Highlights: Ep. 106. Mortgage Lending Standards with Kevin Erdmann

Housing Was Undersupplied During the Great Housing Bubble

• “Just a few cities are at the heart of the housing supply problem, most notably New York City, Los Angeles, Boston, and San Francisco, which I refer to as Closed Access cities. There are two very different housing markets within the United States: the Closed Access market, where new housing is highly constrained, rents rise relentlessly, and households are forced to make difficult choices as housing expenses eat up their budgets; and the rest of the country, where homes can generally be built to meet demand, housing construction is healthy, and housing expenses remain at comfortable levels for the typical household.”
• “If we add these two markets up into an aggregate market, it looks like a market where rents are relatively level over time. In the 2000s, when housing starts were rising and home prices were also rising to unusually high levels, it appeared as if those rising prices were unrelated to rent, and it appeared that prices were rising at the same time that supply was rising. This pattern, rising prices and quantities, seemed to be the result of excess demand—too much credit and too much money funding too much housing.”
• “Yet few places fit that description. For the most part, there were places where housing starts were low, while rents and prices were both rising, and there were places where housing starts were healthy, while rent and price increases were moderate. If we compare median annual rent and median home price within each metropolitan area, it is clear that rents were an increasingly important determinant over the past two decades of home price differentials between different metropolitan areas. And as shown in figure 5, in this regard, the Closed Access cities have become outliers—much higher rents leading to much higher prices.”
• “The Closed Access cities have become new centers of prosperity, but they have limited the growth in their populations through restrictive zoning and bureaucratic obstacles that make it difficult to build housing. This has turned them into enclaves of privilege, only open to the richest newcomers, who spend nearly half their incomes on rent. This pattern has only developed since the 1990s and is neither normal nor natural.”
• “From 1996 to 2005, across the United States permits were issued to build 6.5 homes per 100 residents. The Los Angeles, Boston, and New York metro areas each approved fewer than 2.6 per 100 during that time. San Francisco approved 3.4. In contrast, other economically prosperous cities that attract aspirational families in search of economic opportunity, such as Washington, DC, Seattle, and Dallas, issued permits at rates higher than the national average.”
• “Contrary to Chairman Bernanke’s assumption, at the national level there was no overhang of housing supply that needed to be worked off in 2011. Indeed, even in 2005 there was no national oversupply of housing. Rather, the American economy was burdened by a shortage of housing, especially in the Closed Access cities.”
• “The housing bubble was concentrated in cities in the coastal Northeast, California, Nevada, Arizona, and Florida. Limiting our analysis to the 20 largest metropolitan areas, the Closed Access cities make up three-quarters of the “bubble” cities, in terms of total real estate valuation. Constrained housing supply was clearly the primary source of high prices in those cities, not excess demand. Prices in the Closed Access cities today remain as high relative to other cities as they were during the bubble because constrained supply is the fundamental reason for those high prices, not reckless credit markets.”
• “Even in other bubble cities with generous building policies, the primary cause of rising prices was the severe Closed Access shortage of housing. This is because those other bubble cities were the main destinations for households migrating out of the Closed Access cities. I call those cities Contagion cities, because in spite of their more generous building policies, they were overwhelmed by the problem created by the Closed Access cities. In the years leading up to the financial crisis, the shortage of housing in the Closed Access cities had become so severe that each year hundreds of thousands of households moved away in search of an affordable home. Many of them landed in inland California, Nevada, Arizona, and Florida.”
• “Figure 7 compares net domestic migration of Closed Access cities and Contagion cities. Notice that high rates of out-migration from Closed Access cities correspond to periods of large in-migration to the Contagion cities. Credit markets may have facilitated some of the housing activity during the housing bubble, but at its core this was a mass migration event caused by a lack of housing.”

• “For many people, it seemed obvious that there was overbuilding in places like Phoenix. From 2003 to 2005, Phoenix built many homes. Meanwhile, prices of Phoenix homes rose by about 75 percent in just two years. By 2007, however, the Phoenix housing market was collapsing, buried in a mountain of unclaimed inventory. Surely, it was argued, this was a classic credit-fueled boom and bust.”
• “But, for the boom-and-bust story to add up, Phoenix would have had to build enough homes for all of those new households moving in from California, and then it would also have had to build tens of thousands of units in addition to that. It couldn’t. The problem Phoenix encountered was that the in-migration was so strong that even Phoenix authorities couldn’t approve new supply fast enough to meet demand. Building permits in Phoenix jumped by about 50 percent from 2001 to 2004. By all appearances, that is an extremely frothy market, but as figure 8 shows, the jump in new homes tracked virtually 1:1 with net in-migration.”
• “Many of those in-migrants were coming from California. They were moving to Phoenix largely to reduce their housing expenses. In fact, even though migration from California had continued to rise up through 2005, net migration into Phoenix had leveled off. That is because increasing numbers of households now began moving away from Phoenix, which had seen soaring home prices. From 2005 to 2008, migration into Phoenix declined each year while migration out of Phoenix continued to rise. By 2008, net in-migration into Phoenix was less than 10,000 households. By 2006, Phoenix had a growing number of empty homes and a large inventory of homes for sale. But from 2005 to 2008, the number of new homes approved in Phoenix dropped faster than net migration was dropping. Housing supply had reacted remarkably quickly to shifting demand. Even as housing starts were collapsing, rents were rising, as they were in most cities at the time.”
• “The question that needs to be addressed about the housing bubble and the ensuing bust is not what caused prices to rise so sharply. That is a fairly straightforward question, with a standard economic answer. Fundamentally, there weren’t enough houses. What caused the massive out-migration from the Closed Access cities? The answer to that question is also, fundamentally, that there weren’t enough houses.”
• “This leaves one additional question that has been rarely asked, and which must be answered if we are to come to terms with the crisis that followed. If a lack of housing was fundamentally the cause of the housing bubble, then why had housing starts been collapsing for more than a year before the series of events occurred that we associate with the crisis, like nationally collapsing home prices, defaults, financial panics, and recession? And what caused the Closed Access migration event to suddenly stop at the same time as the collapse of housing starts?”
• “For a decade, the collapse has been treated as if it was inevitable, and the important question seemed to be, What caused the bubble that led to the collapse? This needs to be flipped around. Given the urban housing shortage, it was rising prices that were inevitable. So the important question is, Why did prices and housing starts collapse even though the supply shortage remains? And why were housing starts still at depression levels in 2011?”
• “The surprising answer to those questions may be that a housing bubble didn’t lead to an inevitable recession. It may be that a moral panic developed about building and lending. The policies the public demanded as a result of that moral panic led to a recession that was largely self-inflicted and unnecessary. They also led to an unnecessary housing depression that continues to this day.”

Getting Corporate Money Out of Single-Family Homes Won’t Help the Housing Affordability Crisis

• “Figure 1 shows the quarterly amount of mortgage originations (new mortgages) as a percentage of the total value of owner-occupied housing. (When interest rates decline, many households refinance their mortgages. That explains the bumps in 2003, 2012, and 2019. Note that the 2003 bump predated the subprime private securitization boom, which dominated new mortgage activity from 2004 to early 2007.) Except during the refinance booms, borrowers with credit scores higher than 760 regularly borrowed capital equal to about 1 percent of owner-occupied US housing stock each quarter, both before and after the Great Recession.”

• “Before the Great Recession, borrowers with credit scores below 760 regularly borrowed capital equal to between 2 and 3 percent of the value of owner-occupied US housing stock. That percentage range, as well as the proportion of mortgages originated to borrowers with lower credit scores, didn’t rise appreciably during the subprime boom period from 2004 to 2007. Then, between 2007 and 2009, the proportion of new mortgages going to borrowers with scores below 760 dropped by more than half and remains that low today. (The average credit score among all borrowers tends to be a bit above 700 points.)”
• “This change wasn’t a reversal of the boom era subprime lending excesses. The same pattern showed up at Fannie Mae (Federal National Mortgage Association). As figure 2 shows, from 2000 through 2007, the average credit score on Fannie Mae mortgages barely changed. Figure 2 also compares Fannie Mae’s estimate of the value of homes being funded by new mortgages and the value of homes in Fannie Mae’s book of business (homes with mortgages that had been originated in earlier years). Home prices had risen, but the homes associated with old mortgages had similar values to the homes associated with new mortgages. In other words, even though home prices were rising, the mortgages were going to the same types of borrowers in the same types of homes in 2007 that they had been going to for years.”

• “By 2009, the value of homes across the country, including those with Fannie Mae mortgages, had fallen significantly. But in 2009, the average credit score on new Fannie Mae mortgages was about 40 points higher than it had been previously. And, strikingly, the average value of those new homes skyrocketed to well over $300,000—about 60 percent higher than the homes with mortgages from earlier years.”
• “The profile of borrowers served by conventional lenders changed greatly after 2008. Those with credit scores under 760 were much less likely to qualify for mortgages, so the homes those borrowers were likely to live in had fewer potential buyers.”

Mortgages Outstanding by Credit Score

• “At Fannie Mae, before 2008, roughly 2/3 of mortgage lending went to sub-740 credit scores. Then, it suddenly switched, and since 2008, roughly 1/3 of mortgage lending went to sub-740 credit scores. The proportion of lending going to lower credit scores had been pretty stable as far back as I can track the data. The sub-prime/Alt-A lending boom wasn’t associated with a significant change in lending by credit score.”
• “You can see that in Figure 1, from the New York Fed’s Household Debt and Credit report. There was a refi boom in 2003, which tends to increase mortgage activity among the highest credit scores. Then, from 2004 to 2007, during the subprime/Alt A boom, proportions returned back to the long-term norm. The New York Fed tracks 5 bins of credit scores. I have combined the bottom 2. Total activity before 2008 tended to be roughly the same in the 4 bins, and other data finds those proportions at least into the 1990s. Compare that to today, where all mortgage originations to scores under 740 don’t even add up to the quantity of originations to scores over 760.”
• Figure 1.

• “Since 2008, we’ve had a haves and have-nots market. A third of borrowers are locked out of the market. And borrowers who can still get mortgages had a buyer’s market with, until recently, both moderate prices and low interest rates. It was, by far, the best time in recent history to be a mortgaged buyer, for those who were allowed to be.”
• “The have-nots who lost their homes or are unable to buy homes have gotten the worst of the crackdown, because their rents have generally increased massively. The have-nots who were grandfathered in and managed to hold on to their homes ended up with a bit of a windfall because their homes, especially, have elevated prices because of the effect the mortgage crackdown has had on rents.”

The Moral Panic, Its Perpetrators, and Its Victims

• “I have done something like this before – tracking a poor and a rich ZIP code through the timeline of our offenses. It’s time for an update. I hope that this can provide a concise and comprehensive picture of the confusion and devastation of the mortgage moral panic. Here, I will be comparing 2 ZIP codes in Atlanta. ZIP code 30022, where the average income is currently about $190,000, and ZIP code 30344, where the average income is currently about $50,000.”
• “Figure 1 shows the typical price/income ratio in the rich (red) and poor (black) ZIP codes, over time. The price/income ratio in the rich ZIP code approximates a straight line. The price/income ratio in the poor neighborhood approximates a straight line until 2008. The 3x – 4x levels of price/income in the period before 2008 are pretty common for amply housed cities with normal markets, for those of us old enough to remember one.”
• Figure 1.

• “Figure 2 shows the rate of new construction in Atlanta for single-family and multi-family.”

• “The decline in building was in low-priced new homes. Builders will claim that they can’t build for the low end of the market because their costs are too high. Costs may be high. Productivity in construction has been poor for a long time and there are creeping regulations. But everything looks like a cost problem to builders. In 2012, when the existing homes in ZIP code 30344 were selling for 1.5x incomes, the builders couldn’t profitably build new homes at that price.”
• “Also, at odds with that, apartments, which generally face more stringent local regulations and are generally smaller and more complicated to build, recovered faster and more completely than single-family construction. They recovered more quickly on price too. Figure 3 compares the Case-Shiller estimate for single-family home prices to the Costar estimate of multi-family home prices.”
• “Others believe there had been a bubble and so they still maintain that from 2008 to 2012, home prices in ZIP code 30344 followed a predictable and unsurprising path … There is a supply version and a price version of that assertion. Baker asserted the price version, which is ridiculous on its face when you look at Figure 1. But, the supply version is just as questionable. Where prices had been relatively stable, as in Atlanta, it was because the lending boom before 2008 had supposedly created a massive construction boom. But, construction in Atlanta was no higher in 2005 than it had been in 2000, and then it dropped precipitously in 2006 and 2007. By the time prices bottomed out in 2012, pick any time period length (2 years, 5 years, 10 years, 20 years) and housing construction for that period was the lowest for the period ending in 2012 than it had been in any other period since data have been collected.”
• “The incorrect explanations tend to all suffer from the same problem. They treat a ridiculously deep disequilibrium as a benchmark. On every margin, the housing market had to correct back from that disequilibrium. The prices of homes in ZIP code 30344 could not possibly remain at 1.5x incomes. And so, all the alternative explanations for what happened after 2012 basically observe things that will inevitably happen after a disastrous credit crackdown (more investors, fewer builders, rising rents, etc.) and construct spurious correlations from them.”
• “One more thing to note in Figure 1 is that price/income has recovered in ZIP code 30344 to a level well above the pre-2008 norm. New homes get built when the prices of existing homes are higher than the cost of constructing new ones. If the Atlanta market was producing new housing across both high and low tiers before 2008, then it should be producing new homes today. But, as Figure 2 shows, single family construction is still less than half the pre-2008 norm. This should be inducing new construction.”