defaulting, a naive person looking at these charts from outside the moral panic would clearly and undoubtedly identify the period after 2007 as the period with the extreme shift in lending standards. That shift is universally ignored in this literature as a potential aggravator on economic decline, housing market trends, and default rates. And, don't forget. Even while Fannie and Freddie were partially helping to maintain some stability, the Nevada market was in deep decline, and Nevada had not experienced a building boom, by any stretch of the imagination. The rate of construction in Nevada was no higher in the 2000s than it had been for decades prior. In the bubble markets, population growth suddenly collapsed after 2006, but construction collapsed first and faster. The problem in Nevada in 2007 wasn't too much capital flowing to housing. It was an economic collapse so sharp it could be seen in population trends. | | ---|---|--- Figure 5 Finally, Figure 6 is similar to Figure 4, but with Atlanta in the bottom panel. Think of Atlanta as the control group. As the authors note, most of the underpriced risk was in the bubble markets like Las Vegas. Think of it this way. If Fannie and Freddie and the Fed were aiming for a _proper correction_ in Las Vegas, where one was likely necessary, they should also have been aiming for stability in cities like Atlanta. | | ---|---|--- Figure 6 This paper and practically every other paper in the literature discusses the risks of 2007 with absolutely no accounting of the change in lending and housing markets over the first few years of those loans. The _**obvious**_ counterfactual to consider is how poorly those 2007 mortgages and their insurers would have performed if the housing market had been managed in a way that would have kept Atlanta stable - or, say, _at least_ within shouting distance of previous cyclical low points. The existing literature is not even aware that this is a question to ask. Here is one last figure, showing the homeownership rate in Georgia and Nevada. The activity in 2007 at the agencies was not associated at all with more homeowners. That was largely refinancing activity. And, in fact, the entire subprime boom was not associated with more homeowners. It was largely associated with increased investor activity and refinancing. | | ---|---|--- Figure 7 The deep cut in mortgage access after 2007 was associated with a deep drop in homeownership that coincided with the collapse of housing markets across the country. (Also, keep in mind that age demographics should be associated with a relatively steep upward trend in homeownership over this period.) Changing access for potential homeowners, like the changes being contemplated at the agencies today, has nothing in common with anything related to either the boom or the bust and crisis in 2008. Additionally, the removal of access to first-lien purchase loans for marginal buyers that need reasonable and generous credit access _was a big part of the problem in 2008_. It's reversal is already 20 years late. And, the recent rise in the homeownership rate back toward previous norms is largely due to housing supply that is at such crisis levels, it is preventing renter households from forming. The homeownership rate is rising because of stagnant household formation in the denominator, not rising homeownership in the numerator. One last point that I do agree with the authors about is: "If the purpose of the private mortgage insurance industry is to protect taxpayers from risk, then this approach to risk-sharing has largely been a failure." I agree. Systematic risk is best borne by taxpayers through the agencies. If we collectively decide to shoot ourselves in the foot, we should collectively bear the wound. #### Conclusion Recent announcements about increasing mortgage access will be met with howls of smug complaints. "Here we go again." A moral panic _against_ moral hazard was _far_ more important in leading to crisis than moral hazard was. Those howls are coming from the wolves you should fear. Upgrade to paid You're currently a free subscriber to Erdmann Housing Tracker. For the full experience, upgrade your subscription. Upgrade to paid --- | | | Like --- | | Comment --- | | Restack --- (C) 2025 Kevin Erdmann 548 Market Street PMB 72296, San Francisco, CA 94104 Unsubscribe