Banks, by their nature, are susceptible to bank runs. Depositors panic and demand their money back, the bank doesn’t have enough cash on hand to meet all the demands, this generates even greater panic and even more demands, and pretty soon the bank is selling off assets at fire sale prices in a desperate attempt to placate the depositors. Back before Federal deposit insurance, this used to happen from time to time. According to Yale’s Professor Gary Gorton, author of Slapped By the Invisible Hand: The Panic of 2007, it happened again recently. The great crisis of the past few years was just another bank run, pure and simple.
The only difference is that this time the banks were non-traditional financial firms like Bear Stearns (as opposed to, say, the Bailey Building and Loan Association) and the depositors were large financial institutions like Fidelity (as opposed to the good citizens of Bedford Falls). Here (with a hat tip to Gorton) is how that non-traditional banking system worked: Suppose Fidelity is sitting on, oh, say, half a billion dollars in cash that it would like to deposit in an interest-bearing account someplace. So they open an account at Bear Stearns. Unfortunately half-billion dollar accounts at Bear Stearns (unlike your checking account at the Bank of America) are uninsured, which makes Fidelity a little nervous about turning over so much money in exchange for nothing more than a deposit slip. So Bear Stearns, along with the deposit slip, hands over a half billion dollars worth of collateral, in the form of bonds. Fidelity takes physical possession of these bonds, so that it’s covered if Bear Stearns goes bankrupt or absconds with the money. (Bankruptcy law allows Fidelity to keep the bonds if it’s holding them when Bear Stearns goes under.) When it’s ready to make a withdrawal, Fidelity returns the bonds and gets its cash back. This exchange of collateral is what’s called the repo market.
Now suppose that Fidelity starts to worry about the value of the collateral it’s holding—say because it’s not sure how many of those bonds are backed by hinky subprime mortgages. Sure, the bonds are worth half a billion now, but they might be worth only 400 million by tomorrow morning. They say to Bear Stearns: “Umm. We’re a little worried here. If you want to keep our deposit, we’ll be needing more collateral please”. But if several depositors say this at once, there’s not enough collateral to go around—so depositors start making withdrawals (or, equivalently, depositing less money per unit of collateral). And if you’ve ever seen It’s a Wonderful Life, you know what happens next.
Some of Professor Gorton’s conclusions:
- Bank runs are not driven by irrationality; they are perfectly rational responses to new information about banks’ solvency. Of course “rational” is not at all the same thing as “desirable”.
- It is wrong to say that the crisis was caused solely by the collapse of the subprime mortgage market. That market, by itself, is too small to cause a crisis of this magnitude. Besides, the timing is wrong. Subprime mortgages started going under in January, 2007, but the panic didn’t get underway until the following August. Gorton likens the subprime mortgage market to a small E. coli infection somewhere in the nation’s beef supply. The infection itself is only a small problem, but the ensuing panic can cause the entire beef industry to collapse.
- Where there are banks, there will be bank runs. Therefore this problem is structural; it’s not a one-time thing. On the other hand, imperfect as they are, banks are pretty much indispensable.
To this I would add the observation that fire sales are not unambiguously bad things; they are bad for the sellers but very good for the buyers. So toting up the losses to Bear Stearns and its ilk is a very poor way to measure the cost of this crisis. The real cost has been that with the collapse of the repo market, traditional banks had no place to sell off their loans, and therefore became quite reluctant to initiate new loans, making it difficult for businesses to maintain their operations. Without denying the reality of that cost, it’s worth remembering that it’s a relatively short-run phenomenon.
There is much more here and in Gorton’s book.
I’m having trouble coming up with a model where this would play out as described.
So Fidelity deposits $500 million cash with Bear Sterns and gets the bonds as collateral. At that point, Bear Sterns doesn’t keep the cash “on hand” (hence the problem with bank runs), but it must do *something* with it. Whatever it invests the money in, call that investment X. That’s what is supposed to pay for the interest that Fidelity expected to earn in its interest-bearing account. (And that’s not the same as the bonds that it gives to Fidelity as collateral.)
If Fidelity gets nervous about the value of its bonds as collateral, then Bear Sterns could say, “Well, if you don’t trust the value of the bonds, fine, give us back the bonds, and as alternative collateral we’ll give you our ownership rights in investment X.”
If Fidelity trusts the value of investment X, then that should make it happy. On the other hand, if Fidelity doesn’t trust the value of investment X, then they shouldn’t have invested with Bear Sterns in the first place, because that investment is supposed to pay for the interest that Fidelity earns on its deposit.
So I can’t think of any set of assumptions where Fidelity *would* happily deposit money with Bear Sterns, but would *not* be happy to take Bear Sterns’ investments on their behalf, as collateral (and hence the bank run on Bear Sterns). Can you clarify how this could happen?
It might be as simple as a rule in finance that says something like, “You can’t give someone your ‘rights to an investment’ as collateral.” (Although I thought that’s exactly what the bonds were, that were backed by subprime mortgages!)
Also the model outcomes depend on what the rules of the game are — I know something about game theory but nothing about finance :) — so could you explain:
– About Fidelity’s interest-bearing account with Bear Sterns: it could earn interest, or it could be wiped out completely, but could it do something in between as well? Is it like an investment in a stock, where the $500 million could go down to $300 million, with no insurance to cover the loss, but without Bear Sterns going bankrupt?
– If so, then if Fidelity’s $500 million account with Bear Sterns goes down in value to $300 million, can Fidelity say, “Keep the account, we’d rather have the bonds you gave us?” Or does Bear Sterns say at that point, “No, that’s just collateral in case we go bankrupt or disappear. As long as we maintain your account in good standing, we still own the rights to those bonds, even if your account loses value.”
– Similarly, what about the bonds that Bear Sterns gives to FIdelity as collateral, can they go *up* abruptly in value, just as they can go down? (I would assume so, since if everyone knows they could go down in value, that would lower their inherent value unless it were also possible for them to go up in value.) And again, same question: If the bonds do spike in value, can Fidelity say, “Hey, keep our account, we’ll take the bonds” or would Bear Sterns still say “No the bonds belong to us as long as we don’t go bankrupt or steal your money.”
I had understood that part of the problem was a false assumption that risk had been avoided, when in fact it had been amplified. The incentives had gone very wrong, so outcomes were wrong. The default rate on mortgages was well known, so a bank could parcel up a few high risk ones in credit default swaps, and keep the rest as very low risk. Trouble is, they interpreted very low as no risk, so did not insure them. AIG did provide insurance for some of these, but it was quite expensive and made the whole CDS uneconomic, so most banks did not bother. The default rate was well known, right? There was no way that these were going to default. This might have been OK if the mortgage market had carried on as before, but suddenly, banks could make loads of money, and off-load the risk. They had no incentive to ensure the mortgage was paid back, so they did not bother. They made just as much on a bad loan as a good one. The incentives here are definately very wrong. The buyers of the mortgage bundles had believed that the risk had been eliminated, and didn’t really understand what they were buying. The credit default market had changed the mortgage market, and the original assumptions no longer applied. Suddenly this huge stock of no-risk stuff became risky, and the value dropped like a stone. This caused the run on the banks. AIG had insured too cheaply, and was left burned. This applied to other things as well as mortgages.
Any public owned bank that did not want to take part in this bubble achieved low short-term profits, and so was bought out by one of the gamblers.
As mentioned above, it seems to be in the nature of banks to suffer runs. This is exacerbated by the one-sided nature of the bets taken by the bankers. They will take all the profits, but not suffer all the losses. The incentives are again wrong, so the outcome will be wrong.
I do not know how we can get the incentives right. Short of this, it seem we must regulate to prevent excesses. This will inevitably hinder the operation of banks, but given recent events this is probably a good thing.
The locking up of the credit market will probably be fairly short term, but has very significant human cost. Would we have been better off with more regulation, slower growth and no crunch? I think we would. Even if the total economic output would be the same, we tend to dislike sudden changes. The extra misery caused is not balanced by the extra happiness during the good times.
I may have got everything above completely wrong.
Harold:
I do not know how we can get the incentives right. Short of this, it seem we must regulate to prevent excesses. This will inevitably hinder the operation of banks, but given recent events this is probably a good thing.
Gorton makes the point that pretty much all you can accomplish by regulation is to affect the profitability of the regulated sector, which in turn will affect the size of the regulated sector—that is, it will determine how much activity is pushed off to non-traditional markets that haven’t yet been regulated.
Is there a non-regulated sector for these kind of financial instruments? If you stop the banks doing it, I don’t think anyone else can. Reducing the size and profitability of the regulated sector is pretty much the desired objective. It tends to grow too big through bubbles, then come crashing down. As you said, where there are banks, there will be runs. Regulation can perhaps limit the size and frequency of the runs.
In a proper market, the size of the booms might be limited by the possibility of total loss, but where we are now that cannot happen. I also think that human nature tends to be over optimistic in these things, and will on average underestimate the risks. Perhaps and alternative to regulation is no regulation, but no bail-outs. You have got to make the individuals who take the risks pay the whole costs of the losses, and I don’t quite see how that can be done.
Bank runs are an essential part of the financial sector. Banks should be scared to death that their depositors and lenders will call their cash back at any time and therefore invest accordingly or don’t pay interest and charge a warehousing storage fee.
The problem isn’t the market, the problem is having a lender of last resort and explicit (FDIC) and implicit (bailouts for too big to fail) “safety net” via a central bank that encourages a “no lose” mentality and a fractional reserve banking system.
It is perfectly rational for depositors to not care what is done with their investments under these rules of engagement. However, we don’t want them to walk down that rational path. Nor do we want financial institutions to be that rationally leveraged.
In a free market system, the invisible hand doesn’t “slap us” as Gorton proclaims. But it does give us the finger if we try to circumvent the rules of correctly risk adjusted resource allocation through prices by setting up a monolithic government bureacracy to “manage it all better”. Of course when Government meddles in the market process and things go wrong, it’s the market that gets the blame, not the meddling.
Read Epicurean Dealmaker much lately? *grin*
Anyway, I think I basically agree with Gorton but ANY explanation for the meltdown that ignores the massive and pervasive levels of fraud and outright lying that went on is just missing the point. From retail-level fraud (liar loans) all the way up to the massive near-fraud that made Lehman a hollow shell (as the investigators are finally discovering) the entire system has been riddled with people behaving in the worst possible ways.
Being paranoid about one’s money in such an environment seems pretty rational to me. My guess is that Bear’s customers were shaving coins and doing funny off-book accounting crap so they naturally assumed that Bear was doing it, too and therefore Bear’s paper was crap.
OK, I have no problem with the view that subprime mortgages were a minor factor in a financial crisis that would have happened anyway. But what about the view that
The financial crisis is just a sideshow – the real reason for the economic downturn is the rise and demise of the housing bubble.
http://www.guardian.co.uk/commentisfree/cifamerica/2010/mar/08/financial-crisis-subprimecrisis
Bank runs are only a necessary feature of the inherently fraudulent and insolvent fractional reserve banking system. A system of 100% reserve banking would never experience bank runs.
Snorri Godhi: Where did the money come from to lend to the house buyers? I am sure the housing bubble is important, but perhaps it is as much a symptom as a cause.
Noah Yetter. I presume that the cost of a 100% reserve banking system would be much less money to lend to industry, and much slower or zero growth. One may think this a good or a bad thing. My aim would be to get as much benefit as possible from the banks, whilst reducing the downside.
Bubbles are purely a monetary phenomenon and nothing else. The expanded money supply has to go somewhere. In fact, the Fed wanted it to go to home lending and encouraged new home loan products because they got more bang for each buck artificially added to the system.
Harold – you cannot lend money that hasn’t been saved. Expanding the money supply just deteriorates the value of each dollar and increases prices. Yes there may be less lending but you would need a whole lot less money because the cost of each product and project would be significantly reduced. Expanding the money supply doesn’t increase production or the capital base, you just have more $ notes chasing the same amount of resources.
In a 100% reserve system, banks can only lend money that depositors have set aside for lending. Depositors setting aside savings means that they are reducing consumption and demand of current resources. This means that prices reduce and resources are freed up for other projects.
May I suggest the following book for an enlightening read:
http://www.amazon.com/Money-Bank-Credit-Economic-Cycles/dp/1933550392/ref=sr_1_1?ie=UTF8&s=books&qid=1268843202&sr=8-1
Dave: This is all very complicated. I should study more economics. Under the current system, banks do lend money that hasn’t been saved. I was thinking that this money could “do good” by being put to work productively. This allows faster economic growth, as there is more production. I was also under the impression that Governments just adding extra money was a bad thing – this will unevitably lead to inflation by the expanded money supply you describe. I am now confused. Why should banks adding money be good and Governments adding money be bad? If there is no distinction, then why do we not have 100% reserve systems ypou describe?
says it most entertainingly….
http://freedom-school.com/money/how-an-economy-grows.pdf
I haven’t the slightest clue what everyone means by the term “bubble.” My haphazard guess is that you mean a rapid increase and then decrease in the price of a good. But this can’t be a monetary phenomenon because:
a) This doesn’t explain why the price of one good in particular changes as opposed to the prices of *all* goods changing.
b) “Bubbles” have existed long before fiat currency and central banks.
I think bubble means that the price of something rises dramatically entirely on the expectations of future rises in price. I would think expanding the money supply would cause inflation rather than bubbles, although it could make them more likely.
Dave: The Island described in the link provides a good introduction, but does not entirely answer my questions. Todays banks lend money they don’t have. In the island case, this would be like giving food vouchers instead of fish. Lets say there was a second island that used this system. The bank in the first island has loaned all its fish to the water project. Mr Inventor comes along with a scheme for a fish farm. Sorry, say Mr. Banker, no fish left, you must wait until the water project is finished. However, over on Island 2, the banker says “OK, I don’t have enough fish to lend you, but take these vouchers to the farm, they will provide food for you whilst you build your fish farm, then I will re-pay them in fish later”. Mr Inventor builds his fish farm sooner than on Island 1, and their economy grows faster. I assume Island 1 is a 100% reserve system and Island 2 is more like our current system. Bank 2 is susceptible to runs, as he doesn’t have enough fish to pay everyone back if they all ask at once.
Harold: glad you liked it. As I say, it’s just an introduction. In regards to your question, the certificates on the island are used to buy fish so people can eat while they work rather than have to spend time fishing to eat. ie the fact that there are savings of fish means that people can use their time on other productive activites other than fishing.
If the island 2 bank lends the fish certificates without having savings, then when the people working on the water project go home at the end of the day to redeem their fish for dinner, they will realise that there are none there. So the next day they will abandond the project. At first it will look like there is a boom but when it is evident that there isn’t enough capital to continue the project, then you get the bust. Which is what we see in business cycles.
This example is on the most basic subsitence but makes the point. A society must have enough capital saved before it can get its citizens to abandon survival projects like fishing, building shelter etc for large infrastructure projects.
If it was as simple as just lending money that doesn’t exist, we could save all world poverty using this method and build plush hotels and airports all over Africa and everyone could live like kings tomorrow.
JLA – please provide one example of a speculative bubble that hasn’t occured due to an artifical expansion of the money supply (ie without commensurate increase in savings). Hint: it’s not tulip mania or the South Sea bubble.
Dave, it’s hard for me to do that without you giving a working definition of “speculative bubble” but here are a few possibilities:
http://en.wikipedia.org/wiki/Railway_Mania
http://en.wikipedia.org/wiki/Florida_land_boom_of_the_1920s
If you’re going to argue that monetary expansion causes “bubbles” you need to tell a story why the monetary expansion caused a change in relative prices rather than just raising all prices.
JLA – both of those examples were in the back of huge monetary expansion by a central bank. First the Bank of England, and secondly the newly created Federal Reserve.
With a stable money supply, an increase in prices of one asset class demands a decrease in prices of another. The fact that we didn’t observe that in the last decade is evidence of an increasing money supply.
Speculative bubbles can obviously get out of control with “animal spirits” but without an infinite money supply to fund the speculation, you just can’t get bubbles because the cost of borrowing would increase and other investment opportunties would be far more attractive as their prices would fall/yields would increase both on a relative and absolute basis.
The housing bubble we just went through was a combination of the increasing money supply and legislative incentives for home ownership (tax deductions, no capital gains tax, low interest on housing, Fannie/Freddie bidding up loan books and providing liquidity).
With an increased money supply, it has to go somewhere. If the legislators incentivise certain activities by not allowing the market to generate diminishin marginal returns into an asset class, then that’s where the bubble will form. Other industries latch on to the growth into that sector (eg mortgage broking, building, furnishing etc) that when it becomes evident that there was not sufficient real capital to support the growth and anciliary business, then the bust occurs and everything else that was tied into it follows.
Actually, the U.S. experienced deflation from 1920 to 1921 – right when the Florida land boom was getting started. The land boom turned into a bust in 1925 – right in the middle of a period of loose monetary policy.
The theory that “bubbles” are purely a monetary phenomenon is inconsistent with this observation.
Thanks for making my point. Fed cuts raised drastically in 1921 causing a boom that lead to a bust. This all in the back drop of a lender of last resort via a central bank.
Next.
Can you explain to me how the Fed cutting rates in the latter half of 1921 was responsible for a real estate boom that started in 1920?
Prices were low in 1920….Just because prices were rising doesn’t mean that there was a bubble. there was just demand. Once it got speculative (with an increasing money supply), it shifts into a deflationary bubble with the inevitable bust.
I meant inflationary bubble.
Also with rising interest rates, the Fed is still increasing the money supply with no commensurate increase in savings.
Anyway – if you could supply a source for the boom starting in 1920 exactly would be fantastic (note that this doesn’t mean a price rise).
Kinda feels like your grasping at straws (I note how quickly you moved away from the Railway Mania).
Can anyone knowledgeably address Bennett Haselton’s (first commenter) questions? Pretty much the same questions occurred to me after I read Gorton’s papers. They seem fairly obvious, but he doesn’t address them, and from my Google searches, neither does anyone else.
Thanks!
Ken