*The Bankers’ New Clothes*

That is the new book by Anat Admati and Martin Hellwig and the subtitle is What’s Wrong with Banking and What to Do about it.  Here is their bottom line:

We have argued that if banks have much more equity, the financial system will be safer, healthier, and less distorted.  From society’s perspective, the benefits are large and the costs are hard to find; there are virtually no trade-offs.

I agree with the proposal, though not with the claim that this is virtually costless, as is laid out in their chapter seven (oddly they focus on the question of whether debt and equity “require” comparable rates of return, rather than the general notion of opportunity cost).  In any case this is a major net work on banking and its regulation.  Here is the book’s home page.  Here is Admati on YouTube.


What about George Selgin research that some of the most stable banking systems had not a lot of reserves and no regulation. Is that discussed in the book?

I agree that banks are safer if the have more reserves but how are we gone pick how much that is? Seams a bad idea to let the political process figure that out (Im just assuming that that is what the authors want).

Might you be confusing reserve requirements with capital requirements? Maybe I am wrong. Which one is relevant to this post?

Im not native english speaker, I did not remember what 'equity' means and mistook it as meaning something like 'reserve requirments' witch is what you often hear.

Not sure I understand why more equity would help however.

One of Selgin/White's anecdotes is about Adam Smith, who lost a bunch of money when the Scottish bank he was a shareholder in went under. Shareholders were actually liable for quite a bit. They portray Scottish banking overall being quite safe.

I suspect Gary Gorton of Yale ("Misunderstanding Financial Crises") would take issue with this prescription.

Gorton would argue that the central problem of banking is that it cannot issue riskless debt, and that at some point in the cycle, ordinarily information insensitive debt (like a checking account) becomes information sensitive. At this point, depositors will not be able to determine which banks are solvent and which are not, and may run on them indiscriminately. Banks will not be able to meet the calls for redemption because their debt will be long term (eg, mortgages) even as their liabilities are short term.

Therefore banks will face a crisis of liquidity (which requires cash), and not solvency (which requires capital)--as Admati and Hellwig propose. Gorton identifies a wholesale--not retail--run on the shadow banking system as the root cause of the 2008 financial meltdown. Hedge funds, insurance companies, and mutual funds also need checking accounts, and these were effectively with their prime brokers. When Lehman started to implode, these wholesale customers became unsure as to which brokers might remain solvent, and began to withdraw their funds, leading to a funding crisis in a number of markets. The whole LIBOR rate rigging scandal was also related to the collapse of ordinary interbank lending as LIBOR, a normally information insensitve rate (ie, the participating banks are ordinarily assumed to be liquid and solvent) became information sensitive (ie, providing a market view of the creditworthiness of one or all of the participants).

Gorton notes that there was no retail run on the banks, precisely because deposit insurance reassured depositers than their money was safe. For the wholesale, now shadow banking system, there was no similar mechanism, so tremors in the banking sector were able to transform into runs on the shadow banking system.

Gorton's proposed solution, then, is to provide some form of deposit insurance at the wholesale level. The question is how to do it. He proposes creating chartered banks to deal exclusively as providers of liquidity to the wholesale market. I personally feel this suggestion has sufficient merit to warrant a full discussion in the economics and regulatory sphere.

On the other hand, banks have had capital requirements for years--Basel I, II and now III--and it didn't seem to help much. So I'm with Gorton, and I believe Admati and Hellwig are misunderstanding the financial crisis.

Awesome post S. Kopits, I learned a lot. BTW Anat Admati is the co-author of a 1988 paper, "A Theory Of Interday Patterns..." refereed by Myron Scholes, explaining why interday trading often shows a "U-shape". My one sentence summary: liquidity traders prefer to trade in herds, but following a so-called non-discretionary trader (somebody who has to trade for a reason), and amplifying such traders, with such non-disc traders tending to trade at the beginning or end of a trading session (due to time constraints such as that they have to buy or sell a block of shares that particular day).

Good post, but I have a question.

"When Lehman started to implode..."

Wasn't that a solvency issue? I don't see solvency and liquidity issues as mutually exclusive explanations of the crisis. Their relative importance, I don't know for sure, but I think you are probably right in this case that once things got going it was liquidity. In any case, if Basel didn't work, does that mean that capital requirements don't work or that they were not strong enough?

BTW I once asked Tom Hoenig with the FDIC at a talk what he thought about expanding deposit insurance to wholesale. He was dead set against it.

Solvency. The problem then becomes how valuable is everyone else when they own part of the insolvent and have leveraged that asset, as well as has a large stake in the same asset class that drive the first firm into insolvency.

Equity represents the value of the enterprise in the eyes of the market. In almost all endeavors the size of the project or investment you can take on depends on how deep your pockets are because everywhere else you are required to put some skin in the game. When not if things go sour someone had to absorb the loss; the riskier the endeavor the more backing is required.

There are regulations to this effect in the banking sector but wall street seems to be expert at circumventing them by off balance sheet setups. Shareholders love them because they are profitable as long as everything goes as expected.

Remember what saved Wall Street. Us taxpayers bought worthless paper at above market price. The Fed purposely worked to prevent the true value of these firms to be known. And made almost unlimited amounts of cash available to show up the bank balance sheets. A Potemkin village as a basis for the economy. The inhabitants of this village, as their wont, set about to probe the seemingly infinite depths of the pockets of the Treasury and Fed.

I fail to see how some tweaking of some regulator requirements can do anything.

Yes, Lehman had a solvency issue. However, that wasn't the problem. Had Lehman faced an orderly liquidation and a sale of the operating assets to Barclays--which is what at least some senior management at Lehman wanted--then there would not have been such a crisis in the various financial markets.

But as I recollect it, the traders at Lehman were simply told to pack their things and go home--with hundreds of billions in open positions! So you had transactions in which Lehman was the intermediary, but not the formal counterparty, and these seized up. The risk to the system was not Lehman's bankruptcy per se, but its impact on open positions and the confidence of investors in the ability of the prime brokers to return their money. If you were a hedge fund and had an account with Lehman, then you wanted to get your money out of there for fear that it might get stuck. And they weren't sure who else Lehman might take down with them.

Gorton argues that this sort of crisis was a recurring problem prior to the founding of the Fed. What's required is some confidence by wholesale investors that their cash deposits or their transactions in process will be settled in a reasonable period at par.

As for deposit insurance, Gorton is not referring to the FDIC but rather to a select number of specialized, chartered institutions acting as banks to the wholesale sector. So it's an entirely different institution from the FDIC (and I would probably not encourage mixing these things either) with substantially different structure and rules.

So, instead of a massive Dodd-Frank focused on regulation and capital, Gorton, if I understand correctly, perfers a more surgical and focused approach to dealing with this single problem of wholesale bank runs. That makes sense to me.

I am way past my pay grade here as it is, but I think Gorton, who is--let's be fair--perhaps the country's leading expert on bank runs, should be given a full hearing.

I would recommend that you read his "Slapped" article (link below), or his very good, relatively short, and completely accessible book "Misunderstanding Financial Crises". It's all in there. I would add that the Slapped article was transformative for me, and I consider it the most important economic insight I have gained since I learned about principal-agent theory a couple of decades ago. It speaks to market inversions in which normal free market (libertarian, if you will) principles do not hold because the assumptions underlying those markets have failed. (This is the essence of what happens when an information insensitive security becomes information sensitive.) Thus, the appropriate policy during such crisis is the suspension of convertibility and mark to market accounting. In addition, I would argue that LIBOR should also be administratively set during this period in consultation with the Fed and other central banks. Thus, the suspension of mark-to-market should apply not only to balance sheet categories (notably to liabilities), but also to income statement metrics related to solvency and liquidity.


I agree that nobody is better than Gary Gorton at explaining liquidity issues but I wouldn't trust him on solvency issues. He was neck deep in that mess at AIG.

Sure I understand the runs that these events triggered. I reject any solution that starts with the assumption that a bigger backstop is needed.

These are all adults that were involved in this market. For example, a CFO of a large corporation would have willingly done what is common practice; outsourced payroll where the data and cash is forwarded a week or so before payday. The payroll firm buys some short term paper on the money market, processed the payroll stuff, then cashes in the paper friday morning so the checks and transfers can happen.

Then the money market fund broke the buck.

This money market paper was in fact buying some derivative of a worthless deadbeat pool of mortgages that ultimately was worth nothing. If anyone in this chain had actually read the prospectus and figured out what they were buying they may have not. As fiduciaries they should have.

The only thing that capitalism has going for it is losses. Bad ideas are punished early and harshly. That CFO and CEO who put the operation of the corporation at risk by not doing due diligence on basic functions of the business should be selling apples on the street corner, pour les autres.

The shadow banking sector and deriviative levels are now close to what they were pre 2008. Nobody learned anything, and inevitably some other series of events will collude to create another crisis.

AIG was a quintessential liquidity problem, not a solvency problem. At no point, even after the $200 billion write-downs AIG took, did the aggregate liabilities exceed the aggregate assets. The underlying insurance companies were all solvent. Turning the insurance companies into liquid assets took several years, which was the whole point of the extradordinary process the government undertook.

If the difference between the shadow banking sector and the regular banking sector was merely the distinction between wholesale and retail than your proposal would make sense. However, proving shadow bank run insurance without also imposing the regulatory constraints designed, among other things, to prevent bank runs is foolhardy.

By all means insure large demand deposit accounts, but do it within the FDIC system, not money markets, commercial paper, auction rate securities and exoctica. If that means low returns on cash - well that's the price of safety.

Put simply, many observers think that the financial crisis resulted from a liquidity shortage, meaning that financial firms ran out of cash. That's a problem on the left side of the balance sheet. It can be addressed with reserve requirements, which require banks to hold minimum amounts of cash. When, however, a firm is long-term insolvent, it has too much debt relative to equity. That's a problem on the right side of the balance sheet. It can be addressed, as this book seems to propose, by setting minimum capital (read: equity) requirements, which is another way of saying, maximum leverage ratios. So the essential question is: Was the financial crisis a liquidity crisis or a long-term solvency crisis?

Maybe it's too early to tell, like Mao's French Revolution? Every bank run by definition is a liquidity crisis at first (short term), but, if asset prices don't recover (long term), as you say it becomes a solvency crisis. BTW you see the statistic that during the Great Depression (GD) 20% of US banks failed, but I've also seen the statistic that only ~0.5% of depositors actually lost money (mostly in the Deep South). So bank failures were much ado about nothing. In fact, I've heard about people who lost money in US GD banks, and were given a choice of repossessing real estate owned by the failed banks, which they did and rumor goes these persons got rich decades later when the real estate recovered.

Richard -

Gorton would argue that you are wrong. If a bank set aside every dollar in reserve against a bank run, then it would not be a bank. It would be a private equity fund, because all it's funding would be equity. This would require a higher rate of interest to borrowers--say 12-16%--and offer no convertibility. Thus, you couldn't use it as a checking account, just as the investors in a private equity fund commit their money typically for 5-7 years without the possibility of early withdrawal.

But there's no reason to charge people 12% interest in normal times, and everyone needs a checking account where you have ready access to all your cash. That's why banking was invented: full liquidity can be offered because not everyone wants their cash at the same time. The surplus can be lent. That's what banking's all about.

And that works well most of the time. But it has an Achilles heel, notably that banks cannot issue riskless securities and these securities cannot be called on short notice. So a bank cannot guarantee that a borrower will be able to repay and cannot call loans like mortgages on short notice. Normally, we don't care because our demand deposit is "information insensitive", ie, you can depost your money in any old bank without worry that you'll be able to withdraw it later. You don't ever check the bank's financial statements, for example, before doing so. And that's a great benefit to society!

But if that deposit becomes information sensitive, if you're not sure that you can get your money out, then there will be a run on the bank, and the bank will be unable to find the funds--owing to the very nature of its business--to cover a full run.

That's Gorton's point.

@ Steven Kopits

Gorton would say I'm wrong? I don't see anything that you wrote that contradicts anything that I wrote. I was just laying out some definitions.

Isn't the equity premium higher because there is more risk?

Exactly. We want high percentage of equity and high reserves. But the return on equity is around 15%. And the authors claim that there are no trade-offs. Fascinating.

There is a trade off. In the event of things going sour, which have well established characteristics; sudden, everything, today; the firm will not be in a situation where their losses can cause the systemic failures that we saw. Essentially the firm would not have been able to expose themselves to larger losses than could be absorbed. That does mean the equity holders are wiped out.

There is no real tradeoff (unless you want to invoke some kind of agency friction). Even though equity is "more expensive" than debt, increasing the amount of equity on the balance sheet lowers the cost of debt and leaves the total cost of funding the bank unchanged. It's called the Modigliani-Miller theorem.

Except where tax systems allow the cost of servicing debt to come from untaxed income, but dividends are paid from taxed income. It's called the "tax shield" and it means the Modigliani-Miller theorem does not fully hold. When you adjust for the tax shield on debt, corporate debt financing is cheaper than equity.

Right, but equity holders also don't like their companies to be over-leveraged, which somewhat offsets the "tax shield" effect.

It's the debt holders who don't like over-leverage

Talk about mixing the obvious with the absurd.

For banking, equity is essentially equivalent to capital, and there has been a drive for higher capital requirements for some time. Naturally this reduces systemic risk in more ways than one. First, it reduces leverage. Second, it increases shareholder risk and, ostensibly, increases shareholder oversight and market discipline.

But to say the tradeoff is costless is absurd. Higher capital requirements reduce return on equity making investments in banks less attractive vis a vis alternatives. Investors in banks have to be compensated relative to the risk and return profile of all other choices.

Higher capital requirements can also have perverse consequences. Banks might be too conservative, and their liquidity will be threatened by low earnings. Banks might increase their share of MBS in order to get more bang for their regulatory buck.

There's no question that we will be making the banking system safer in the short run. The question is whether tighter banking will boost economic growth as much as we would like. Part of the reason for the slow recovery is from people not taking enough risk at a time when risk appetite should be high.

In a systemic event, increased capital does not mean that risk is eliminated, or even significantly reduced. The only thing capital does is reduce the amount FDIC might have to pay out in excess of capital (assuming the bank had losses) in the event of a bank liquidation. That's an accounting matter without implications for the broader banking sector. It's like a manufacturing company going bust. You offset the liabilities with the assets and close the doors. If there's a deficit to the depositors, the FDIC makes up the difference.

But that's not a systemic crisis. In a systemtic crisis, capital will be of little help, argues Gorton. The remedy for that is not increased capital, but rather the administrative imposition of what would be "information insensitivity". So you prevent people from withdrawing funds, you cease providing accounting information, and you have all the banks or the Fed back each other. You do not trade and you do not allow assets or liabilties to be valued during periods of panic. Instead, the Fed provides essentially unlimited liquidity until eveything calms down. You reassure demand depositors--whether individuals or corporations--that they will get their demand deposits back at par value. And you do that leaning on the central bank--which is what happened, after the Lehman debacle.

Now, for the Fed to act in time and with confidence, it needs to know what sorts of liabilities are out there and who holds them. In designating special, chartered banks to deal with wholesale demand deposits, the Fed would be able to exert control over only a few banks to have control over the wholesale liquidity situation overall. Hence Gorton's proposal, if I understand it correctly.

Gorton's considered position is that only governments can create safe assets. Government debt is used as collateral in the shadow banking system and it is this that renders it "safe", just as FDIC insurance makes retail deposits safe. Gorton is taking as a given the continuing existence of a plentiful supply of safe government debt. He is adding to that the shadow bank equivalent of a lender of last resort, that's all. I would personally be extremely cautious about Gorton's ideas, and I certainly wouldn't regard them as a private-sector solution to the bank run problem. Because of the shadow banking system's dependence on safe assets, the taxpayer is still on the hook really.


Private equity investors only "require" 12-16% at the moment because their equity is usually highly leveraged. Think about a world in which there was no leverage on banks. The equity rate should then be equal to the rate of return on the underlying assets - usually around 1 or 2% (real) depending on the risk. My guess is that you work in the finance industry - it is hard sometimes to see the water when you are a fish.

The real issue is that a "bank system" with long term funding using short term deposits is actually a really crappy way to organize lending to industry and individuals, the mismatch between long and short lending will always create problems. Its like an engineering system that has a fundamental instability built in to the design/ As well as being fragile, the current system suffers from huge agency issues and moral hazard (what Government wants to see a banking crisis on their watch). We need to say enough is enough and shut the whole thing down.

Deposit taking (and access to ATMs and other retail banking) is really a totally different business than lending, and confusing the two businesses (while in the interest of the banking industry principles) is the root of the problem. My alternative system is to have short term deposits be required to be invested in the US treasury (which can print money in case of a bank run). In the same way we make Ponzi schemes illegal, we can make other forms of deposit taking illegal.

We can then make funding for industry and individuals separate from deposit taking, this could be from equity or very restricted debt, with limited redemption. The ideal situation would be that all debt would be traded, so that if you wanted to exit ahead of the debt term, basically you would sell the debt into the market place.

There is a claim that not allowing the banking industry to access a low cost pool of savings, such as banking deposits, would increase long term interest rates and so reduce development. I submit that this is not so, for instance the cost of housing is determined by ability to pay. Increase the costs of funds, land prices will come down and just as many houses will be built. We have had a great natural experiment for this, in the last few years house prices fell due to credit scarcity rather than holding constant. Funding for large businesses won"t be a problem, they get their money from bonds already usually. The classic function of banking, lending to small local businesses, barely exists anymore - but could continue to be done by local banks but more explicitly by raising funds. A lot of the arguments against this basically boil down to people saying that transparency in lending is a bad idea - as it reduces lending. Which I don't think is a very solid idea.

Richard - Mea culpa.

Frances - That's correct. You're still backstopping the system using government resources. But, yes, if there's no trust in government securities, then that could be an issue.

ChrisA - You want all demand deposits to be covered by US government securities. Well. So everyone's checking account is now used to fund what must be a truly enormous budget deficit. (No need for a trillion dollar coin!) Of course, the government can lend this money to the private sector. Thus, the government would be the bank, and in fact, there are state-owned banks in many places. Now, let's consider whether one could have a run on a state-owned bank. As it is, I have some experience in this matter. When I worked at Deloitte in Hungary, I was a technical lead on the privatization of the postal bank there--the second biggest bank in the country. One day, emails starting filtering in to the Hungarian staff rumoring a bank run. Within two hours, emails were flooding in to the various firm employees members to get their funds out of the bank, and by mid-afternoon there was a full blown run on the bank. The resulting damage cost the economy 1.5% of GDP. (Needless to say, the bank was not privatized at the time, and it didn't do any favors to Deloitte either. But that's a topic for a different comment.)

In any event, just because a bank is stated-owned, don't assume it can't be subject to a bank run.


You are misunderstanding my proposal. The suggestion is that private banks effectively keep their depositors money in cash in the US Treasury. They would not be able to lend them out, unless they specifically asked the depositor for permission and included restrictions on the ability of the depositor to withdraw at notice. You shouldn't have a bank run in this circumstance, as all deposits are 100% credibly guaranteed (except in contrived circumstances). The banks would presumable charge for the cost of checking, use of ATMs etc, branch costs etc. Of course at the moment these costs are charged in lower interest so it would appear charges would be increased, which might annoy some people. But eventually transparency should reduce these costs.

On funding the deficit, at the end of the day, the Treasury makes cash, so as we have seen with QE, so whether or not my proposal occurs the Treasury can easily fund any deficit. The deficit only really matters for inflationary purposes - not for solvency. I guess there is a concern, along the lines of the starve the beast lines, that the lower the deficit the less constraints on government and the bigger wasteful government, but this is arguable (and is argued a lot). I don't think this is working in reality anyway, as noted the people in power know they can always print money anyway.

On your question of the risk of lack of confidence in the US treasury, surely if that occurs there would be no solvent banks anyway, so the whole question is kind of silly. Can we really have a banking sector if the issuer of money is not credible? I would guess the issue you encountered in Hungary was more people panicking and a lack of appropriate response by the Government/Treasury? They couldn't just issue a 100% guarantee like happened in UK with Northern Rock 2008? Perhaps also people were banking in non local currencies?

I seem to remember the Bush administration trying to get Fannie and Freddie to raise more capital. I think certain Congressmen objected. Good luck with that one.,

ChrisA -

In cash in the Treasury? So my checking account balance is money that is retired from circulation? You have no lending to businesses or individuals?

So if I need a working capital loan, I have to obtain that through equity finance? Wow. That's a pretty restrictive system. And you think that's better overall then a normal banking system with a government backstop? It's seems like you're using a nuke to destroy a wolf. Effective, but probably not optimal for the neighborhood.

On the other hand, if the government does the lending, then we're back to the problems noted above, no?


In my proposal you (as a bank) can lend to business for a working capital loan - but the loan needs to originated from equity funds not from bank deposits. Don't they teach nowadays that there is no difference in funding between debt and equity anyway? The main reason debt is used (or overused) by companies is the tax advantages, you can deduct interest costs. If you don't think that equity investors will be interested in this type of investment, maybe it shouldn't be provided at all?

As to whether this is using a nuclear bomb to kill a wolf - I guess the finance industry liked the results of the last few years but it sure cost the rest of us a lot (and I don't mean in cash only). And it will happen again and again. There is no "regulation" of normal banking (i.e. what we have today) that can prevent this happening again, it is built into the design. When you say to one person - "yes you can have your money back any time you want" and another person "I will lend this money to you for ten years" you are creating a situation that is bound to fail. I agree this proposal will be very bad for the finance industry but that's kind of the point.

Also, back on the issue of this will reduce the total amount of funds available for lending and create a great recession then we can just adapt NGDP targeting.... actually there is a simpler answer - if deposits are lent to the government, government debt will be reduced and the displaced funds go into the market place looking for an investment home so net funds available for investment remain the same. You have to argue that somehow as equity these funds require a greater return than debt to argue the amount of funds will reduce, which goes against the principles of modern finance (ignoring tax issues which can also be fixed at the same time).

Well, the way I understand debt is that it is normally a collateralized obligation; whereas equity is not.

So you're telling me that MM Theory tells us that my bank should charge the same interest on my home loan if it's collateralized by the house or not. Is that right?

Just want to comment that this exchange between ChrisA and S Kopits is one of the best I've seen here in a long time. I hope they continue, I'm learning a lot.

I'm inclined to favor Kopits' argument more. ChrisA's reminds me of many other revolutionary proposals, often from libertarians, which might work well in a blank slate building a society from scratch way but really couldn't be implemented in the actual world we all live in. Path dependence is a very real thing.

Gosh mskings, thanks. Also, sadly you are probably right that my proposal is probably impractical due to where we are today and entrenched interests. But look at what has been tried in the past to solve this problem, everything for the "one branch" solution to Glass-Stegall, to massive regulatory apparatus, Basel I, II and III and so on. Just to pick one example, the one branch approach (Texas) is arguably more radical that what I an suggesting, basically they tried to make depositors do due diligence on their banks because they would know that the bank would be so small there would be no chance of a bail out. It worked for a long while but then after a while people forgot why they did this and allowed consolidation, plus it was a real pain once you went out of state. Also the problem remains for the one branch approach, when the economy goes bad even good banks can fail due to the mismatch between long and short. This was very much the story in the 1930's.

I am definitely out of my league here but I believe that the absolute rate of interest (or dividend demanded) does depend on the collateral offered, but the risk adjusted rate is the same. But you can structure equity injections so that you own the underlying assets (actually this is normal for equity). But actually this is missing my point - I am not proposing to ban loans, just ban loans made from short term deposits. So a bank can make a working capital loan to a company with all the same current guarantees and collateral. But the loan cannot be originated out of depositors funds. It needs to come from an equity fund. In today's world this loan would come from a mix of equity and debt - the equity only gets a high rate of return in this scenario because it take the first loss risk. In my proposal the equity funders get exactly what they would have got from a debt fund, including access to the underlying loan collateral if that's what they want, but the rate of return is lower because they take less risk. Of course if you want to get fancy you can structure different levels of equity, with tranches that take the first loss but get compensated by higher rates of return. This begins to look more like debt financing, so I (as the theoretical regulator) would be taking a close look at these kind of structures before I allow them.

I truly believe that most of the financial industry we have today is about taking advantage of people who are confused or don't understand about leverage, they believe somehow that adding in debt can magically make an underlying asset return better. It is the risk adjusted return that matters, not the absolute return. But try telling someone that the 15% return he is getting on his highly levered equity is actually a lower return than the 2% he gets on savings bonds when adjusted for risk (especially after fees). People are greedy and there is always someone out there ready facilitate that misunderstanding.

Comments for this post are closed