Thursday, June 5, 2014

The Economist on Narrow Banks

The Economists Free Exchange blog covers narrow banks, and parts of my "run free" paper in a post somewhat mean-spiritedly -- or perhaps unintentionally self-descriptively --  titled "Narrow Minded." I always appreciate publicity, but a few parts seem wrong enough to address.

After nicely covering the history of the idea, the Economist writes,
such a plan raises huge practical questions. The first is implementation: how to get from today’s system of highly indebted banks to one in which they are financed chiefly by equity. 
That's not hard. We're slowly raising capital requirements, and all we have to do is to keep raising them. My Pigouvian tax on debt would help a lot -- I think banks screaming how hard it is to issue equity or how terrible not to pay dividends for a while would suddenly find it much easier if paying 5 cents for each dollar of debt issued. Announcing that institutions above 50% equity and with less than 20% short-term debt are exempt from Basel and Dodd-Frank asset regulation might cause a rush for the exits.
Politically, there would be formidable opposition from vested interests. 
And this is, somehow, an argument against the plan rather than for it? There is formidable opposition from vested interests against abolishing agricultural subsidies, trade protection, occupational licensing, and taxicab monopolies. Dear Economist, when did feeding the cronies become an argument for keeping bad policy in place, not a main indicator of needed change?

Economically, the transition would require banks to dispose of a vast stock of loans, or raise an equivalent amount of long-term debt and equity.
The first is simply untrue, and the second is deeply misleading. For every dollar of long term debt or equity that must be raised, one dollar of short term debt is paid back. No extra funds from investors are required, and no selling of assets is required. It's just a Modigliani-Miller / Yogi Berra reslicing of the same pizza.
A second concern is whether a split between narrow banks and wider lending-and-investment firms would actually eliminate runs. If other institutions replace banks in making loans, they could end up creating fragilities of their own. Mutual funds, for example, are financed by shareholders, not creditors; but if such shares are seen as stable and safe, investors will treat them as deposits—and try to withdraw their investment if that safety is threatened.
This is just simply wrong, and in the "Economist should know better" camp. You cannot "withdraw your investment" from a floating-value  fund.  The fund makes no fixed-value promises. It cannot fail. It cannot suffer a run. Look up the definition of run, dear Economist! A floating-value fund, and especially an exchange-traded fund with no one-day NAV promise,  is the paradigmatic example of a run-proof institution.

Yes, investors can all try to dump stocks, either held directly or held through funds, and stock prices can go down. There is no failure, no bankruptcy, and no crisis in this. We want a system that allows booms and busts without crises, not the promise that wise regulators will step in to stabilize stock prices!
After this happened even once, people would simply flock to the narrow banks, and there would be no source of lending.” To prevent this, the authors argue, governments would have to intervene to save the “not-so-narrow intermediaries”.
Now we're deep into the silly season. The intermediaries do not need any saving. They have not made any promises. A floating value fund cannot go bankrupt! Yes, stock prices can fall, and your fire sale is my buying opportunity. Do we really want Governments and their central banks buying stocks to prop up their values? Do we really want governments allocating credit? Have we so lost sight of what a "crisis" is, and is not?
Third, such a system would still need plenty of regulation.
The fact that we need some regulation -- that I don't produce a libertarian-anarchist nirvana solution in which absolutely zero regulation is required -- is somehow a defense of the current monstrous setup? I think we need cops at stoplights. Is this a defense of Dodd-Frank? Come now, it takes about 1/10th the regulation, because we can throw out all regulation of the safety of bank asssets, all the risk weights, all the stress tests, all the "resolution," and so on. The perfect is truly the enemy of the good at the Economist.
But given the growing cost and inefficiency of today’s regulatory regime, the concept of narrow banking surely deserves more serious consideration.
I'll take the grudging endorsement and return a grudging gratitude for the mention of the idea!


  1. The big weakness of these proposals is that they want to give us bank stock in place of the inside money we hold now, but they don't anticipate the consequences of this shift. I'm trying to read thinkers like Perry Mehrling to better understand the role of money, and why we have it at all.

    1. “but they don't anticipate the consequences of this shift”. What are you on about? It’s stark staring obvious what the consequences are: the cost of borrowing will rise a bit. But that only reflects the removal of the subsidy which the existing banking system benefits from. What’s wrong with removing a subsidy for which there is no justification?

    2. I agree that banks should not be subsidized. The consequences I have in mind are explained in the work of Rene Stulz and Harry DeAngelo. Look at their paper titled "Liquid-Claim Production, Risk Management, and Bank Capital Structure: Why High Leverage is Optimal for Banks". Prof. Cochrane proposes a reduction of liquid claims to the liabilities of the US Treasury. It is reasonable to ask if that is sufficient to meet transactional needs, especially when a lot of that debt is demanded for other purposes.

      Prof. Cochrane wants a run-free financial system, but he also wants firms to be subjected to market discipline, and allowed to fail if they are run poorly. Raghuram Rajan and Douglas Diamond (both are Chicago guys, qualified to evaluate the Chicago plan for banking) have a theory of banking that places more emphasis on the second point than on the first. They say that allowing runs is a way to impose market discipline.

      I tweak their argument by requiring banks to buy macroeconomic insurance. Then if a run happens, it would not simply be a result of an economic downturn. In other words, we a more effect contract with bankers. And bank failures would not be so correlated, which means that it is more likely that we can simply allow them to happen, without bailouts or implicit bailout subsidies.

    3. Anwer,

      I’ll have a look at the Stulz and De Angelo paper.

      Re your first paragraph, if the economy did not have enough money for “transactional needs” that would show up as deflation (in the “demand reducing” sense of the word). All government / central bank would need to do would be to create and spend more into the economy (and/or cut taxes). Indeed under the EXISTING SYSTEM, where the private sector thinks it has an inadequate stock of cash, spending falls, and that “shows up as deflation”.

      Re your second para, if we allow runs with taxpayers reimbursing depositors who stand to lose out (which is the system in the UK - and doubtless elsewhere) that comes to much the same thing as the full reserve system (where money which depositors want to be totally safe is lodged with government.) The only difference is that under former variation on the theme one is subsidising private banks, which I don’t agree with.

      Re Raghuram Rajan and Douglas Diamond, I don’t have the respect for them that you do. I demolished some of their arguments respectively here:
      and here

      Re the claim you attribute to the above two authors, namely that allowing runs imposes market discipline, discipline is imposed under full reserve by letting shares in lending entities fall in value when they make silly loans. So I don’t see what Diamond’s system achieves that full reserve doesn’t also achieve.

    4. Ralph, I do believe that there should be a full-reserve system. In fact I like the one described in Prof. Cochrane's post on floating-rate treasuries, which also proposes a reform of government debt. But there is increasing evidence that there isn't enough safe government debt to meet ever-growing transactional needs worldwide. They are increasing the supply of US debt with quantitative easing, but there are obvious limits to that program. That's why I'm trying to think of a more adequate system of monetary plumbing, and one that doesn't contract during crises and exacerbate them.

      So in addition to the full-reserve system, I like the idea of a fractional-reserve system for those willing to take a little risk and hold Trills. Fractional reserves of dollars makes banks fragile, but fractional reserves of Trills provides them with macroeconomic insurance. It seems like a definite improvement over the existing system.

    5. Anwer, Government debt is very rarely used for “transactional needs”. It’s MONEY that is used for transactions (central bank created money or commercial bank created money). And if the private sector thinks it doesn’t have enough money, that will show up as reduced spending: a sort of Keynsian “paradox of thrift” situation.

      But that’s easily dealt with simply by having government and central bank create and spend money into the economy (and/or cut taxes).

      As to government debt, that’s similar to money in that it fulfils the “store of value” function of money (but as I just said, not the transaction function). A shortage of debt would also show up as reduced aggregate demand. In fact advocates of Modern Monetary Theory refer to base money and government debt as “private sector net financial assets”.

      Re a “monetary plumbing” system that doesn’t “contract during a crisis”, that characteristic is found in a full reserve system. Under full reserve, the only form of money is central bank money (base money), and the volume of that only contracts if government / central bank decide to reduce or contract it. In contrast, privately created money or fractional reserve money collapses in a crisis.

      Re the DeAngelo and Stultz paper, I’m not impressed. It’s written in near impenetrable jargon. And as far as I can see they just make the point that absent a central bank, private banks perform a useful service: the creation of money or liquidity. Now the latter point is true, but the authors don’t address one of the main merits of full reserve (which I mentioned above), namely that money or “liquidity” is provided by the government / central bank under full reserve. The paper just doesn’t feature phrases like “central bank”, “base money”, etc. A bit of blunder, I think.

    6. Anwer,

      "Fractional reserves of dollars makes banks fragile, but fractional reserves of Trills provides them with macroeconomic insurance."

      How so? The return (coupon) on Trills is positively correlated with economic activity (nominal GDP level). The only benefit they provide is to stabilize the federal government's fiscal position. Higher nominal GDP generates higher tax revenue allowing government to pay higher coupon.
      For insurance, a bank would want an asset with a return that is negatively correlated with nominal GDP.

      It seems that fractional reserves of Trills would make banks even more fragile. During a recession, the nominal value of an individual dollar held as reserve remains constant. The nominal value of a Trill held as reserve could conceivably fall.

    7. Frank, by "fractional reserves of Trills" I mean that banks should issue more Trills as liabilities than the amount of government-issued Trills they hold as reserves. Currently the dollar-denominated liabilities of banks remain fixed during downturns, while their assets typically shrink in value, due to poor performance.

      If both assets and liabilities were denominated in Trills, they would both move in a similar fashion with macroeconomic fluctuations, and the solvency of banks would not be as correlated with these fluctuations and with each other. This makes systemic crises less likely, and reduces the need to bail out troubled banks.

      The government-issued Trills held by banks obviously would not be for insurance purposes (it's the Trill bank *liabilities* that do that). Rather, such holdings would be necessary if Trills were part of the payments system, so that liabilities could be traded between banks at par. We currently have a similar relationship between bank-issued dollars and those issued by the government and held in accounts at the central bank.

      As I said before, I do like the narrow bank proposal that makes all government debt perpetual, with all of it being usable for transactions. Currently, short-term debt is more useful than long-term debt for transactional purposes. But Prof. Cochrane combines that proposal with a requirement that bank lending should be funded 100% with equity, and suggests that this bank equity could also be used as money. I think we agree that the liquidity of this bank stock would be reduced at times, which makes me doubt that payments networks would accept it as a monetary unit. Its use as collateral would be similarly questionable.

      Another problem with the 100% equity-funded bank is that sometimes it would need to raise new equity when there is widespread confusion about the value of bank stock. The alternative would be not to make any more loans until there are better prospects for new issues of stock to get the dollars needed to loan out. What effect would this constraint have on macroeconomic fluctuations? Currently, banks can supply loans very freely, by creating dollar deposits whenever they want - for approved borrowers. If loans and deposits were issued in terms of Trills, banks would have the same flexibility that they do now, and lending would not be constrained by the liquidity of their stock.

    8. Anwer,

      "If both assets and liabilities were denominated in Trills, they would both move in a similar fashion with macroeconomic fluctuations, and the solvency of banks would not be as correlated with these fluctuations and with each other. This makes systemic crises less likely, and reduces the need to bail out troubled banks."

      "If loans and deposits were issued in terms of Trills, banks would have the same flexibility that they do now, and lending would not be constrained by the liquidity of their stock."

      I presume you are talking about Trills as both a medium of exchange and a unit of account here.

      How would a negative return on a Trill work? Presumably you could still have recessions with a falling GDP level. Does $50 in Trill currency become $45 in Trill currency during a recession? Why wouldn't this create a negative feedback loop?

  2. " That's not hard. We're slowly raising capital requirements, and all we have to do is to keep raising them. My Pigouvian tax on debt would help a lot -- I think banks screaming how hard it is to issue equity or how terrible not to pay dividends for a while would suddenly find it much easier if paying 5 cents for each dollar of debt issued. Announcing that institutions above 50% equity and with less than 20% short-term debt are exempt from Basel and Dodd-Frank asset regulation might cause a rush for the exits."

    It all comes down to growth. If there is growth then banks should have no problem issuing equity becuase they will be more profitable. If there isnt growth then no system of banking may be viable including the current one.

    "Economically, the transition would require banks to dispose of a vast stock of loans, or raise an equivalent amount of long-term debt and equity."

    Obviously the transition could be a challenge but couldnt special measures be made for the transition until the new system is established? For example the central bank could provide the long term loans for a transitionary period.

  3. Wasn't that Yogi Berra? Not Casey Stengal.

  4. Re the transition to full reserve banking, Milton Friedman said: “There is no technical problem in achieving a transition from our present system to 100% reserves easily, fairly speedily and without any serious repercussions on financial or economic markets.”
    As to Grumpy’s point about “vested interests”, personally I wouldn’t use the phrase “vested interests”: I’d say something like “the criminals, fraudsters and blatant liars that make up the senior ranks of the banking industry”. Anyway…

    Milton Friedman made a similar point on the question as to why full reserve had not been adopted so far. He said,

    “The vested political interests opposing it are too strong, and the citizens who would benefit both as taxpayers and as participants in economic activity are too unaware of its benefits and too disorganised to have any influence.”

    The quotes are from Friedman’s book, “A Program for Monetary Stability”.

  5. To summarize: Our current system of social democracy, fiat money issuance and fractional reserve banking inevitably seems to wind up creating more fixed claims than can be satisfied by the real economy. Some day, some how, those claims will need to be "broken up" and reallocated and, hopefully, institutions can be put into place to prevent them from getting out of hand again. Narrow banks are at least a step in the right direction.

  6. In my view, it is likely that your narrow banking solution fails to take into account the economic function of banks.

    I have a theory of money, banking, and economic performance, expressed in this paper,, which argues that modern economic growth was made possible by the maturing of banking in late 18th c. Britain. This evolution created an environment where incentives were carefully aligned to enable the private sector to issue “safe” assets abundantly, and allow a vast number of individuals to overcome the liquidity constraints that characterize most of our economic lives. In short, I argue that the existence of a functional banking system is what makes neoclassical economics (with its complete absence of liquidity constraints) imaginable.

    In the process of providing a solution to the general public's liquidity constraints, the banks must take on liquidity (and credit) risk. This is their economic function.

    In my view, the problem with modern banking is the destruction of the incentive structure that made the issuance by the private sector of “safe” assets incentive compatible. This destruction has taken the form of legal and regulatory changes, most of which took place from the 1980s on – and included dramatic changes to the concept of a “lender of last resort” and its duties.

    While I can agree that continuing with our current financial system is a recipe for continuing financial crises, I have profound doubts that your solution is consistent with the rates of economic growth to which we have become accustomed in recent centuries.

    1. Thanks for the link and great comments. I do think that technology undermines many of the problems we had before. Widespread equity finance, and using floating-value securities as money was simply not possible before computers and instant communication.

    2. You've made this point in earlier posts: modern technology allows us to use equity as money, because we can get quotes instantly from high-frequency traders willing to buy our securities instantly. That works most of the time, but during crises bank stock can become illiquid. If banks are financed purely with equity, then they won't have solvency crises, but their equity value might be in question during a fire sale.

      If they issue "safe" liabilities, then we only need to make a judgment regarding solvency before accepting these liabilities from them at par, as payment - whether that liability is denominated in dollars or in Trills. Before accepting equity as payment, during a crisis, we might need to negotiate a haircut, and that could greatly affect the functioning of payments and collateral systems.

      Some have observed that HFT gives us plenty of liquidity at times where liquidity isn't needed so much, but none when it is needed desperately. "Safe" assets (in Gary Gorton's sense) are a much better way to get liquidity, and we need more of them due to ever increasing transactional demands worldwide. You say that people can simply use US Treasury debt for that purpose. Well, they already are doing that. Yesterday I read an article that traders in China are even using (and re-using) metal inventories as collateral to facilitate business loans.

    3. Traditional banking works precisely because the banker couldn’t sell debt without guaranteeing it. Thus, banking, unlike a market in debt, was designed to align incentives so that bad debt was not issued. Because true sales (in accounting terms) are antithetical to the alignment of incentives in the origination of debt, I don’t see how your market-based risk allocation process can possibly generate the same amount or quality of debt (with or without technology).

    4. I can see from the abstract of your paper that you are very well-versed in the work of Perry Mehrling; indeed you cite him many times in the paper itself. I'll try to go through your article so that I can understand your objections better. BTW: I don't like shadow banking either, and am trying to think of reforms that attract more people to the regulated system.

  7. John: I guess that one of your responses to the following is that the authors are ignoring the importance of the floating-value constraint for the mutual funds they say would be subject to runs under narrow banking:

    1. Exactly. At worst there is a problem that floating NAV funds promise a price for one day. That's easy to fix, and has been at most equity funds. ETFs don't even have that. See above on "floating value fund cannot go bankrupt." They're confusing a stock price plunge, which certainly can happen, with a run, which cannot happen absent a fixed - value promise. I left out the economists' cite of this work on the hope it had been misquoted.

    2. The above "money and banking" article contains a whole string of errors: too many to deal with here. So I've just done a post on my own blog about it.

  8. When I read this back in April (after having read some of your narrow banking stuff), I was confused trying to figure out why two experts like Cochrane & Cecchetti disagree on this point. Now I understand better. Thanks.

  9. John, you mention technology with regards intermediation. Have you also considered technology with regards money creation?

    You write "Our government should take over its natural monopoly position in supplying interest-paying money." This sounds a lot like the UK pressure group "Positive Money" who have been caricatured as wanted a roomful of men to decide how much money there should be. This seems overly centralised and not nearly sensitive enough to moves in the economy, as compared to the distributed lending decisions of thousands of bank branches.

    It seems clear to me that we would rather have money-creation tied to the "needs" of the economy, than at the whim of committee, however well-intentioned. However I believe I am agreeing with you, when I say that putting the power of money-creation in the hands of banks, is asking for trouble.

    I would propose going direct to source, and tying money creation to money velocity. This would require a realtime aggregate data of money turnover, but fortunately the blockchain can offer this (not with the popular bitcoin implementation, but the refinement is trivial). There are more details here, in a proposal that combines your full-collateralised banking proposal, with the transition to eMoney (Kimball et al) This gets us NGDP targetting, all the benefits of fiat (limiting deflation) & gold (controlling inflation).

    The effect on the man on the street would be to see a very small amount of money credited to his emoney account every night. Hardly the stuff of revolution.

  10. Hi John, great work. Q: is there a way of thinking about the equilibrium level of reserves. My difficulty is, why 100%? Why not 90%? or 80%? etc. Absent an "equilibrium", what is the difference between 100% backing a "high" liquidity ratios?

  11. Mr. Cochrane refuses to answer the fundamental question of why anyone would buy bank equity if the return on that equity is less than the yield on the risk-free government bond. As I pointed out in Mr. Cochrane's previous post on run-free system, during the period 1996 to 2005, the approximate average yield on the 10-year treasury bond was about 5%. The yield on mortgage loans was about 6%. Even if the bank had zero non-interest expenses (which we know is not possible), the after-tax return on loans made by banks would be 4% i.e. 6% x (1-tax rate) which I, for the sake of convenience, assume is 33%. Since, the bank is 100% equity-funded, the 4% also happens to be the bank's ROE. So, I repeat my question. Why would investors buy bank equity if they know that their return is going to be only 4% especially when they can get 5% on the risk-free bond? Mr. Cochrane's system works only when the spread between loans and the risk-free bond is wide enough to offset the impact of corporate taxes. In an algebraic sense Yield on Loans x (1- corp tax rate) > Risk-free rate or put another way, Yield on Loans > Risk-free rate / (1- corp. tax rate). Let us also not forget that I have made a highly dubious and simplifying assumption that non-interest costs are zero. If you bake that in, the system almost certainly doesn't work unless Mr. Cochrane proves otherwise.

    1. "refuse" nearly got you in the deleted pile.
      I dealt with this at length in the paper. Equity of less leveraged banks is much safer. Yes, MM doesn't hold, because banks lose their subsidies. Good.

    2. Are you telling us that investors would be willing to accept a 4% return on a bank stock when they can get 5% on a risk-free bond?

  12. Also, a simple suggestion: what if the FDIC announced that all large banks have to offer T-bill accounts to all corporate customers, making it clear that these are by definition 100% guaranteed by the government. Then simultaneously remove the FDIC guarantee on all corporate deposit accounts, making it clear that they are no longer guaranteed. If this works, extend to households.

    1. Eric Lonergan,

      Re your first sentence, there are different ways of implementing full reserve banking, but what you suggest is exactly what many full reservers propose.

      Re your second sentence, I think you’ve confused two issues. I.e. the fact that a bank account is held by a corporation is not a reason to deny the corporation a totally safe bank account. That is, under full reserve, there would be (1) totally safe accounts available to anyone: households, corporations, you name it, and (2) accounts where relevant sums are loaned on or invested, but the account holder carries the risk, not the taxpayer.

  13. Also, I'm not totally convinced that floating NAV funds can't have a run. Yes in contrast to a deposit they promise to redeem based on mk-to-mkt. However they can have huge liquidity mismatches. Consider the following, plausible, scenario: 1) A very large single manager who has outperformed over a long period of time gathers a dominant share of assets in debt securities, the fund provides daily liquidity; 2) an event/rumour triggers daily withdrawals; 3) the underlying assets are highly illiquid (and other market participants know this); 4) nav falls rapidly as the fund sells into v wide bid-offer sprds; 5) price declines trigger accelerated withdrawals ... That does not sound like financial stability? Ironically, "riskier" but highly liquid funds might be less "run-prone" than "safe" (in terms of credit risk) but illiquid - that's obvious on reflection, but implies regulation?.

    1. Agree, Eric. Retail investors have no idea of the potential lack of liquidity inherent in floating NAV funds, especially for those funds invested in liquid assets like leverage loan ETFs. Yes, you can always find a buyer, but only with a huge haircut. Buyer beware? Perhaps the investor base will a more narrow voting block then homeowners, so the Fed will stay on the sidelines this time, even in the face of massive wealth destruction. But I there will be calls for another bailout and to your point, at the very least, this proposal is hardly financial stable.

  14. Let me correct the first paragraph: Cochrane suggests that bank stock is as usable a liquid claim as bank deposits are, now that we have HFT. In other words: you can hold bank stock, and sell it when you need to make a payment to someone. But HFT activity dries up at some times. Does that mean we must let the payments system break down at those times?

    What we need are claims that remain liquid even during crises. In order to accept a checking deposit as payment, at par, I only need to decide that the bank is solvent. But precise valuation of bank stock requires a more careful valuation of bank assets, and these asset values might be uncertain during a crisis. Perry Mehrling explains the new role of the government during crises as a dealer of last resort for assets of financial intermediaries, including MMFs handling massive redemptions.

    1. Anwer,

      Presumably John does not want a central bank as "bank stock buyer of last resort" so yes, the payment system would break down when the liquidity of bank stock becomes reduced.

    2. Anwer says “What we need are claims that remain liquid even during crises.”

      That’s exactly what full reserve banking involves!! That is, the value of what’s placed in totally safe accounts won’t drop if there is a stock market crash. And anyone with any sense will use those accounts for “payments system” purposes.

      If someone wants to use “bank stock is as usable a liquid claim”, they’re free do so. But they’re taking a risk, and no one is obliged to come to their rescue if it all goes wrong.

      The latter policy is the equivalent of me keeping nearly all my assets in non-cash forms: equity, house, car, antiques, etc. If I run short of cash as a result, that’s entirely my fault.

  15. Anwer: T-bill floating-nav fund: run proof. FDIC shld require that all regulated banks provide them with checking and payment facilities.

    1. Prof. Cochrane makes a similar proposal in this post:

      I agree that it would keep working during a crisis. Note that he mentions the significance of "safe assets" in that post, but his usage of the term is restricted to risk-free assets i.e. US government debt.

  16. I agree that the last two criticisms are invalid, however there is truth in the first one in that your proposal is so radical that its impossible to implement, so better to at least ask for more equity.

  17. I agree with your underlying proposition, but I think the two ideas, elimination of run prone liabilities and the provision of money-like assets, could be merged into a single and easier change.
    You say that banks should be funded by equity or by some long-term debt. This certainly eliminates run prone liabilities. Equity is simple solution, but long-term debt can be made to work too and provide a new and more soundly based form of money.
    If each bank issued a single tranche of perpetual floating rate debt and the interest rate on that were set by an online auction, its value would remain at par. The rate would immediately reflect the credit of each bank. It would not be run prone as it would have no redemption at all or only after a long notice period. Under Basel III pressure, these types of bank deposits are already becoming popular as evergreens or notice period deposits. The rate would be true continuous floating rate unlike current periodic reset floating rates. The interest rate or margin would be capped so that, in a crisis, the liquidity risk would be widely dispersed among holders until that bank’s position was resolved.
    This has the advantage of:
    • No double taxation problem with equity funding.
    • Being at always par, it would fit the payments system easily in a similar way that demand deposits are settled, or it could be sold on the online trading system, for a demand deposit for a few seconds to allow the current payment system to operate. If however, other variable price securities (e.g. Apple stock) are used by the payer, they will need access to an always open and fast market maker to trade the Apple stock for the security that the payee wants (say Brazilian bonds). The market maker would need to provide both spot and future prices for future payment obligations (which are very common in trade) and then, problematically, the parties would both to need to have credit limits.
    • The rate could replace LIBOR (compromised) as the basis for floating rate term bank loans
    • This deposit is a senior long-term claim on a bank, with a senior claim on a higher amount of long-term loans, with a senior claim on an even higher amount of long-term real assets. It is matched (after the junior buffers) to the value of the long-term real assets. Long-term real assets are the largest pool of assets globally. Short-term real assets (trade credit and inventory) are a much smaller pool and are the only sound basis for short-term debt.
    • Banks then become managers and aggregators of loans made against real term assets. Credit is their main risk, against which their equity is a sound hedge. They no longer need to engage in MT (as Maturity Transformers). That said, since central banks granted Lender of Last Resort backing, they have actually been MT (Maturity Transferors). The shadow banks just engaged in MT (Maturity Take no notice of it). As you rightly say, notwithstanding how many times, like an article of faith, the ability to transform maturity has been claimed, maturity risk cannot be transformed it can only be sliced and diced pooled and tranched. As physicists will attest, time (maturity) is a rather immutable property of the universe and all the accountants and regulators, much as they might pretend otherwise, cannot change that.
    • The funds raised do not go to the Treasury who would use it to fund more government spending. To make a globally applicable solution, the solution needs to be unreliant of huge government debt - not all counties have that and of those that do, not all of their debt is reliable
    • It is not fiat money as it can be exchanged (and eliminated) for the extinguishment of loans against a senior claim to real assets – much larger, and more diverse, asset base than gold.

    1. Mark,

      Bank borrows for a perpetual amount of time to lend for a fixed amount of time. If term premium holds, why would't the banks cost of funding be higher than it's return on investment?

  18. Mark,

    "If each bank issued a single tranche of perpetual floating rate debt and the interest rate on that were set by an online auction, its value would remain at par."

    Who would be the buyers for such debt? Presumably banks would not be given special status as too big to fail and would not be permitted to collude together, and so who would put their money (principle) at risk for perpetuity?

  19. If you consider present options, you can have money on demand or for a fixed and illiquid term. once the liquidity of the auctioned deposit is well established you would get a return in excess of demand deposits which these days pay next to nothing as the funds have to be deposited in High Quality Liquid Assets. this deposit would give the bank stable funding which could be invested in real earning assets and therefore pay a decent rate. From the depositors perspective they take little price risk (only if the rate hits the documented maximum as it might in crisis) unlike a conventional term deposit. I think that will lead to a lower rate than term premium ordinarily suggests. I know that seems odd but it comes back to the distinction between money and other assets as per Mehrling's "hierarchy of money". As he points out assets which qualify as money do get a special status and can have yields below short term Treasury assets. The other great example is the perpetual zero coupon bonds we all carry in our pockets and value so highly at par - cash solely because it is so liquid. All that said, I think these deposits would need a redemption option, certainly initially, to give confidence in case trading did fail for some reason. This would be an option to redeem after giving notice of initially 31 days to satisfy LCR and later this could be extended to 12+ months to satisfy NSFR. Upon giving notice, the deposit would convert into a conventional deposit redeemable after 31 days or 12+ months from the notice date paying a lower rate as it no longer is as stable funding for the bank. This structure is becoming very common as a floating rate evergreen or notice period deposit. All I am suggesting be added is the web based trading of these. The redemption option described is normally "out of the money" and irrational to exercise provided trading is still operative. Another way to look at instrument is as one where the return on capital is decided by a continuous auction, in a sense, a voting mechanism between existing and potential holders. Debt normally has fixed or formulaic returns and equity a division of the remainder set as dividends by the BoD. This is yet another mechanism. In the late 1980's banks, did sell a lot of perpetual debt for a while and it was even subordinated. They later tanked and banks repurchased them at a discount. These deposits however would have a senior claim on income and capital if redeemed. They would also be eligible for FDIC insurance, though I don't think that is necessary for them to be viable.

  20. Apologies my statement "These deposits however would have a senior claim on income and capital if redeemed" is not quite correct.

    It should have read "These deposits however would have a senior claim on the bank." as interest payments and any redemption payments are not dependent on there being profits and income as for dividends.

  21. My letter to the editor at the Economist in response to their article. I was advised by them to put it here.

    Truly narrow-minded

    SIR – your Free exchange on making finance safer (“Narrow-minded, 7th June) was aptly named. You consider John Ochrane’s “21st century” version of the Chicago plan, which involves levying a tax on bank debt to counter the negative externalities of leverage. However, you fail to put his proposals in the context of recent regulatory changes à la Basel III (or its risk-weighted shortcomings) - even though this is likely to be the biggest “implementation” problem with Mr Cochrane’s idea.

    Why? As Basel III capital requirements push banks into holding more of their assets in the (perceived) safety of sovereign-debt (itself an implicit asset tax); the Pigouvian tax on debt-liabilities would add to the burden on bank profitability, and undermine their efforts to build capital. Furthermore, even if it successful in herding banks towards more equity capital, this would undoubtedly raise funding costs. Banks might compensate for this by increased risk-taking elsewhere to remain attractive to shareholders.

    A truly “radical proposal” would be to reconsider the bias in Basel capital regulations for banks’ investment in sovereign debt through zero capital charges. History, and more recently Greece’s debt restructuring, has shown that such securities are far from risk-free. Such a re-think would force banks to actually raise capital – rather than simply adjust balance sheets to exploit this (flawed) regulatory loophole. Meanwhile, additional interventions that skew an already biased regulatory framework are unlikely to make the financial system safer.

  22. I came to this blog from the Economist article, and have been reading the comments with interest; although I confess to not yet reading Professor Cochrane's original article, something I hope to remedy soon.

    I have a simple, potentially stupid question. The implicit assumption in the Economist article seemed to be that we would be shifting from the current system to this new system over some short timeframe. Indeed, many of the comments on this blog talk about the potential shortfalls and benefits of one system vs another.

    What seems obvious to me is that, with all due respect both to Professor Cochrane and the many experts providing comments here, is that we probably don't know enough about how the new system would operate to make systemic change. We probably know a lot more than the progenitors of the current system back several centuries ago (my thanks to the poster above for the link to his? article which I'll read soon), but I suspect not enough to make a solid case for systemic change; particularly when those who benefit from the current system have such a vested interest in highlighting every potential flaw, real or imagined.

    My question, then, is whether the best approach forward might not be systemic change, but piecemeal change? That is, set up one bank, perhaps not even a large one, along the lines that Professor Cochrane suggests, and see how it evolves its way through some of the problems people have enunciated in posts above. I appreciate that banks under the current system enjoy certain subsidies, but if these can be quantified, and the innovative bank given some appropriate government support to obviate the subsidy other banks obtain (support which fits within the overall framework Cochrane suggests, not support which turns into a crutch which can never be removed) such that it competes on a level playing field with them, then this provides the basis for a fair experiment to test Professor Cochrane's ideas in the field. If it works with one bank, then extend it to two and so on. At the same time, wind back the subsidies by making incremental changes to the existing institutional system which provides subsidies to the current generation of banks.

    I realise there are a host of problems in the suggestion above, from quantifying the subsidy to choosing the first bank (although we auction the rights to government subsidy streams all the time in unrelated areas such as public transport). I realise also that you don't get many of the benefits Professor Cochrane highlights until you have systemic change. However, I can't help but wonder if this idea might get more legs in a policy sense if Professor Cochrane and those supporting it turned their attention to the question of how we might create one viable "narrow bank" operating as Professor Cochrane suggests within the constraints of the current system, and remove barriers preventing the evolving innovation from spreading, rather than the question of how a new, fully-evolved system might operate. This much smaller question, although of less interest to economists who love to debate about how competing systems might operate, is probably a much easier sell to policymakers, and a much harder target for detractors to hit.

    I would be very interested in any comments you, collectively, have. Hopefully I don't find out when I get to the relevant articles that the first comment ought to be "we already thought of that, read paragraph XX."

    Nick Wills-Johnson

  23. I came to this post after reading the Economist article, and though I admit to not yet reading the underlying academic papers (in the interests of saving other reader's collective time in deciding whether it it worthwhile reading the rest of this comment), I have a simple, and perhaps stupid question.

    The Economist article, and most of the posts above, focus on the current system versus the system Professor Cochrane proposes, and the Economist article in particular seems to have an underlying assumption that there would be fairly rapid change from one system to another, thus focussing on the problems associated with such a change.

    I realise that focussing upon, and arguing about how different systems might operate is interesting, and useful in the longer term. However, if the aim is to effect change, I wonder if a different question might be asked. Instead of asking "how would this new system work?", ask "how might one of these 'narrow banks' operate within the current system?".

    That is, focus upon the kind of case that would need to be made to policymakers to get just one of these new banks up and running, and competing on a level playing field with existing banks operating within, and taking advantage of, the current institutional framework. This would require some support from government, because of the subsidies noted by Professor Cochrane, and one would need to think very carefully about what those subsidies would look like, so that they could be scaled if necessary when the experiment is tried with the next new bank, but such that they peter out as a shift in systems occur so that they do not become a lasting feature. They would also have to be something that one bank gets, rather than something (like the Pigovian tax suggested by Professor Cochrane) that is imposed on everyone else.

    Focussing on one bank, though not without its problems (how to create the subsidies, which bank to choose - and how - and whether the benefits of Professor Cochrane's ideas can indeed be seen if only one bank switches), is a much easier sell to policymakers than systemic change, and is a much smaller target for detractors. Additionally, as is apparent from the debate above, and with due respect to the erudition of Professor Cochrane and the various posters above, it is abundantly clear that we do not know exactly how this new system would operate, and what problems it would throw up. In such a situation, an evolutionary approach seems more appropriate.

    I would be very interested to read any comments on how one might make the "just one bank changes" case; hopefully I am not just floating an idea which has already addressed in one of the many papers above which, in my haste, I haven't read yet.

    Nick Wills-Johnson

    1. Nick Wills,

      I don't think you could do it on a one bank basis. At the heart of Cochrane's argument lies the elimination of the central bank as lender of last resort. The problem is that there is more than one central bank (U. S., Canada, Europe, Japan, England, etc.).

      Eliminate the central bank of the U. S., and short term borrowers would simply move on to a different central bank.

  24. Yes I'm suggesting Trills as a unit of account for the fractional reserve system, while narrow banks would use dollars backed by the perpetual floating-rate securities described in some of Prof. Cochrane's blog posts. Both would be media of exchange.

    During an economic contraction, there would be a rebate for debtors instead of a required coupon payment. And if long-run economic prospects have deteriorated, the principal value of debt would also diminish with the decline in Trill valuations. This is arguably a more efficient distribution of macroeconomic risk than we have in the current system. For arguments in favour of debt forgiveness linked to economic conditions, I refer you to the work of Atif Mian and Amir Sufi.

    There is a bad feedback loop in the current system. I think economists describe it as "debt deflation", where the real value of debts increase as the economy contracts, leading to even greater contraction. Mian and Sufi advocate greater risk-sharing with more complex debt contracts (home loans, student loans) whereas market monetarists want to adjust the general price level to attain the same goal. Kevin Sheedy has produced analyses of the latter policy. Of these two approaches, I prefer contractual innovation (by changing the unit of account) while keeping prices stable.

    1. Anwer,

      "During an economic contraction, there would be a rebate for debtors instead of a required coupon payment. And if long-run economic prospects have deteriorated, the principal value of debt would also diminish with the decline in Trill valuations."

      The way you initially described it both assets (medium of exchange) and liabilities (debt) are denominated in Trills, and so while the value of debt falls with falling GDP so does the value of the medium of exchange used to discharge the debt - you get nowhere.

      The way you are now describing it now, debt contracts would have floating rates that are adjusted based upon a macroeconomic condition (nominal GDP level). In the extreme case, the flow of money from debtor to creditor reverses.

      That creates a really perverse incentive to borrow money and bury it in the ground. I borrow money at a an interest rate, I take borrowed money and bury it in the ground, I wait for a recession and claim that I should be paid by my creditor because GDP has fallen.

      "There is a bad feedback loop in the current system. I think economists describe it as debt deflation, where the real value of debts increase as the economy contracts, leading to even greater contraction."

      You are correct on debt deflation. But a couple of things about that:

      1. Debt deflation does not affect the federal government - tax revenue used to service government debt is independent of the price level / inflation rate

      2. Private companies can resort to equity financing when the real interest rate becomes onerous.

      3. Private individuals / companies realize their cost of debt service on an after tax basis meaning that interest payments are tax deductible

      "Of these two approaches, I prefer contractual innovation (by changing the unit of account) while keeping prices stable."

      My own preference is for contractual innovation as well, but I prefer government equity. Federal government sells equity claims against future tax revenue with a potential rate of return that meets or exceeds the real interest rate. Potential borrower hedges against deflation by buying government equity. His / her pretax real debt cost may be onerous but his / her after-tax real debt cost can be whatever the federal government decides it should be (preferably zero).

    2. Anwer,

      We are not really that far off from each other. I think tax reduction securities (government equity) would be easier to implement on a legal basis. To implement what you are describing, all existing debt contracts would need to be rewritten or renewed to incorporate an adjustment based upon the level of GDP. I think that would be a legal nightmare of epic proportions.

      Implementing tax reduction securities would be implemented by simply having Congress (Legislative Branch) granting the Treasury Secretary (Executive Branch) authority to sell equity. That authority could be granted explicitly or implicitly. Implicit authorization would involve the Congress setting the level of spending and the level of taxation and then refusing to borrow to make up the difference.

      Implicit authorization would be politically easier. Congress promises it's constituents to not raise taxes and Congress promises it's constituents to not cut spending. Congress also refuses to authorize a debt increase. Department of Treasury is left to proceed at it's own discretion.

    3. I think that eventually there will be more agreement on the need for contractual innovation. In his paper Kevin Sheedy notes that there are missing financial markets and people do not use optimal debt contracts, and it then falls on monetary authorities to stabilize the real value of debts by adjusting the price level. At some point in history, that will no longer make sense.

    4. Anwer,

      " then falls on monetary authorities to stabilize the real value of debts by adjusting the price level. At some point in history, that will no longer make sense..."

      Agreed. A really, really big increase in productivity would do that. Something on the order or direct energy to matter conversion (nuclear reaction in reverse) or teleportation (see Star Trek) would push the legal protections of nominal debt to their breaking point.

  25. Anwer,

    That DeAnglo and Stulz paper is nonsense. All they’re saying is that commercial banks perform a useful function in that they supply the economy with a form of money or “liquid financial claims” to use the pseudo technical language at which the authors excel.

    I.e. what they miss is that advocates of full reserve are well aware that in a full reserve scenario, private money production is curtailed or banned altogether. Instead, what full reserve advocates propose is that central banks supply the nation with it’s money.

    Ironically and thanks to QE, the proportion of money in the US which is made up of Fed produced money has shot up over the last three years or so, and the sky hasn’t fallen in far as I know.

  26. Ralph, how would you apply this idea in the Eurozone? Without a fiscal union, they really need money produced privately by banks. I would have those banks each issue Trills of their home countries (where most of their loans are made). This would also prevent the capital flight problem they have now with banking based on Euros, when a member country experiences an economic downturn.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.