Bank equity: a mechanism for aligning incentives

Kudos to John Cochrane. To address criticism of his narrow banking proposal, he is willing in a recent blog post to “Suppose it really is important for banks to ‘create money,’ and to take deposits, and to funnel those into risky, illiquid, and otherwise hard-to-resolve assets.” He asks the question: How can we fund banks with deposits while at the same time ensuring that banks are easy to resolve? and argues that the solution is to issue the deposits not at the level of the bank (which funds itself only with equity), but at the level of a holding company whose only asset is bank equity, but which is funded by both deposits and equity. Resolution takes place when holding company equity is wiped out and remaining losses are put to depositors.

Cochrane misses the point about what it is that banks do. As I write in a hopefully soon to be published paper:

The fundamental problem of finance is that even though debt plays an important role in production and growth, what constitutes ‘safe’ debt for a lender is precisely the opposite of what constitutes ‘safe’ debt for a borrower. Because the future is uncertain, borrowing is safe when debt is long-term, unsecured, and expected revenues more than cover interest payment on the debt. Lending on the other hand is safe when debt is extremely short-term and grants the lender the immediate right to seize the borrower’s assets as soon as there is any risk of default. The banking system offers a modicum of safety to both lenders and borrowers, not just by transforming the maturity of the debt, but more importantly by placing the burden of loss due to default on the bank owners, whose interests thus lie in making the system of maturity transformation work.

Banking was able to develop in an environment where the owners and the managers of the banks were the same people — that is, banking developed in an earlier era without access to the corporate form and where unlimited liability was the norm. Thus, an important function of the liability structure of the bank was to align the interests of the owners — and originators of the bank’s assets — with those of the depositors. Depositors could only loose money after every single one of the bank’s owners had declared personal bankruptcy (for a famous 1878 example see here).

So what are the economic functions of a bank? First, to offer both lenders and borrowers an asset that is somewhat safe for each of them; and, second, to warrant the safety provided to the lenders by engaging to bear the first loss on any bad loans. A key function of  banking is the alignment of the incentives of the originators of the illiquid, hard to value assets that finance entrepreneurial activity in the economy with those of the lenders who provide the funding for these loans. The fact that the originators bear the first loss is extremely important to the incentive structure of banking.

The U.S. experimented with corporate banking early in its history, but after a first major crisis that extended from 1837 well into the 1840s, adopted a modified form of the corporate structure. From the 1840s until the reforms of the 1930s U.S. bank owners typically faced double liability — so that an amount equal to the par value of the shares could be called up in the event that a bank failed. The assumption built into the 1930s reforms was presumably that bank regulators would be an adequate substitute for the incentives formerly created by the liability structure of the bank.

Thus, Cochrane’s plan has two flaws: First, he insulates the owners of the bank from the consequences of originating bad assets. Instead of being in a first loss position with respect to the bad assets, the owners share only proportionately in those losses. This structure ensures that their interests are not aligned with those of the depositors. Second, his system puts losses on bad assets in excess of holding company equity immediately to the depositors. But this surely means that the depositors will be unwilling to fund traditional illiquid, hard-to-value bank assets. Given the risk of loss, a reasonable depositor would surely insist that the bank’s portfolio be transparent and full of easy to value assets. As a result, Cochrane’s plan would almost certainly result less funding availability for traditional bank borrowers.

In short, while one appreciates that Cochrane acknowledges that his plan needs to address the role played by banks in ‘creating’ money, he has not yet mastered the ideas underlying this view of banking. I look forward to evaluating his next effort.



29 thoughts on “Bank equity: a mechanism for aligning incentives”

  1. I agree totally that loan originators should bear the first loss, and that was the problem I had with Cochrane’s suggestion that the holding company should hold bank equity. We should keep funding loan originators with debt, which is not very difficult to value when we are in information-insensitive circumstances. The issuer of deposits can hold this debt issued by loan originators.

    1. I don’t understand what you mean by “information insensitive circumstances.” I thought “information insensitive” was Gorton’s term for illiquid, hard to value assets. Given this definition surely the standard state of the economy is “information insensitive circumstances.”

      1. Bank deposits are information insensitive when the economy is doing well. They are easy to value – in fact they trade at par, as long as we are sure that the bank is solvent. We don’t need to probe very deeply into the bank’s investments to arrive at a precise valuation for something that trades at par.

        If we remove the option to swap for cash, senior bank debt should still have these properties. That debt is what I would hold in the bank holding company, money fund, or whatever we are using to provide deposits and run the payments system.

  2. I don’t buy in to Gorton’s banking theory. I think it is at best a model of a very troubled banking system. In his model bank deposits can only be information insensitive because the bank’s assets are illiquid, opaque and very hard to value. Trading at par is just an interim condition before there is a crash and depositors (or taxpayers) lose their money. In Gorton’s environment we are never really sure that the bank is solvent; it’s just that most of the time it’s too expensive to bother determining whether the bank is solvent.

    Gorton’s whole model hinges fundamentally on the assumption that unsecured short-term credit is an impossibility. Given how common unsecured short-term credit is (e.g. credit cards) both in the past and in the present, I don’t think the model has much to do with the real world.

    1. That’s an interesting critique of Gorton’s theory. Perhaps I complicated the discussion unnecessarily by mentioning his concept. But I’d like to return to my point of agreement with you: contract theory tells us that loan originators should bear first loss. John Cochrane keeps overlooking this point somehow with his insistence on equity funding for loans.

      What I like about Cochrane’s post is the separation of the banking system into two layers – with loan origination on the bottom layer, using “safe” (in your terms) funding of some form. I personally like long-term debt, rather than Cochrane’s equity. The higher layer of the system provides this safe funding, and funds itself mostly with deposits. He has been writing about a “run-free” system for a few years, while my own motivation is to greatly reduce the supervisory burden on the payments system. In the US system, as it is organized now, thousands of loan originators are supervised and participate in the payments system. This requires excessive effort and a high number of supervisory personnel, and yet it remains fragile. We can make a system that is simpler, better, and perhaps better able to cross national borders by using more careful and deliberate organization than we have had in the past.

      1. It’s not clear to me what advantage comes from your two tier structure. Illiquid, hard-to-value, but often short-term loans are funded by e.g. 10% equity and 90% long-term debt. A holding company holds all of the long-term debt and issues mostly deposits and a little equity against it. It seems to me that by inserting long-term debt into a banking system that usually transforms deposits (zero-term debt) into short-term and medium-term debt (e.g. lines of credit, credit cards, etc.) you will introduce destabilizing maturity mismatch rather than solving any problems.

  3. The advantage is that the optimal scale of these two activities is different. Lending benefits from specialization, in terms of geography, industry, and other variables. Deposit-keeping benefits from diversification, especially over geography. The way to reconcile these two scales is to have two layers to the system. If we still need supervision of depositories, they will be larger and fewer in number than the thousands that participate in the US payments system now. And if their holdings are limited to bank debt that is traded, the task of supervision is relatively simple. I will go out on a limb and say that they might simply police each other, by deciding when and when not to accept each others’ deposits at par.

    1. I’ll have to think about this. I tend to think of deposits as the consequence of lending and therefore of the two as naturally belonging on the same balance sheet. It’s not clear to me that the types assets that traditionally sit on bank balance sheets are good candidates as backing for marketable debt. I think of bank balance sheets as inherently opaque.

  4. Yes I remember that you want to restrict the types of assets that are eligible for monetary finance, especially to exclude capital investments that are subject to large fluctuations in value. I don’t have strong views on such restrictions, but the underlying concern seems to be orthogonal to the one John Cochrane addresses with his dual-layer structure – which is to simplify and facilitate the resolution of failed lenders. He wants to fund lending with equity, to eliminate this possibility of failure, whereas I am fine with long-term debt funding as an improvement over funding with deposits and enmeshment in the payments system.

    I don’t know if you follow fintech developments in India, but they have a new regulatory category of “payments bank” which is part of the payments system, but restricted to liquid investments. Closer to home, we have businesses like Paypal which offer deposits, and keep looking for good regulatory accommodation. These are examples of the emerging second layer that I have in mind.

    1. It’s not clear to me in what sense the asset structure of a bank can possibly be orthogonal to its liability structure. I do understand that Cochrane’s focus and yours is the resolution of failed banks. The question I am trying to raise is what implications the proposed change in the liability structure of banks have for the asset structure of banks. I am far from convinced by the implicit claim that the two sides of bank balance sheets can be addressed independently.

  5. I think you and I have similar views on the proper liability structure of banks: they should be funded with “safe” (in your terms) debt, so that they bear the first loss on their investments. And this type of funding is indeed connected with the type of investments they make; the difficulty of monitoring obliges us to rely heavily on optimal contracting. I would make the debt safe by using long-term contracts, while you have a theory of incentive-compatibility, that relies on mechanism design strategies.

    Bank deposits can also be called by individual depositors. Perhaps it is wrong for me to say that the allocation of these call options is orthogonal to the restrictions you want to put on bank assets, because there are theories that justify callable bank debt as an enhancement to the contract between the principals and the agents of bank investments. I think your own theory of incentive-compatibility also supports callable deposits.

    But we can imagine a different arrangement for contingent funding, as part of a separation of banking into two layers – as John Cochrane contemplates in his post, and I do as well in my comments here. There is a post by Prof. Jayanth Varma discussing bank deposits without a connected swaption for cash: “Bank deposits without those exotic swaptions”.

    Even if we accept the arguments for a call option (to improve incentives) he notes that consumers typically don’t exercise the option in an optimal manner. I’d rather assign the right to revoke additional funding (without triggering fire sales) to the asset allocation professionals at the second layer of the banking hierarchy, and let them exercise it in a way that reduces the likelihood of irrational panics and the need for last-resort lending by central banks.

    So to sum up: callable deposits issued by the higher layer, and revocable credit lines provided for the lower layer, which frees the thousands of US banks currently at this lower layer from the need for supervision, and last-resort lending.

    1. I don’t believe it’s possible to have the kind of financing of entrepreneurship that banks have traditionally provided without also having a need for a lender of last resort. The process of making working capital a liquid asset comes at the price of generating pure liquidity crises (by which I mean that origination methods can be perfect and you will still have liquidity crises). The lender of last resort will be needed or the liquidity of working capital will be lost. I don’t see any middle ground between these two options.

      Loans financing the working capital of small entrepreneurs are by their nature not loans that can be priced in markets. The banking system is basically a huge mutual society that provides credit to traders large and small (see )

      Trying to finance typical bank loans using equity or long-term debt is just a means of eliminating the finance of small entrepreneurs. Which is actually a process that I would guess has been underway in the U.S. for the past few decades and is one of the explanations for why productivity has been so poor. The shift from bank-based to market-based finance means that our financial system is providing less and less of the kind of finance that supports a dynamic economy.

      My current work is focused on explaining why it is the creation of a lender of last resort that makes modern growth possible. To eliminate the need for a lender of last resort is to eliminate the kind of debt that made 19th and 20th c. markets liquid and to eliminate the kind of debt that supports modern growth.

  6. I agree totally that we need an elastic supply of loans to small entrepreneurs, and that these loans cannot be priced in the market. But bank debt can be, and is priced – whether it is issued to depositors, or whether it is a credit line issued by an entity that can create money.

    The problem I have is with making the debt of loan originators callable. This makes them dependent on the central bank and its supervision regime, when depositors exercise their options en masse, in a run. I can’t see many reasons to think that giving a large and dispersed group of depositors the option to run is better than giving asset managers the right to revoke further use of a credit line, if a lender is performing poorly. One can make an argument that we don’t want large asset managers to have too much power. But their power is a fact today, and not something that I am proposing as a change.

    1. Not quite sure what you mean. Bank debt that is backed by loans that cannot be priced trades for nothing (or a small fraction of the loans’ value) on a market. Therefore it cannot be used to finance the issue of those loans. The economy will have to do without this type of loan if we shift to market based finance.

      On asset managers: given the devastating problems created in the recent crisis by the sort of wholesale funding that you are talking about, it would appear that there is a far greater danger of runs by asset managers than there is of runs by depositors. The massive shift away from such wholesale funding and back to deposits is an important factor creating greater stability in the U.S. banking system today.

      1. Are you not counting bank deposits as bank debt? That is what I’m talking about, and the problem of including a swaption for central bank money. Presumably we pay for the option with reduced interest rates, as well as a costly supervisory apparatus that still fails on occasion.

        I would say that the problem with wholesale funding is the lack of safety, and not the simple fact that it is wholesale. Like retail bank depositors, other lenders have the option to run, which means a withdrawal of their investment, and not simply a stop to further use of their credit facilities. We should agree on that as a structural flaw.

  7. Bank deposits are non-marketable bank debt. They serve a completely different function and I would argue are more stable and more capable of funding illiquid entrepreneurial loans than marketable bank debt.

    I don’t agree that the “runnability” of bank deposits is a structural flaw mainly because I view it as an unavoidable component of the kind of finance the economy needs. In order for it to be a “flaw” there must be a way of “fixing” it that doesn’t have more costs than benefits. Runnability is a characteristic of banking for which the lender of last resort was developed. Only after we had an de facto lender of last resort (Bank of England from about 1763 on) was modern economic growth possible.

  8. “Suppose it really is important for banks to ‘create money..”. Cochrane’s whole argument rests on that assumption. It’s a false assumption for the simple reason that a central bank & government can create whatever amount of money is needed to keep the economy at capacity. I.e. privately created money and the risks that go with it are wholly unnecessary.

    The main factor in the way of a ban on privately printed money was pointed out by Milton Friedman in the preface of his book “A program for monetary stability”. As he put it, “The vested political interests opposing it are too strong….”.

    1. Thank you, Ralph, for the comment. I believe Friedman failed to grasp what it is that banks do, and unfortunately passed his confusion on to large segments of the economics profession. I’m curious what evidence has convinced you that ” a central bank & government can create whatever amount of money is needed to keep the economy at capacity.”

      My model of banking is available here:

      1. I believe Ralph is using Friedman’s monetarist thinking: that aggregate demand can suffer due to a shortage of base money (which is produced by the state). I remember you being very critical of this view in the blog post before this one (on banking theory), where you work very hard to stress the link between bank lending, the price level, and aggregate activity.

  9. Carolyn, You ask “…what evidence has convinced you that ” a central bank & government can create whatever amount of money is needed to keep the economy at capacity.” My answer is thus.

    First, there is no question but that central banks have created astronomic and unprecedented amounts of money in recent years to fund QE and rescue dodgy banks. Second, there is nothing to stop government and central bank simply printing money in any quantity they like and spending it on whatever: infrastructure, health, education, etc. Alternatively they can use the new money to cut taxes: i.e. increase households after tax incomes – a form of helicopter drop.

    Indeed, traditional fiscal stimulus combined with QE (what we’ve actually done in recent years) comes to the same thing as the latter “print and spend”.

    And there is no question but that given enough of the latter “print and spend” stimulus, the economy can be kept at capacity – not that I’m suggesting that is an easy trick to pull off.


    I’m certainly invoking Friedman to SOME extent. That is, I don’t see how it can be denied that the amount of base money in private sector hands has some sort of stimulatory effect.

    However, strikes me that in addition to that monetary effect, the simple fact of government ordering extra infrastructure or whatever, also has an effect. That’s what might be called a fiscal effect: something Friedman ignored far as I can see. But I could well be wrong there – perhaps he did mention it somewhere.

    Plus the empirical evidence supports that monetary and/or fiscal effect: the Bush tax cuts induced households to spend more.

    Next, you say in reference to Carolyn, “you work very hard to stress the link between bank lending, the price level, and aggregate activity.” I don’t know of anyone who denies that link. Certainly I don’t. Steve Keen has done a lot of research on that point.

    Unfortunately private banks use their power to influence “aggregate activity” in a totally irresponsible manner. That is, during a boom, they create and lend out money like there’s no tomorrow, thus exacerbating the boom. Exactly what we don’t want them to do. Then come the bust, they call in loans, which has a deflationary effect. Again, exactly what we don’t want them to do. I.e. they act in a pro-cyclical manner.

    In addition, they go bust regular as clockwork and spark off bank crises followed by years of excess unemployment. All in all, I agree with Friedman and various other Nobel laureate economists (James Tobin, Maurice Allais, etc) and James Cochrane, that private money creation should be banned, with the state being the sole issuer of money.

    Incidentally, I left a comment on JC’s blog about 24 hours ago, to which he gave a favourable response.

    1. Ralph I can see why you and John Cochrane want to eliminate private money creation for safety reasons, because the banking sector can be so fragile. But you are very dismissive of the possibility of macroeconomic regulation through a central bank that targets, let’s say, NGDP. The previous blog post here on banking theory discusses such a partnership where the state controls the value of the accounting unit, while the total amount of money (i.e. bank lending) is endogenous. John Cochrane btw regularly writes in support of inflation targeting, as an end in itself, without spelling out a link with broader macroeconomic regulation.

      Now if you and John Cochrane propose the outright ban of bank money, this leaves open the question of how exactly the regulatory mechanism for the economy will operate. I know that you are critical of the jobs guarantee suggested by the post-Keynesians, though you do mention fiscal policy. As for John Cochrane, I’ve pointed out that the academic literature criticizes narrow banking proposals for their difficulty in responding to economic shocks in an elastic manner. He doesn’t respond to such comments, which leaves a significant gap in his proposals.

      1. Deepwatrcreatur,

        I’m baffled as to why you think “…you are very dismissive of the possibility of macroeconomic regulation through a central bank that targets, let’s say, NGDP..”.

        I specifically set out such a system above. It consists of “central bank and government create new money and spend it into the economy” either via extra public spending and/or via tax cuts. Positive Money and others have been advocating exactly that system for years. See:

        Click to access 3a4f0c195967cb202b_p2m6beqpy.pdf

  10. Ralph, I am certainly willing to grant that CBs/governments have the ability to create as much money as the economy needs. Indeed, I would argue that this capacity is essential to the operation of the banking system.

    In my view, however, the challenge is not the creation of money, but the allocation of it. The reason money creation has been outsourced to banks since the dawn of modern growth (approx. 1750) is that a well-managed banking system’s job is to master the “science” of credit (as Thornton 1802: 175-76 puts it). I think Friedman’s formula for government-based money runs into the same problems as any command economy: just as prices are needed to allocate inputs and outputs, so banks are needed to allocate funding.

    I agree that modern banks are catastrophically bad at performing banking functions, but I believe that is due in no small part to Friedman’s success and the dismissal of banking’s importance leading to poor bank regulation when what we should have been doing over the course of the past 80 years is studying how to make banking work better.

    1. Speaking of poor bank regulation, the folks at Basel have taken the system in a direction probably opposite to what you would recommend, based on your historical studies. Their risk weights strongly encourage real estate lending and there’s nothing in their system about real bills or trade credit. It really makes me wonder what type of useful intervention any of us can make wrt the regulatory apparatus.

      1. I suspect that Basel’s greatest error was in treating unsecured lending as most risky and marketable collateral as a means of offsetting the risks. This had the effect of encouraging banks to rely on the markets that they exist to support. Basel encouraged banks to unlearn how to do classic bank lending and to focus on making sure that the bank was protected by “marketable” collateral. The effect was to destabilize markets for collateral.

        Basel is a classic case of experts thinking that they know more then they actually know. Brings to mind Confucius: “To recognize what you know as what you know and to recognize what you do not know as what you do not know, now that is knowledge.”

    1. Absolutely. Deliberately treating business loans as less favored than mortgages is simply bizarre. Somebody didn’t think it through.

      1. And it has been 30 years since that was approved in Basel I and yet no one has been questioned or much less been held accountable for that mistake.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: