Professor Krugman on maturity transformation

There’s always time for a short lecture on UR’s favorite topic. It’s important to remember that Paul Krugman is an idiot, but he’s not a fool:

Like a lot of people, my insights draw heavily on Diamond–Dybvig, one of those papers that just opens your mind to a wider reality. What DD argue is that there is a tension between the needs of individual savers—who want ready access to their funds in case a sudden need arises—and the requirements of productive investment, which requires sustained commitment of resources.

Banks can largely resolve this tension, by offering deposits that can be withdrawn on demand, yet investing most of the funds thus raised in long-term, illiquid projects.

I face a similar tension between going to Reno every weekend, and buying strained carrots to feed my one-year-old. The Fed could largely resolve this tension, by printing fat stacks of Benjamins for me to blow on coke and whores before I hit the Safeway for some Gerber. I note also that the policy would create demand—a favorite consequence of Professor Krugman’s.

Just as the entire purpose of a monetary system is to decide who gets to go to Reno and who has to scrimp and save for strained carrots, the entire purpose of an interest-rate market—as well understood when Professor Krugman was eating strained carrots—is to match the supply and demand of loanable funds and eligible borrowers at every duration.

If you convert loanable funds of one duration into loanable funds of another duration, either by wholesome George Bailey banking or by synthesizing collateralized instruments (a category which logically includes nominally zero-term demand deposits), you are taking this elegant market signal, the yield curve, and raping it in the ass. You will give it AIDS. It will give you AIDS back. This will become known as the “business cycle”—a sort of historical quartan ague. Though no one understands it, it exists.

And both Wall Street and Main Street will exhibit a pattern of unending financial crises for all of modern Anglo-American history—from Walter Bagehot to Secretary Geithner. I’m sure this couldn’t be due to a defective, archaic banking system which wasn’t even redesigned for the 20th century, let alone the 21st.

Indeed, as Bagehot’s Wiki—not cattily—notes:

Bagehot’s observations on finance remain relevant and cited by central bankers, most recently in the wake of the global financial crisis that began in 2007.

Indeed. I could hardly put it better myself. I’ve proposed before that classic Lombard Street banking, borrowing short and lending long, should be forever known as a “Bagehot scheme.” Isn’t it fascinating that while so many other 19th-century English institutions—like slavery, the gallows, and impressment of sailors—have met their demise, this one continues merrily on? Talk about a barbaric relic!

Professor Krugman is unintentionally wonderful on this point:

The problem, of course, is the vulnerability of such a system to self-fulfilling panics: if people believe that a bank will fail, everyone will in fact want to withdraw funds at the same time—and because the bank’s assets are illiquid, trying to meet those demands through fire sales can in fact cause the bank to fail.

Actually, Professor Krugman, the Diamond–Dybvig model is not a true multiple equilibrium. The maturity-matched model—in which long-duration asset prices are much lower—is the only free-market equilibrium.

If you can fix asset prices, of course—whatever. But in a free market, the value of a synthetic asset is always epsilon less than the value of the equivalent real asset. By definition the synthetic asset can default, whereas the real asset can’t. (In a George Bailey bank, for instance, your demand deposits are loans to the bank collateralized by the bank’s portfolio of burned-out Section 8 New Deal ghetto towers.)

No perfect system of collateralization can be constructed. Epsilon exists. As free markets become frictionless, epsilon becomes tradable. The bank run happens automatically. Intervention is required to prevent it:

This then leads to the need for policy: deposit insurance and/or lender of last resort facilities to head off bank runs, and bank regulation to reduce the moral hazard from these explicit or implicit guarantees.

Unless the lent asset is USG shares—i.e., dollars. USG can issue and lend as many USG shares as it wants. By engaging in this practice, it can lower interest rates to zero across the duration curve. Indeed, it is in the process of doing so. In theory, Google could just as easily operate a Bagehot scheme in GOOG shares. They are neither idiots nor fools, so they don’t.

In the end state of this pernicious practice, there are no “private” banks at all. There is just one big bank: the government. Congratulations, Professor Krugman! You’ve reinvented the Soviet Union. Could you get a second Nobel for this mighty discovery? When the zero bound is hit across the curve, there is no lending even at zero interest rates, and capitalism is officially flatlined. Instead of infinite stimulation, this is the point of infinite stagnation. All economic organization becomes the task of the government. Soy Cuba! Yo, Cuba!

A “lender of last resort facility” is a crucial piece of machinery in the Bagehot scheme. In all cases, “loan guarantees” can be modeled simply as loans. If A guarantees B’s loan to C, what is really happening is that B lends to A, and A to C. A in this case being our friend, tha USG. Or more specifically, the Fed.

So, when you “deposit” dollars “in” a bank, not only are you really lending them to the bank—you’re really lending them to the Fed. Moreover, when a bank lends you dollars, you are really borrowing from the Fed. Yo, Cuba! Never in the history of Bagehot schemes has this been more clear. Fortunately, at least we’re not on a gold standard, under which no Bagehot scheme can survive (USG being a perfect credit risk for USG equity, and nothing else). Naturally, this is Professor Krugman’s most devastating argument against the gold standard.

What I love—what really illustrates the difference between an idiot and a fool—is the bizarre overall thrust of the Professor’s argument, which clashes so baldly with the rest of his political idiocy. Graphite-cooled plutonium reactors, Professor Krugman tells us, are essential, because they generate electricity. How else can you generate electricity?

(Obviously, if we did not synthesize loanable funds at 30-year duration, no one would invest in 30-year mortgages, because no one saves money with the intent to spend it 30 years later. Thus, the price of a 30-year dollar would be a present nickel. And we will all be zillionaires when we retire, because each of our 2011 nickels will buy a 2041 dollar. Hey, wait…)

But graphite-cooled plutonium reactors also have a tendency to burn when they melt down, contaminating half the Ukraine. We need electricity, though! So we’ll just have to cover the Ukraine with plastic sheeting, which teams of Mexicans in bunny suits can wipe down every time there’s another Chernobyl. As an added benefit, this will create demand and stimulate the Mexican economy. Also, tomatoes can be grown under the sheeting.

No fool could come up with this. But an idiot could. Also, one lovely benefit of this dangerous brush with reality is that we have the opportunity to hear the Professor’s rare, shy and beautiful libertarian side—the prothonotary warbler of the liberal conscience. Regulation, you see, is hard:

So, are you going to ban fractional reserve strategies by money market funds? Are you going to ban repo? Auction rate securities? Where does it stop?

Professor! I know! I know! Hey, look at me! “Yes,” “yes,” “yes,” and “a long way farther on.”

They have an oddly antiquated notion of what money and finance are about, one that misses the “virtualness” of the modern world. They still think of money as being pieces of green paper, rather than what it mostly is now, zeroes and ones in some server somewhere. They still think of banks as being those big marble buildings, in a world in which most banking is a lot more abstract than that.

This is, after all, the 21st century. Things have moved on a bit.

No, actually. They haven’t. They should, though.

Now, to be fair, Professor Krugman is a fool but not an idiot. If there are any Austrians left who can’t understand that fractional reserve is a special case of maturity transformation, and they prefer to consult living professors, they can ask Philipp Bagus. If they prefer the idol himself:

For the activity of the banks as negotiators of credit the golden rule holds, that an organic connection must be created between the credit transactions and the debit transactions. The credit that the bank grants must correspond quantitatively and qualitatively to the credit that it takes up. More exactly expressed, “The date on which the bank’s obligations fall due must not precede the date on which its corresponding claims can be realized.”

In other words, a healthy bank—virtual or marble-pillared—expects to meet all its obligations from cashflow, without new borrowing to pay off old loans. Well, knock me over with a feather. If that’s not simplistic, I don’t know what is. That’s what accounting should be—simplistic.

The genius of Professor Krugman is that he goes so near the truth that he makes it obvious even to his commenters—who typically are both idiots and fools, but several of whom spontaneously exhibit the same insight themselves:

Why can’t we regulate or even ban the maturity mismatch? Savers would have to make the maturity choice themselves and it would be transparent. Currently, the savers don’t understand the huge run risks that the banks have by funding with demand deposits and lending long. It’s hiding the risk.

And almost cogently:

I’m not a hard-money crazy, but I have wondered if there’s a weaker version of hard money that makes sense: forcing duration matching. Is borrowing short-lending long actually a required service? What if the bank only lent long money borrowed long such that the assets and liabilities matched in duration?

It’d still be fractional reserve banking, but it would not be subject to bank runs. While we’re on the subject of hard-money feasibility, can you comment on this variant?

Somehow I don’t expect an answer.