Science1 publisher3 min readPublished
Loan-to-value ratios held steady through the US housing boom and crash, a UVA study finds
UVA's W. Ben McCartney found low-down-payment mortgages common across 25 years of US data, before, during and after the housing bubble. The lenders changed, so explanations of the boom turn to income limits, paperwork and buyers' expectations.
The Scientist · Science desk
What happened
- W. Ben McCartney of UVA's McIntire School of Commerce examined 25 years of US mortgage data to test whether easy small-down-payment loans drove the housing bubble.
- Loan-to-value ratios stayed remarkably stable over the period, including in places where home prices boomed and crashed hardest.
- Since 2020, as home prices rose sharply, loan-to-value ratios declined somewhat, so mortgaged buyers financed a smaller share of their homes with debt.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- constraint Lessons that pin the 2000s bubble on looser down-payment rules lose their evidence, because the share borrowed did not rise even where prices swung hardest.
- exposure Borrower leverage looked the same throughout, so accounting for who was exposed in the crash means tracing which lenders and agencies backed the loans in each phase.
- decision Anyone testing causes of the boom now has to examine debt-to-income limits, documentation, loan products and buyer expectations; loan-to-value data cannot measure them.
- cost A buyer putting down a larger share since 2020 can still carry a bigger dollar loan at a higher rate, so the falling ratio says little about affordability.
The design tests a prediction. If looser down-payment rules had pushed prices up, buyers should have borrowed a growing share of their homes' value as the boom built. That growth should have been largest where prices ran furthest. "If easier down-payment requirements were pushing prices higher, we should see buyers borrowing an ever-larger fraction of their homes' values during booms. We don't," McCartney told UVA Today [6][14]. Comparing places is the nearest thing here to a control. Even in markets with the sharpest boom and crash, the loan-to-value ratio stayed stable [7].
The story under test is the standard account of the early-2000s bubble that crashed in 2007-08 and set off a foreclosure crisis and a deep recession [1]. In McCartney's paraphrase of it: "Suddenly, people could buy homes with only 5% down, which brought a flood of highly leveraged buyers into the market, and home prices skyrocketed" [3]. Five percent down is a 95% loan-to-value ratio [16]. He said that story "is essentially at odds with the data" [3].
The lenders changed. Before the boom, the Federal Housing Administration and the Department of Veterans Affairs backed many low-down-payment loans. During it, private lenders took over much of that market, and after the crash the two agencies stepped back in [5]. "But the use of low-down-payment mortgages hardly changed at all," McCartney said [4]. I think that is the part of the work with the most bearing on the lessons drawn from 2008. It moves the risk question from how much buyers borrowed to who took on the risk, and under what other terms [13][5].
"Our results tell us more about what didn't cause housing booms than they do about exactly what did," McCartney said [9]. He named credit terms that a loan-to-value ratio does not capture: debt-to-income limits, documentation standards, mortgage products and other lending terms [10]. A 95% loan made with full income paperwork and one made with none show the same ratio [16][10]. He also pointed to expectations. Buyers convinced that prices will keep rising may pay much more without being allowed to borrow a larger fraction of the home's value [11].
The denominator matters for the recent data too. Since 2020, as home prices rose sharply, loan-to-value ratios declined somewhat, meaning buyers with mortgages put more equity into their purchases [8]. McCartney cautioned that a smaller ratio on a far more expensive house can still be a much larger loan in dollars, and higher interest rates make those dollars cost more to borrow [12]. He said higher rates pull two ways: they cut what buyers can afford, and they can reduce supply because owners with low-rate mortgages may be reluctant to sell [15].
The interview does not give the loan-to-value figures themselves, name the dataset, or say where the research was published [14].
What to watch
- Publication of the full paper, with its dataset and loan-to-value figures, and whether the stability holds across the distribution of borrowers as well as on average.
- Studies that test debt-to-income limits, documentation standards and loan products against the same 25-year record.
- Whether the FHA and VA share of low-down-payment lending shifts again as private lenders respond to higher rates.