The attached policy study originally appeared in the Spring 2018 issue of Housing Finance International. 


Housing finance, as we all know, is lending on fundamentally illiquid assets, taking risk on their prices, which are subject to boom and bust cycles, and doing so on a highly leveraged basis for both the borrowers and the lenders. Naturally this business of ours gets us periodically into severe problems, as has been experienced in numerous countries over time.

Adair Turner, the former chairman of the British Financial Services Authority, goes further. In his provocative book, Between Debt and the Devil (2016), he puts the principal culpability for financial crises – and thus the identity of the Devil – on real estate lending.

He points out that banks in recent decades have changed from being primarily lenders to commerce and industry, to being primarily real estate lenders. In the U.S., this fundamental shift in bank credit toward concentration in real estate dates from the 1970s.

Lord Turner writes:

“In 2007, banks in most countries had turned primarily into real estate lenders.”

“Before the mid-twentieth century, banks in several advanced countries were restricted or at least discouraged from entering real estate lending markets.”

“Lending against real estate… generates self-reinforcing cycles of credit supply, credit demand, and asset prices.” (The interaction of real estate prices and lending is without question a key risk dynamic.)

“At the very core of financial instability in mod-ern economies thus lies an interface between an infinite capacity [to inflate mortgage credit] and an inelastic constraint [real estate].”

Thus, the conclusion: real estate finance and mortgages “are not just part of the story of financial instability in modern economies, they are its very essence.”

Quite an indictment. If it is not the whole truth, it has at least an important element of truth.

In this context, we should consider the instructive history of the laws governing real estate lending by U.S. national banks. These are the banks chartered by the U.S. Government and regulated by the Comptroller of the Currency in Washington, D.C., as opposed to the banks chartered by individual states of the United States. Both exist, but before the American Civil War of 1861-65, all banks were state banks. There are now 943 national banks in the U.S. with aggregate assets of $11 trillion, and 4,075 state-chartered banks with assets of $5 trillion.

National banks make a good study in real estate lending because we can go right back to their creation by the National Currency Acts of 1863 and 1864, later renamed the National Banking Act.

The authors of the original National Banking Act took an unfavorable view of having real estate loans and real estate risk included in the assets of the new national banks, the liabilities of which were going to form the nation’s new currency. They addressed their concern in a simple way: the new national banks were prohibited from making any real estate loans at all!

This seems amazing now, when national banks have $2.5 trillion of real estate loans, or 43 percent of all their loans. On top of that, they own $1.3 trillion of securities based on real estate (mortgage-backed securities), which represent 58 percent of their bond portfolios. (For state banks, real estate loans are 57 percent of total loans and mortgage backed securities are 55 percent of their total bonds.)

The prohibition of real estate loans for national banks lasted about 50 years, until 1913. Although the sponsors of the National Banking Act had intended for national banks completely to replace the state banks, instead the state banks survived and then multiplied, and the national banks felt the competitive pressure.

The first statutory permission for national banks to expand into real estate came as part of the Federal Reserve Act of 1913. This allowed national banks to make real estate loans on farm land only. (In those days, half the population of the U.S. was rural. Congress would expand agricultural lending further with the creation of the Federal Farm Credit System in 1916.) But loans from national banks were limited by the law to 50 percent of the farm property’s appraised value – very conservative, we would say.

The 1913 Act included another basic financial constraint: that real estate loans had to be explicitly tied to more stable bank funding. So, at that point, total real estate loans were limited to a maximum of 33 percent of a national bank’s savings deposits. The idea was that deposits payable on demand should not be invested in real estate financing. The same idea was shown in traditional mortgage lending theory with what used to be called the “special circuit” for funding housing finance. This meant using more stable savings accounts, often in earlier days viewed as “shares,” a kind of equity, and not as deposits – the point being to match more appropriate funding to longer-term residential mortgages. Today we pursue the same goal by the creation of mortgage-backed securities or covered bonds.

An additional limitation of the law was that real estate loans were limited to 25 percent of a national bank’s capital. In contrast, for national banks as a whole today, they represent 256 percent of the tangible capital. For state banks, this ratio is 359 percent.

The limitation to farm real estate for national banks lasted only to 1916, when the law was changed to allow loans on nonfarm real estate, but with a maximum maturity of one year. In 1927, this was expanded to five years on improved urban real estate, with the loan still limited to 50 percent of appraised value.

Vast defaults and losses on real estate lending marked the Great Depression of the early 1930s. Jesse Jones, the head of the Reconstruction Finance Corporation, memorably described “the remains of the banks which had become entangled in the financing of real estate pro-motions and died of exposure to optimism.”

However, in following decades the long-term trend for more expansive real estate lending laws continued apace. Allowable loan-to-value ratios increased to two-thirds, in some cases to 90 percent, maximum maturities were increased to 30 years, and the limit on total real estate loans to 70 percent and then 100 percent of time and savings deposits. In 1974, unimproved land was added as acceptable collateral for national banks. In 1982, the final step in statutory evolution was taken: all statutory real estate lending ratios and formulas were removed by the Garn-St. Germain Act of that year. The 1980s and early 1990a featured euphoric real estate credit expansions and then multiple real estate busts.

In 1994, pursuing further expansion of real estate credit, the administration of President Clinton adopted a political real estate lending campaign: the “National Homeownership Strategy.” The idea was to promote so-called “creative financing” – in other words, the U.S. government was pushing for low and no-down payment mortgages and other risky and low-quality loans. The authors of the National Banking Act would have been appalled by this project. They would have accurately fore-casted its disastrous outcome, which arrived in due course as a contributor to the Great Housing Bust and panics of 2007-08.

That, of course, was the crisis which gave rise to Lord Turner’s book, its diagnosis so unflattering to real estate lending, and to his key prescription:

“To achieve a less credit-intensive and more stable economy, we must therefore deliberately manage and constrain lending against real estate assets.”

In this context, “we” means the government, which must, on Lord Turner’s view, constrain real estate lending, not promote it.

Representatives of housing finance like us may or may not agree that this is the right answer, but we can observe that it is consistent with the statutory limitations on real estate lending provided in the National Banking Act as originally designed and during its first century.


Image by ra2studio