## Maturity transformation considered harmful: an unauthorized biography of the bank crisis

“You may not be interested in war,” Trotsky once said, “but war is interested in you.” Finance, too.1

Perhaps you are under the impression that banking, accounting and economics are perfectly understood by the bankers, accountants and economists. Or did this week cure you of that? If it didn’t, next week might. Generals, too, deserve deference for their understanding of war. But not unlimited deference.

I am not a banker, an accountant, or an economist. I am a computer programmer. My approach to financial engineering is to analyze it from the outside, taking nothing for granted, treating it as I would a new operating system or programming language.

Our financial system is not a new operating system. It is a very old operating system. Worse, there is only one of them: the whole world runs the Anglo-American banking system, more or less as described by Walter Bagehot in Lombard Street (1873). Lombard Street is our Windows. There is no Mac. There is no Linux. Our experts in finance are not experts in finance. They are experts in Lombard Street finance. Asking them to imagine an alternative is like asking a Windows programmer to imagine OS X—except that Windows isn’t 314 years old.

(The closest alternative to Lombard Street finance is a relatively obscure branch of economics called the Austrian School. The basic problem with Austrian economics is that it has never been tried in practice, and it has not advanced much since Ludwig von Mises wrote Theory of Money and Credit in 1912. (A more readable modern text is Murray Rothbard’s Mystery of Banking.) The discussion below is generally Austrian in method and theory, but simplified and generalized—and my conclusions are very different from, say, Ron Paul’s.)

In any case, if you’re running Windows and you suspect that you might want to “switch,” a Windows expert is the last expert you want to ask. What needs to happen now, I feel, is that a very large number of very smart people who know nothing about finance need to do what I did, and try to figure the issue out from scratch. Hopefully they will not get the same results.

Because my results are… disturbing. If you suspect that they might be right, please try thinking through the problem for yourself. Let’s go straight to the disturbing results, and then we’ll try to justify them.

1. We do not have a free-market financial system.

2. We have never had a free-market financial system.

3. Leaving the financial system to “work things out on its own” will not produce a free-market financial system. It will produce a smoking heap of rubble.

4. Paulson’s bailout is, if anything, far too weak. Our financial system is part of the government. The proper first step is to stop lying about this. This means nationalizing the banks. This is not an expansion of government, but a recognition of its actual size. It is not an expenditure, but a revision of accounting to reflect reality.

5. A free-market financial system would be way cool. More important, it would be extremely stable. But the only way to create one is to build it right from the start. If you have a car and you want a motorcycle, sell your car and buy a motorcycle. Don’t decide to call your car a “four-wheeled motorcycle,” and don’t think unscrewing two of the wheels will solve the problem.

6. Therefore, the government should close down the financial system we have now and replace it with one that doesn’t suck. What is the probability that this will happen? Zero. But at least you know.

These are the results. Now, the explanation.

When we ask: “what caused the bank crisis,” we need to distinguish between proximate and ultimate causes. Our focus today will be on the ultimate cause. But first, let’s get the proximate cause out of the way.

The proximate cause of the bank crisis is the gigantic vote-buying machine we know and love as the “Democratic Party.” This gave us something called the Community Reinvestment Act, which compelled banks to steer over a trillion dollars in flagrantly bogus loans to the Democrats’ electoral base. The scam is described, and its outcome predicted, in this article from 2000. Here is Barack Obama’s lead economic advisor, endorsing it—in 2007. Doh.

If the gigantic protection racket known as the “Republican Party” had obstructed “diversity lending” in any way, that might be a reason to support them. They didn’t, and it isn’t. Indeed, Steve Sailer has uncovered a hilarious, and apparently improvised, W. speech endorsing the program:

All of us here in America should believe, and I think we do, that we should be, as I mentioned, a nation of owners. Owning something is freedom, as far as I’m concerned. It’s part of a free society. And ownership of a home helps bring stability to neighborhoods. You own your home in a neighborhood, you have more interest in how your neighborhood feels, looks, whether it’s safe or not. It brings pride to people, it’s a part of an asset-based to society. It helps people build up their own individual portfolio, provides an opportunity, if need be, for a mom or a dad to leave something to their child. It’s a part of—it’s of being a—it’s a part of—an important part of America.

To open up the doors of homeownership there are some barriers, and I want to talk about four that need to be overcome. First, down payments. A lot of folks can’t make a down payment. They may be qualified. They may desire to buy a home, but they don’t have the money to make a down payment. I think if you were to talk to a lot of families that are desirous to have a home, they would tell you that the down payment is the hurdle that they can’t cross. And one way to address that is to have the federal government participate.

Truer words were never spoken. The federal government is certainly participating now! So, we have our proximate cause: the Dempublicans. Or possibly the Republocrats. Do we care? Does anyone with any brains still believe in any of these swine?

The ultimate cause, however, is a matter of financial engineering. It has nothing at all to do with elections or politicians. Politics explains where the bad mortgages came from. Politics does not explain why they caused our financial system to lock up like a clogged fuel pump, or why no one can price or sell these instruments. Pricing dubious and complicated securities is what a financial system does. So why isn’t it happening? What is the engineering mistake that caused the financial system to be so sensitive to this relatively minor piece of graft?

The engineering mistake is an accounting practice called maturity transformation. In the best CS tradition—I consider it harmful.

Maturity transformation might also be called monetary time travel. It is an accounting structure which permits a financial institution to pretend that it can teleport dinero from the future into the present. High-tech modern finance can do many cool things, but this is not one of them.

The price we pay for this illusion is a fundamental instability in the lending market. To most economists, this instability is a Diamond-Dybvig dual equilibrium. To Austrian economists, it’s the Misesian theory of the business cycle. And in plain English, it’s your common or garden bank run. The present crisis, which is by no means over, has many fascinating and devilish modern characteristics—but old Beelzebub is easily discerned beneath its raiment.

Modern computerized finance is especially susceptible to the bank-run bug, because in the last twenty years it has adopted an awesome array of incredibly beautiful and fragile modeling techniques. Unfortunately, all of these models assume a stable or single-equilibrium pricing environment. They do not allow for a phase change between dual equilibria. Doh!

Why has this mistake gone unrecognized? The basic problem with maturity transformation (MT) is that, while it is a bad accounting practice, it is not a new bad accounting practice. MT is not some horrendous monetary Guantanamo unleashed upon us by the corrupt Bushocracy. Nor is it a form of financial Bolshevism devised by the Mussulman Candidate, Barack Osama. Oh, no.

Maturity transformation is the heart and soul of the Anglo-American model of banking. Our current round of MT dates to the founding of the Bank of England in 1694. Lombard Street (which is still a good read) describes it in a nutshell. MT is inextricably woven into our political system, our accounting system, our profession of economics, even our Wikipedia. Despite the fact that it has been causing booms, busts, and crashes since Pennsylvania Avenue was a swamp.

Or so, at least, I assert. Obviously it’s a somewhat dramatic assertion. But I think it’s possible for you to use your own brain cells to understand why MT is harmful. You don’t need to trust me, or anyone else. The problem is just not that hard to follow.

You may know maturity transformation as “fractional-reserve banking,” which is one common case of the practice. A financial institution practices MT whenever it “borrows short and lends long,” i.e., promises to deliver money in the short term based on the fact that it is owed money in the long term. For example, in a classic fractional-reserve bank which takes checking deposits and uses them to fund mortgages, the bank’s promises have a term of zero (your money is available whenever you want it), and its mortgages are repaid across, say, 30 years.

But few of us have operated a bank. I want to explain intuitively why maturity transformation is a basically corrupt practice. And for that, we’ll need a more down-to-earth example.

Suppose you lend a friend of yours from work, Bobby, a thousand dollars. He agrees to give you the money back, plus fifty for interest, next week. The week is the term or maturity of the loan.

Bobby then lends your K to a friend of his, Dwight. Dwight is about to ride up to Humboldt to spend three boring weeks with his loser parents. Which will suck, but which will also enable him to score a pound of weed and hitch back with it. The weed can be moved for $1500—but the three weeks is non-negotiable. So the glitch arises when you buttonhole Bobby by the water fountain and mention that it’s Monday. “Yeah, right,” he says. “Man, hey, do you want to go in for another week? I’ll give you another fifty. That’s a good rate, man.” This is, you agree, totally fine. You have just rolled over your loan to Bobby. Next week, you do it again. And the week after that, Dwight gets in. He pays Bobby$1300. And Bobby comes to you with $1150, still smelling faintly of Humboldt. You have just had a successful customer experience with the Bank of Bobby and Dwight. You profited. Bobby profited. Dwight profited. But is what Bobby did cool? Is it right? Did it display probity? We’d have to say—no. At least, we would have to say this intuitively. We have not yet applied our logical faculties to the matter. We have just used the awful Bobby, and the still more horrendous Dwight, as propaganda props. Can we smear the noble art of banking with these dank characters? The basic problem with this transaction is that when Bobby said he wanted to borrow a grand for a week, he was lying to you. Bobby is a user. He knew he could get away with stretching the loan from one week to three, and he did. If he’d just asked you for a three-week loan, everything would have been fair and square. But you might have said no. What do you want to know when you lend someone money—whether it’s Bobby, or Citigroup? You want to know that they’ll pay you back. Or at least, that the probability of repayment is high enough to be justified by the interest rate on the loan. To standardize this decision, our ancestors developed the art of accounting. Bobby, or Citigroup, opens the kimono and shows you two lists. One is the list of promises Bobby, or Citigroup, has made. For instance, Bobby has promised to give you$1050 in a week. This is sometimes known as a liability. The second list is the list of things Bobby, or Citigroup, owns—for instance, Dwight’s promise to deliver $1300 in three weeks. This is an asset. As a lender, what you want to know about Bobby or Citigroup is that they are solvent. In other words: Bobby has enough money to pay you back. Solvency is computed by subtracting liabilities from assets, the result being equity. For instance, if these are the only transactions on Bobby’s balance sheet, and if we assume (a big if) that Dwight’s promise is actually worth$1300, then Bobby’s equity is positive $250, and he is solvent. Note that when you consider solvency, you are not asking whether Bobby can pay you back. You are asking whether Bobby can pay all his creditors back. For instance, if Bobby has also borrowed$1000 from your officemate Dave, but lent it to his ex-girlfriend Angelique, who spent it all on meth, his equity is negative $750, and he is insolvent—or bankrupt. He can pay you or he can pay Dave, but he can’t pay both of you. There are two keys thing to remember about insolvency. One, it is a sort of event horizon; you cannot borrow yourself out of bankruptcy. No one has any good reason to lend to an insolvent party. Two, it is a collective decision: Bobby is either insolvent with respect to both you and Dave, or he is solvent with respect to both. He can either make good on his promises, or he can’t. But absent Angelique, when we use the solvency test to evaluate Bobby’s business, it looks like a good one. Bobby is solvent. But he also had to lie to you to get the loan. Something is not quite right here. If Bobby shows his balance sheet to a trusted third party, perhaps an auditor or regulator, the auditor will agree that Bobby is solvent. If you ask Bobby whether he can be trusted, he refers you to the auditor, who tells you that Bobby is solvent. But if you saw his actual balance sheet, you wouldn’t do the deal. Our accounting model is simply not doing its job. Or is it? Enough with Bobby. Let’s look at an actual bank. FooBank, a classic fractional-reserve bank, has a billion dollars in demand deposits—that is, liabilities of zero maturity, which are continuously rolled over when its depositors fail to withdraw them from the ATM. It has$50 million in the vault, and $1.1 billion in fixed-rate 30-year mortgages, giving it equity of$150 million.

Note that $1.1 billion is not the total amount of money owed on FooBank’s mortgages. It is the current market price of the mortgages. I.e.: if FooBank sold the mortgages to some other financial institution, it could get$1.1 billion for them. (Because interest rates are always positive, that means the total payments over 30 years will be much more—let’s say, $1.5 billion.) So our question is: should you put your money in FooBank? If you give FooBank a demand deposit, can you be sure that it will be there when you demand it at the ATM? Our first answer is: yes. Whatever your deposit is, it’s a lot less than$50 million.

Our second answer is: no. Because FooBank, like Bobby, is a maturity transformer. It owes $1B, now. It expects to receive$1.5B—over the next thirty years. Dwight is sure taking a long time up in Humboldt! FooBank has only $50 million in the vault to pay its$1B in demand deposits. If more than 5% of its customers demand the payments they are contractually owed, FooBank is screwed. If you are not among the first 5%, you’re screwed too. What’s FooBank going to do? Will the ATM print out a message telling you to go up to Humboldt, and find Dwight?

Remember, back when we were considering Bobby’s solvency, we established what you want to know as a lender: that the borrower (Bobby) will be able to fulfill all his promises, not just yours. There is no such thing as selective default. And there is absolutely no reason to assume that your fellow depositors won’t all show up at the same time.

When you don’t turn up and withdraw your demand deposit, you are effectively rolling over a loan to the bank. And from the bank’s perspective, there is no difference between rolling over a loan and finding a new lender. So FooBank is staking its undefeated record in promise fulfillment, which might be impressive in a Bobby but is pretty much required in a bank, on paying back its old loans by finding new lenders. Um… where have I heard that before?

Our third answer is: yes. Because we’ve forgotten something, which is that FooBank is solvent. It may not have $1 billion in the vault. But it can raise$1 billion—quite easily, by selling its $1.1 billion worth of mortgages. In a modern digital market, this can be accomplished on the spot. It is a simple matter of software. As customers line up at FooBank’s ATM, red flags go up, mortgages are sold—presumably to FooBank’s competitor, BarBank—and trucks full of cash from BarBank arrive. At the end of the day, FooBank has no deposits, and$150 million in assets. It returns these to its stockholders and closes down. Call it an immaculate bank run.

This is a worst-case scenario. It won’t happen. And the fact that, in this worst-case scenario, everyone gets their money back, is what makes it not happen—because the motivation for a bank run is that, in a bank run, not everyone will get their money back. Problem solved.

Our fourth answer is: no. Because we’ve forgotten something else, which is that maturity transformation doesn’t actually work. At least not in a physical, literal sense. You cannot actually teleport money from the future to the present. FooBank can’t—and BarBank can’t.

To understand the problem here, let’s think a little bit about what makes FooBank’s mortgages “worth” $1.1 billion. We tend to use words like “worth” and “value” as if they represented absolute, objective, physical characteristics. A kilogram of iron will always weigh a kilogram. But what makes a package of mortgages “worth”$1.1 billion? Merely the fact that if you put it on eBay, the high bid will be $1.1 billion. This, obviously, depends on the bidders. Mortgages, like everything else, are priced by supply and demand. Even if we hold the demand for mortgages constant, pushing$1.1 billion of them onto the market in one day is likely to depress the market price. And why should the demand be constant? We certainly have no basis for this assumption. Let’s poke a little harder on this one.

First, we need to think a little harder about interest rates. We are used to thinking of interest rates from the customer’s perspective, in which they are a return on investment. From a banker’s perspective, however, an interest rate is the price of future money in present money. For example, a 10% annual interest rate means that I can buy $110 of 2009 money for$100 in 2008 money. This is an exchange rate, just like the exchange rate between dollars and euros.

To know the correct price of a future payment—such as the payments on FooBank’s mortgages—we need to know two variables: the probability of default, and the interest rate. Let’s assume (this assumption is not valid in reality, as we’ll see, but breaking it will only make the problem worse) that the bank run has no effect on the probability of default. Let’s assume, also, that FooBank’s mortgages are perfectly good and have a minimal default probability.

But the interest rate for our 30-year mortgages is set by supply and demand. Clearly, because we are in a maturity-transforming banking system, a considerable quantity of that demand comes from maturity transformers such as FooBank. And ultimately from its depositors. I.e.: maturity transformation in the mortgage market, by magically transmuting demand for zero-term deposits (money right now) into demand for mortgages (money 30 years from now), has vastly lowered 30-year interest rates.

So when FooBank’s depositors pull their money out, mortgages go on the market and mortgage interest rates go up. This increases 30-year interest rates across the market—lowering the price of mortgages. That $1.1 billion isn’t$1.1 billion anymore.

Worse, we have arbitrarily assumed that the run does not extend to BarBank. In fact, we have assumed that BarBank, in some way, pulls in $1.1 billion of new deposits—because that cash in those trucks needs to come from somewhere. But, since our bank run is not the result of any problem restricted to FooBank (whose mortgages are good), it can only be a systemic run. That is, all depositors everywhere realize that the banks simply do not have the present money to repay them—and their so-called “solvency” is a result of long-term interest rates that do not reflect the actual supply and demand for 30-year money. Thus, the entire banking system is certain to implode. And implode instantly. The result: a landscape of shattered banks, people who have lost their deposits, and very, very high (but perfectly market-determined) interest rates. Moreover, housing prices will decline—because they, too, are set by supply and demand, and high interest rates mean expensive mortgages. Another way to think of this is to realize that the Diamond-Dybvig model (see the original paper, here) is not quite right—there is no “good equilibrium.” The so-called equilibrium in which depositors leave their money in the bank is unstable. You can see this by observing that the probability of a bank run is never 0, and the value of a deposited dollar cannot exceed the value of a non-deposited dollar—i.e., the exchange rate between a dollar in the bank and a dollar under the mattress cannot exceed 1:1. But since a bank run can occur, a dollar in the bank must be worth slightly less than a dollar under the mattress—to compensate for the probability of a bank run. E.g., if the probability of the run is 0.001, and the bank returns only 50 cents on the dollar after a run, the value of a dollar in the bank is 0.9995 dollars in cash. This disparity creates withdrawal pressure, which is a feedback loop, and the run happens. If this hurts your head, just think of the maturity-transforming banking system—FooBank, BarBank, MooBank, and all their competitors—as a single bank. This system has, like Bobby, made promises that it cannot physically fulfill, because physically fulfilling them would require actual, physical time travel. There’s simply no way this can be good accounting. Essentially, what we’re looking at is the collapse of a market-manipulation scheme. The banks have been collaboratively bidding up 30-year money by buying it with 0-year money that belongs to someone else. The situation would be no different if they were bidding up, say, copper, or Honus Wagner baseball cards. They have created an artificial pricing environment, based solely on the carelessness of their depositors in rolling over loans. Once the scheme is exposed, the price of Honus Wagner cards crashes, the banks have spent the depositors’ money on goods whose price has considerably declined, and massive insolvency is revealed. It helps to understand the use and misuse of the word liquidity in this environment. The proper and original meaning of liquidity is the existence of a market with instant trading and small bid-ask spreads, such as the stock market. For example, a house is not a liquid asset in this sense, because setting up a housing sale is very difficult and expensive. But liquidity has come to mean something else: the presence of maturity-transformed demand for long-term assets. As we’ve seen, the price of 30-year money in a market where banks can balance 0-year liabilities with 30-year assets is one thing. The price of 30-year money in a market in which all the demand for 30-year mortgages comes from 30-year lenders (for example, a 30-year CD—an instrument which does not even exist at present) is very different. Thus, when a maturity transformation scheme breaks down, the market is said to be illiquid. In fact it is perfectly liquid in the first, original sense of the word. It is just revealing the actual market price of 30-year money as set by 30-year supply and demand—an interest rate so horrifyingly high it makes any 30-year mortgage at 6% more or less worthless. (Actually, the “fire-sale” market for mortgages today is still well above this price—it is set more by speculation that the MT switch will flip back on, I think.) This phenomenon appears to resemble genuine liquidity, because assets are “hard to sell”—but they are hard to sell only because those who now hold them have them on their books well above the market price, and don’t want to sell for less. Professor Bernanke’s recent mention of “hold-to-maturity” price reflects this intentional misunderstanding. Economist Willem Buiter clarifies: The MMLR [lender of last resort] supports market prices when either there is no market price or when there is a large gap between the actual market price of the asset, which is a fire-sale price resulting from a systemic lack of cash in the market, and the fair or fundamental value of the asset—the present discounted value of its future expected cash flows, discounted at the discount rate that would be used by a risk-neutral, non-liquidity-constrained economic agent (e.g. the government). I.e.: the long-term interest rate on Treasuries. Which is still set by maturity-transformed demand (especially, via the central banks of China and the Gulf states, which back their currencies with Treasuries). Professor Buiter’s definition of this rate as “fair” reflects the institutional assumption of the economics profession that MT is a good, clean, and healthy thing. In reality, it is concealing the most important price signal in the world: the present demand for future money. MT adds present demand for present money into this market, utterly and irrevocably jamming the signal. The basic problem with the toxic assets that are clogging up the banking system today is that there is no market mechanism that can reveal Professor Buiter’s “fair value.” Any such mechanism needs to solve for two variables at once: it needs to create a market in toxic mortgages in which the players are banks paying with maturity-transformed money. Otherwise, the default risk of the loans cannot be calculated by comparing their price with the price of Treasuries. The risk-free interest rate in a market in which MT has broken down is much higher than the risk-free Treasury rate. This rate is unknowable. And you can calculate default risk from price only if you know it. Moreover, banks have no incentive to buy these toxic mortgages, turning MT back on in the market—at least not until the price hits its rock-bottom point, at which MT is completely off and 30-year supply is met by 30-year demand. Nor is there any way to see when this point has been hit, because there is no genuine maturity-matched market, and there are plenty of speculators betting on some kind of intervention. And the implied interest rate at this rock bottom is so high that the houses which collateralize the mortgages may be almost worthless. And our fifth—and final—answer is: yes. Because FooBank, BarBank, and MooBank are all FDIC-insured. Actually, this is not even quite right. At least according to this story, FDIC is basically empty. It had only$45 billion last time it reported, and it surely has a lot less now. This to “insure” something like $4 trillion in deposits. You might as well defuse a car bomb by wrapping it in toilet paper. FDIC “works” because it is backed by the Treasury, which of course has the Fed’s “technology, called a printing press”—i.e., Ben’s helicopters—behind it. In other words, when you make a bank deposit, you have acquired not one but two securities. One is a promise from FooBank to pay you on demand. The other is a promise from USG to pay you if FooBank doesn’t. Because the promise is denominated in dollars, USG can print dollars, and USG has no reason to welsh on this promise, the dollar on deposit is truly worth$1. Not $0.9997,$1.

Note that we have just discovered the missing dollars from last week’s post—or some of them, anyway. These dollars are contingent—they only come into existence in special cases, such as a bank run—and they are informal. But as informal, contingent dollars, they exist nonetheless. If the Fed prints money to bail out FDIC, it is a bailout—just like Paulson’s bailout, the Fannie and Freddie bailout, the AIG bailout, etc. There is no law requiring bailouts. But they seem to happen anyway.

For that matter, Treasury obligations are risk-free for exactly the same reason. This is how the US can run a $10 trillion risk-free debt in a world with only 825 billion actual dollars. It is simply assumed that well before Treasury would default, the “technology, called a printing press,” would be used to bail it out. There is no formal guarantee of this. There is simply every incentive to do it, and no incentive not to do it. This whole beast is an incredibly cumbersome and bizarre accounting structure. And it simply cannot be understood without reference to these kinds of informal obligations. In accounting, the word “informal” is essentially equivalent to “criminal.” Simply put: the whole thing stinks. Moreover, we can construct a formalized equivalent that has the same outcome, uses maturity-matched accounting, and is completely unacceptable from a political standpoint. In fact, it’s more than unacceptable. It’s ridiculous. Consider FooBank, with its FDIC guarantee in place. The infinite printing press guarantees FooBank’s liabilities to its depositors. This leaves FooBank free to use the depositors’ money to buy 30-year mortgages, while assuring them that they can redeem at any time. A much simpler approach is for FooBank to simply store the deposits in its vaults, and have FDIC make the mortgage loans—in a quantity equalling FooBank’s deposits. This produces: exactly the same safety for FooBank’s depositors; exactly the same demand for mortgages; and exactly the same risks for FDIC (which is, of course, exposed to the mortgage risk). And it’s also utterly ridiculous. Basically, it means that mortgages are cheap because Uncle Sam is printing money and lending it. When you want a mortgage, you apply to the government. Moreover, the connection to FooBank’s deposits is utterly unnecessary. There is no reason that FDIC has to restrict its mortgage issuance to FooBank’s deposit base, tying it to the irrelevant convenience question of whether depositors prefer their money in a vault or under the mattress. What this thought-experiment tells you is that, if you believe in maturity transformation, you believe it’s good public policy for the government to print money and lend it. Homeownership, after all, is important! President Bush says so, so it must be true. This raises the question of why, if it’s good for Uncle Sam to engage in this practice, it’s not good for everyone. Why not license private entities to engage in the essential public service of counterfeit lending? The counterfeit spender drives up prices and is bad for everyone, but the counterfeit lender creates a liability for every dollar emitted… anyway. We are in the realm of absurdity. We are now in a position to understand the crisis as a whole. What happened is that a shadow banking system appeared, which performed maturity transformation without formal FDIC backing. Participants in this market assumed that large or “too-big-to-fail” institutions, such as Lehman, Bear, AIG, etc., were effectively recipients of an informal Federal guarantee, just like FDIC itself, the traditional banks, Fannie and Freddie, etc. But they had reached the edges of informality and walked off the cliff into delusion. Furthermore, the financial models that players in the shadow-banking market used simply did not incorporate the possibility of a maturity-transformation crisis. They lived in a world of Lombard Street accounting, in which MT was just normal. The fact that MT only works with a lender of last resort who can print money (or, under the “classical gold standard,” compel market participants to accept paper at par with gold, as the Bank of England did during the Napoleonic wars), did not exactly bring itself to their attention. Nor did the fact that they were treating private corporations as perfectly-insured counterparties. Essentially, Wall Street (a) thought it was a free market, and (b) assumed that MT in a free market just works. Both of these assumptions were wrong. The financial system was both free and protected, but the part that was free was not protected, and the part that was protected was not free. And the border between the two was completely informal, and was set in an ad-hoc, after-the-fact way by panicked bureaucrats in Washington. We are now in a position to consider solutions. Our first solution is a free-market solution. In the free-market solution, Washington renounces all bailouts, guarantees, nationalizations, etc. There is an easy way to do this: break the Fed’s printing press. Pass a constitutional amendment limiting the number of dollars extant to the number of actual dollars in the world: M0, 825 billion. That’s about$2750 for every American—although not all of these dollars, of course, are in America.

Result: the mother of all bank runs. All bank deposits are vaporized. The assets backing them become nearly worthless. Have fun paying off that $300,000 mortgage, with your$2750. Even Treasury obligations trade at pennies on the dollar—have fun paying off that $10T national debt, in a world with only 825 billion dollars. The good news: hyperdeflation. If you have a dollar, an actual physical greenback, you can eat for a day. If you have$20, you’re set for the month. A benjamin is unimaginable wealth. Gas? Five cents a gallon. Gold? Worthless. Ammo? Priceless. Basically, we’re looking at Mad Max Beyond Thunderdome, with 1915 prices. I’m sure this would make some people very happy. I am not one of them.

Our second solution is the Paulson plan—with its wonderful acronym, TARP. TARP is effectively a new bank with no liabilities, $700 billion of equity (owned by USG), and a mission to buy mortgages. I am not going to go out on a limb and say the Paulson plan won’t work. But I will be surprised if it works. Basically, its goal is to pump enough money into the market for “troubled” assets—assets in which a bank run has occurred, and MT has broken down—to pull them out of the “troubled” category. My suspicion is that the result will be just one more bank which holds troubled assets, held on the books above their market price. So what? Why should this price be the “official” price? Add this to the fact that the troubled assets are wildly heterogeneous—every mortgage-backed security is different. Just because TARP overpays for one, doesn’t mean everyone should overpay for any. And they won’t. And our third solution is Plan Moldbug—nationalize all banks and other maturity transformers, exchanging their shares for cash at the present market price, acquiring their assets and accepting their liabilities. Consolidate the entire financial system onto USG’s balance sheet. While we’re at it, merge the Fed, Treasury, Social Security and Medicare into one financial entity. Clean up the whole maze of interlocking quasi-corporations. The US Government is one operation. It should have one balance sheet. And yes, the banking system is part of it. Under the pretext of “regulation,” the banking system is already federally managed. Even its executive pay is apparently about to be set by Congress. But most important, the banking system is part of USG because USG has already chosen to accept its liabilities—by guaranteeing them. When A guarantees B’s liabilities, B needs to be on A’s balance sheet. This is accounting 101, folks. This unity is even recognized in the conventional nomenclature for money supplies; M0 is the supply of zero-term Fed liabilities (i.e., Federal Reserve Notes, i.e., dollars); M1 is the supply of zero-term bank liabilities. The exchange rate between an M0 dollar and an M1 dollar is 1:1 and guaranteed to stay that way. So in what way are they different things? As we discussed last week, FRNs are more like equity than debt—they are not promises to pay anything. And if an FRN is a share, a Treasury bill is a restricted share. Thus, under Plan Moldbug, USG has a single balance sheet with no debt. It can, of course, continue its present practice of diluting its equity (selling Treasuries, printing FRNs, etc.) to pay off its operating deficits. But it would be better advised to issue a large pool of shares to itself, to assist it in phasing out this pernicious habit. Also, Plan Moldbug has a critical fairness advantage: it is portfolio-neutral. The day after this action, everyone’s portfolio statement will show exactly the same number. If (God forbid) you hold bank shares, those shares will be converted to cash, just as if the bank had been acquired by another bank. If you hold mortgage-backed securities—but who holds mortgage-backed securities? Not anyone I know. And if you pay a mortgage, in all probability you are now paying it to the government. Do you care? This is an important point, because Plan Moldbug involves a gigantic increase in M0. M0 is the sum of Fed liabilities. By consolidating the balance sheet, we are moving bank liabilities—the rest of the M’s, and beyond—into this category. In any naive analysis, this seems “inflationary.” But naive analysis neglects the actual relationship between money supply and prices. Prices increase when people have more money, ceteris paribus, to spend on the same goods. If the change is portfolio-neutral, by definition it cannot affect prices. In reality, all we are doing is recognizing the actual supply of dollars, which is much greater than$2750 per American.

Naturally, the point of any such nationalization is not to have to do it again. Applying to the government for loans is a little Soviet for my taste. Our present system, in which Fannie and Freddie more or less are the government, is already a little Soviet.

The end goal is to phase out this lending-counterfeiter business, and construct a new financial system—the motorcycle—in which lending is really, truly private, and financial intermediaries match their maturities. If Bobby needs money for three weeks, he asks you for a three-week loan. He does not ask you for a one-week loan and then get a surreptitious, covert, informal three-week loan from the Fed’s “technology, called a printing press.”

In any such financial system, we would see the true yield curve, the graph of interest rates at every maturity, uncontaminated by maturity mismatching. My suspicion is that at least at first, long-term rates would be quite high. Which means lower house prices. In the spirit of portfolio neutrality, USG might want to print some more money and kick it back to homeowners, such as, of course, myself…

But at this point we are just fantasizing, because none of this is going to happen. If the American public balks at “King Henry” (I believe one poll showed 7% support for the feeble TARP)—they’ll have none of Emperor Moldbug. If printing a mere \$700B turns their stomach (the “cost to the taxpayer”—of course, no one will raise taxes to “pay” for this use of the “technology, called a printing press”), moving a few T from M1, M2 and M3 to M0 is a nonstarter.

What we’re seeing here is a failure of democracy. Americans have a certain picture of their financial system. They believe that a dollar is, in some way, an asset like gold that cannot be printed at will. They believe that banks are free-market corporations, not branded branches of the State. They believe in regulation, but they don’t believe in socialism. And so on.

Any politically realistic policy has to be framed in these terms, which are not realistic. The result: stalemate, ineffectuality, drift, and disaster. In other words: EP1C FA1L. I don’t know what’s going to happen, but I know it’s not going to be good.

1. There is a good discussion of this post chez Arnold Kling. —MM