“Flip This House”: Investor Speculation and the Housing Bubble
Andrew Haughwout, Donghoon Lee, Joseph Tracy, and Wilbert van der Klaauw
The recent financial crisis—the worst in eighty years—had its origins in the enormous increase and subsequent collapse in housing prices during the 2000s. While the housing bubble has been the subject of intense public debate and research, no single answer has emerged to explain why prices rose so fast and fell so precipitously. In this post, we present new findings from our recent New York Fed study that uses unique data to suggest that real estate “investors”—borrowers who use financial leverage in the form of mortgage credit to purchase multiple residential properties—played a previously unrecognized, but very important, role. These investors likely helped push prices up during 2004-06; but when prices turned down in early 2006, they defaulted in large numbers and thereby contributed importantly to the intensity of the housing cycle’s downward leg.
Virtually everyone who buys a house is hoping for prices to rise, and most use leverage (debt)—in this case, a mortgage—to allow them to buy more housing than they can afford to pay for in cash. While the majority of borrowers have a consumption motive—as “owner-occupants,” they intend to live in the house—some borrowers own housing purely as an investment. As mortgage lenders have long known, investors are more likely than owner-occupants to walk away from an underwater property. So when a borrower acknowledges on the mortgage application that she won’t live in the house, the lender will typically require a higher down-payment and charge a higher interest rate to reflect the additional default risk. Within the category of real estate investors, some buy properties with the intention of renting them out, while others intend to simply “flip this house,” selling quickly and reaping a capital gain.
The charts below show the shares of new-purchase mortgage debt going to investors with multiple first-lien mortgages on their credit reports, as reported in the FRBNY Consumer Credit Panel (CCP). Since each property can have just one first lien—the primary mortgage—the number of first liens reflects the minimum number of properties a borrower owns (“minimum” because people can of course own houses free and clear). In the charts, the colors of the bars indicate how many properties the borrower owns after taking out this new-purchase mortgage debt; we show data for the country as a whole on the left and for four states that experienced especially pronounced housing cycles—Arizona, California, Florida, and Nevada—on the right.
The charts reveal some astonishing facts. At the peak of the boom in 2006, over a third of all U.S. home purchase lending was made to people who already owned at least one house. In the four states with the most pronounced housing cycles, the investor share was nearly half—45 percent. Investor shares roughly doubled between 2000 and 2006. While some of these loans went to borrowers with “just” two homes, the increase in percentage terms is largest among those owning three or more properties. In 2006, Arizona, California, Florida, and Nevada investors owning three or more properties were responsible for nearly 20 percent of originations, almost triple their share in 2000.
Investors Are Different from Owner-Occupants
Whether they were buying vacation homes or flipping houses, real estate investors behaved very differently from borrowers with just one first lien, a group almost certainly dominated by owner-occupants. For one thing, investors are very unlikely to move to the house they bought, especially if they own three or more properties. In rising markets, “buy-and-flip” investors typically want to hold properties for relatively short periods, and we show in our study that multiple-property owners in the mid-2000s tended to sell their properties much more quickly than those with just one first-lien mortgage. Importantly, we also show how buy-and-flip investors can make higher bids on houses, even if they had relatively little cash, by using low-down-payment loans. Nonprime credit—mortgage lending to borrowers who were unable or unwilling to qualify for cheaper, prime loans—enabled optimistic investors to speculate by making highly leveraged bets on house prices.
Because investors don’t plan to own properties for long, they care much more about reducing their down-payments than reducing their interest rates. The expansion of the nonprime mortgage market during the 2000s provided the perfect opportunity for optimistic investors to get low-down-payment credit, albeit at high interest rates. As shown in the charts below, investors were far more likely than owner-occupants to use nonprime credit to make their purchases, especially at the peak. Again, the colors show the number of properties the borrower owns, but this time we leave in the single-property borrowers for comparison.
We created these charts by matching our CCP sample (which is based on Equifax credit reports) with data from CoreLogic's Loan Performance data set of securitized nonprime mortgage loans. An additional benefit of this matched sample is that it helps to explain why we are among the first to notice the importance of investors in the mortgage market during the 2000s. The next two charts report shares of mortgage lending using two distinct measures of investor status. The first measure, shown by the dashed lines, is what’s reported in the Loan Performance data: whether the borrower self-reported to the lender that she wouldn’t be an owner-occupant. The second measure, represented by the solid lines, is our preferred measure from the CCP: Does the borrower own more than one house when this purchase is complete? In the crucial nonprime sector, the share of borrowers who acknowledged their investor status was relatively low and constant—less than 20 percent throughout the boom. But by our definition, at the peak the investor share in this market was more than double the self-reported rate.
Did It Matter in the Bust?
So far, we have half the story: Optimistic investors—speculators—used low-down-payment, nonprime credit to place highly leveraged bets on the housing market, perhaps facilitated for some by reporting an intention to live in the house. Because they didn’t have to put much money at risk, these investors were able to continue to buy housing even as prices rose further. All of these developments were especially noticeable in Arizona, California, Florida, and Nevada. Longstanding tradition in the mortgage lending business and the predictions of economic models hold that investors will quickly default if prices begin a persistent fall. This is what happened starting in 2006, as the charts below show.
An interesting feature that’s hard to see here but that we document in our study is that borrowers with multiple mortgages start out being better risks—their loans were less likely to become seriously delinquent before 2006—but end up accounting for a disproportionate amount of defaults thereafter. What changed in 2006? Prices started to fall. In 2007-09, investors were responsible for more than a quarter of seriously delinquent mortgage balances nationwide, and more than a third in Arizona, California, Florida, and Nevada. While this sharp change in the risk assessment of owners of three or more properties may not seem surprising, it also applied to investors with “just” two homes. If there were reason to believe this latter group was less prone to act like investors, the data don’t support this view.
Lessons to Learn?
We conclude that investors were much more important in the housing boom and bust during the 2000s than previously thought. The availability of low- and no-down-payment mortgages in the nonprime sector enabled investors to make these bets. This may have allowed the bubble to inflate further, which caused millions of owner-occupants to pay more if they wanted to buy a home for their family. In the end, even the value of the 20 percent down-payments made by responsible, prime borrowers was wiped out—leaving the housing market, and the economy, in the vulnerable state we find them in today.
The idea that asset price booms may be driven by optimistic or speculative investors who make highly leveraged bets on asset prices, then quickly default if their expectations are not realized, is not new. Indeed, John Geneakoplos of Yale University argues that such behavior is a fundamental driver of what he calls the “leverage cycle.” To our knowledge, our study provides the first direct evidence that such behavior may have been important in the 2000s housing cycle.
But what, if anything, does it teach us about policy? We conclude that it’s very important for lenders (and regulators) to manage leverage as asset bubbles are inflating. In the 2000s, securitized nonprime credit emerged to allow leverage to increase, with effects that extended far beyond this sector, including spillovers from defaulted mortgages to the value of other properties (see Campbell, Giglio, and Pathak ). Effective regulation of speculative borrowing, like what is being attempted in China today, may be needed to prevent this kind of crisis from recurring.
The views expressed in this blog are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of New York, or the Federal Reserve System. Any errors or omissions are the responsibility of the authors.