Mortgages Without Maturity Transformation

Carl Milsted, Jr on Jul 5 09:10:17

View QTML

Maturity transformation is the root of many financial evils. If an isolated bank lends out money from checking and passbook savings accounts to homebuyers wanting 30 year mortgages, all is well as long as deposits keep up with checks and withdrawals. Should too many customers want too much cash at the same time, an embarrassing situation results. Should other customers hear of the problem, a bank run ensues. Watch “It’s a Wonderful Life” for a dramatic illustration.

And so our government tries to dampen the inherent instability of the arrangement by connecting banks together. We have interbank lending, the Federal Reserve System, deposit insurance, FANNIE MAE, FREDDIE MAC, and more. When one bank runs low on cash, it can borrow against its portfolio of outstanding long term loans. With enough banks, the Law of Large Numbers should prevent bank runs as long as overall cash reserve rates are kept high enough. Cash poor banks should be offset by cash rich banks.

But they aren’t. The economy is connected. The Law of Large Numbers only works when the large number of entities is uncorrelated. In our interconnected system, bank runs have been replaced with widespread crashes in the financial system followed by bailouts, inflation, and/or nationalization of large chunks of the economy.

To have a stable market-based financial system, long term loans need to come from patient capital. A 30 year mortgage needs some investors willing to wait up to 30 years to get their money back. But who wants to do that?

To start with, we need to bring back the yield curve. Long term loans should have significantly higher interest rates than short term loans. (And to allow that we need to do something about the national debt, else the government goes bankrupt.) With a steep enough yield curve, bank IRAs become a worthy tool for retirement planning. Conservative investors could go to their local bank and invest in Main St. instead of putting their money into Wall St. index funds (which perpetuate today’s big corporations at the expense of new businesses).

With a crackdown on maturity transformation the government can get out of the business of inflating the currency. You won’t need a degree in high finance in order to plan for retirement.

But what about the American Dream of home ownership? Can there be enough money to finance 30 year mortgages without the financial wizards playing a perpetual game of chicken with impatient money? To arrive at a happy conclusion, we need to resort to a bit of accounting magic.

A Bit of Accounting Magic

From the consumer perspective a home mortgage is a large loan where you pay the interest as you go, along with a bit of principal. The arrangement is reasonably well understood by consumers, and allows the government to subsidize new home owners via the mortgage interest deduction. (Whether the latter feature is a good idea is another matter…)

For the bank perspective, let us use a different picture. Instead of thinking of mortgage payments as mostly interest on a very slowly decreasing pile of principal, treat each payment as a zero coupon bond. That is, for each “bond” compounded interest and principal are paid in one lump sum.

With this accounting perspective, the first payment is simply a one month loan. And the payment is mostly principal on said loan. A bank can safely use impatient money for the principal of the early payments, and medium term CDs to finance the first few years of a mortgage. Since impatient money receives a low interest rate, the bank earns a nice spread on the interest portion of the early payments.

The last payment is a thirty year bond. The bank does need some very patient money (retirement accounts or its own equity) to finance the principal for this payment. But note that the bank only needs money for the principal. Unless interest rates are very low, this last payment is primarily accumulated interest.

Let’s run some numbers. To keep the math and graphics simple, I’ll illustrate with a simplified mortgage with annual payments instead of monthly payments, and use annual compounding of the interest. Using P for payment, E for principal portion of a payment (i.e., home equity), and i for interest rate (as a fraction), we have:

$P_1=E_1(1+i)$

For the last payment, we have:

$P_{30}=E_{30}(1+i{)}^{30}$

Ir rearranging:

$E_{30}=\frac{P_{30}}{(1+i{)}^{30}}$

For a 4% loan compounded annually, this is:

$E_{30}=\frac{P_{30}}{(1.04{)}^{30}}$

or$E_{30}$ is 30.8% of the last payment. Most of the last payment is accumulated interest. The bank only needs patient money for the principal portion. We can graph all the payments as follows:

The green area is the initial money the bank needs on hand to make the loan. Green money to the right needs to be more patient than that on the left.

What happens when patient money becomes scarce? Interest rates go up. Let’s look at the same graph for a 7% interest mortgage:

We need less patient money as the interest rate goes up. We thus have a negative feedback loop. Negative feedback is the key to stability in any system, including a financial system. There are actually three channels of negative feedback:

1. With higher interest rates for long term deposits, people have a greater incentive to put away money for the long term.
2. When patient investors are more generously rewarded, they become richer, which allows them to deposit more.
3. Long term loans require a smaller proportion of patient money, since the later payments are mostly accumulated interest.

The Cost of Accounting Magic

At this point I suspect some there are some skeptical gold bugs and Austrian School economists in the audience who think I’m hawking financial fairy dust. Not true! This change in accounting is more than a paper gimmick; accounting has consequences.

Consider those first payments financed with cheap impatient money. While it is true that the bank earns a huge spread on the interest portion of those payments, the interest portion on those payments is tiny. The bank needs to wait for the real profits to roll in. That means no paying out bonuses, dividends or taxes to speak of for those early payments. Mortgage lending is a long game and should be treated accordingly. To abide by this discipline there may need to be changes in accounting standards and tax law. I’ll leave the details to the accountants in the audience.

Another consequence is that a missed early payment counts against the bank’s pool of principal (vs. profits). Alarm bells go off early. This should make banks more careful on whom they lend to.

Finally, banks need to be able to offer financial instruments more complicated than long term CDs to get patient deposits.

When I said that we could treat each payment as a zero coupon bond, I oversimplified. Homeowners have been known to refinance their mortgages, or even sell their houses before 30 years is up. A better model for these payments is to treat them as callable zero coupon bonds. (I think that’s the right terminology. Finance wizards, please correct me.)

So what happens when interest rates drop and homeowners rush to refinance using this accounting?

Banks go under, that’s what.

Banks need to be able to issue callable long term CDs.

Yes, this complicates banking from the customer side. But with the Internet, consumers could easily roll over their portfolios online as long term CDs get called. You could also have online auctions to determine interest rates. Those trying to cash out their portfolios could do so as part of the online auction system vs. paying bank-set penalties for early withdrawal.

That said, for individual retirement accounts, banks could offer intelligent agents to automate the rollover process as well as offer graceful exits by offering cash when replacement deposits arrive.

We now have the technology to implement stable, local banking.

Tags: sustainable economics banking Austrian economics