What is really behind the mushrooming rate of mortgage foreclosures since 2007? The evidence from a huge national database containing millions of individual loans strongly suggests that the single most important factor is whether the homeowner has negative equity in a house — that is, the balance of the mortgage is greater than the value of the house. This means that most government policies being discussed to remedy woes in the housing market are misdirected.

Many policy makers and ordinary people blame the rise of foreclosures squarely on subprime mortgage lenders who presumably misled borrowers into taking out complex loans at low initial interest rates. Those hapless individuals were then supposedly unable to make the higher monthly payments when their mortgage rates reset upwards.

But the focus on subprimes ignores the widely available industry facts (reported by the Mortgage Bankers Association) that 51% of all foreclosed homes had prime loans, not subprime, and that the foreclosure rate for prime loans grew by 488% compared to a growth rate of 200% for subprime foreclosures. (These percentages are based on the period since the steep ascent in foreclosures began — the third quarter of 2006 — during which more than 4.3 million homes went into foreclosure.)

Sharing the blame in the popular imagination are other loans where lenders were largely at fault — such as "liar loans," where lenders never attempted to validate a borrower's income or assets.

This common narrative also appears to be wrong, a conclusion that is based on my analysis of loan-level data from McDash Analytics, a component of Lender Processing Services Inc. It is the largest loan-level data source available, covering more than 30 million mortgages.

There's a very interesting graphic that goes with the story, titled "No Skin in the Game" summarizing the data.

The analysis indicates that, by far, the most important factor related to foreclosures is the extent to which the homeowner now has or ever had positive equity in a home. The accompanying figure shows how important negative equity or a low Loan-To-Value ratio is in explaining foreclosures (homes in foreclosure during December of 2008 generally entered foreclosure in the second half of 2008). A simple statistic can help make the point: although only 12% of homes had negative equity, they comprised 47% of all foreclosures.

Further, because it is difficult to account for second mortgages in this data, my measurement of negative equity and its impact on foreclosures is probably too low, making my estimates conservative.

What about upward resets in mortgage interest rates? I found that interest rate resets did not measurably increase foreclosures until the reset was greater than four percentage points. Only 8% of foreclosures had an interest rate increase of that much. Thus the overall impact of upward interest rate resets is much smaller than the impact from equity.

To be sure, many other variables — such as FICO scores (a measure of creditworthiness), income levels, unemployment rates and whether the house was purchased for speculation — are related to foreclosures. But liar loans and loans with initial teaser rates had virtually no impact on foreclosures, in spite of the dubious nature of these financial instruments.

Instead, the important factor is whether or not the homeowner currently has or ever had an important financial stake in the house. Yet merely because an individual has a home with negative equity does not imply that he or she cannot make mortgage payments so much as it implies that the borrower is more willing to walk away from the loan.

Update, and thanks to Mark Field in the comments, here is Barry Ritholz responding:

As to prime versus sub-prime, it appears the Mortgage Bankers Association, data dispute the professor's. Jay Brinkmann, chief economist for the MBA, noted in May 2009 that in 2008, prime, fixed-rate loans were only 19% of foreclosure starts nationwide, while Subprime adjustable-rate mortgages were 39%. More recently, the two levels have come together: prime loans are up to 29% of foreclosure starts while subprime adjustables came down to 27%.

But reporting only in percentages can be misleading. As Floyd Norris noted in August of 2008, "There are far more prime mortgages than subprime, of course, and subprime loans are much more likely to get into trouble. But this does show how the foreclosure problem is spreading."

Agreed.

But the claim that during this crisis it has been Prime and not Subprime is simply unsubstantiated by the timeline or data. Subprime went bad first, then Alt-A, and then prime followed it later. Sub-prime and Alt-A went bad due to poor lending standards; Prime went bad in part due to job losses and as the economy got worse.

If anything, there is a stronger argument to make that the problem is worse from 30 year fixed versus ARMs. Here is the MBA data from September 2008:

For prime loans, foreclosure starts on fixed rate loans were 0.34 percent, an increase of five basis points, while prime ARM foreclosure starts were 1.82 percent, a 26 basis point increase. For subprime loans, fixed rate foreclosure starts increased 27 basis points to 2.07 percent and subprime ARM foreclosure starts increased 31 basis points to 6.63 percent

Sub-prime worse than Prime, ARMs much worse than fixed.

Of course, it is true that 100% LTV mortgages are a problem. But you need some context to understand how they came about. And while the professor does correctly identify underwater mortgages as a major factor — he seems to place the blame squarely on 100% LTV. Perhaps another question worth exploring is the boom/bust issue: How did those home prices run up so much, only to reverse back towards normal, historical pricing metrics? For that, you need to look at many factors.

A more comprehensive 40,000 foot view would note that 100% LTV is a symptom of the larger problem of a) abdication of lending standards, caused by b) enormous demand for securitized loans, enabled by c) rating junk as AAA, in order to satisfy the demand for higher-yielding, non-junk paper, all of which traces its roots to d) Greenspan's ultra low interest rates.

Yes, bad lending standards, no money down, lack of income verification or debt servicing ability were key culprits. But to claim that it was more Prime than sub-prime is belied by the history of foreclosures. And, it ignores all the other moving parts to the equation.