avatar

Capital and Cash Reserves

by George Selgin June 11th, 2011 1:14 pm

I promise to make this my last post for a while concerning the matter of 100-percent versus fractional-reserve banking. However, in addressing some comments on my recent posts it occurred to me that some very serious misunderstanding is at play concerning the difference between a bank's capital and its cash reserves. The distinction between these is important, because in an important sense, and particularly with respect to comparisons of fractional and 100-percent reserve institutions, the two are substitutes.

Capital serves as a cushion for the purpose of absorbing the effects of adverse shocks to a bank's asset portfolio, so that the bank's creditors won't suffer losses in connection with such shocks, except in rare instances. This in turn allows the bank's IOUs to continue to hold a fixed value in terms of the underlying reserve asset despite changes in the market value of the bank's non-reserve assets. The larger a fractional-reserve bank's capital cushion, the greater the adverse shocks it can handle without becoming insolvent, that is, without finding that it has run out of equity. In a free market, an insolvent bank ceases to be a going concern: it must either find a buyer or go into liquidation.

A 100-percent reserve bank hardly needs any equity capital, because the need for such capital arises only when the bank faces shocks that can reduce the value of its assets. By holding only cash reserves, such a bank eliminates most of the portfolio risks that fractional reserve banks encounter, making equity redundant. In fact past 100-percent institutions had very little if any equity capital.

In a world without insurance, on the other hand, a risky fractional-reserve bank can only attract deposits by putting its owners' capital at stake. Other things equal, the more capital a bank has, the more likely it will succeed in attracting risk-averse depositors. Fractional reserve banks can also attract such depositors by holding low-risk assets, such as U.S. government securities. In the past, banks used both sorts of strategies to gain market share, often taking out advertisements in newspapers in which they supplied pertinent balance-sheet details. (Although it deals only with the U.S. the best book to read about this, and how it all changed after deposit insurance was introduced, is Jim Grant's terrific Money of the Mind.)

Of course, not all banks catered to depositors whose primary interest was safety: there was a market for riskier bank deposits also. But, despite what apologists for central banking and deposit insurance claim, it was not especially difficult to tell safer banks from less safe ones. The problem in places like the U.S. before 1934 and England before 1826 was not so much one of distinguishing relatively safe banks from relatively risky ones, but one of legal restrictions that prevented well-capitalized banks from emerging in many communities. In the U.S. the restrictions consisted of laws preventing branch banking; in England they consisted of laws preventing English banks other than the Bank of England from having more than six partners. (In 1826 other public or "joint-stock" banks were permitted, but only if they did not operate in the greater London area--itself a major limitation; while in 1833 other joint-stock banks were admitted into the London area, but only provided they gave up the right to issue banknotes.) These regulations limited the capitalization of U.S. and English banks while at the same time limiting those banks' opportunities for financial diversification--a recipe for failure. In both instances the regulations were products of politicians' catering to rent-seeking behavior on the part of banking industy insiders. Yet the resulting, unusual frequency of bank failures and substantial creditor losses stemming from such failures helped to sustain the belief that fractional reserve banking could only be made safe by means of further government intervention.

Where laws did not prevent banks from diversifying their balance sheets, especially by establishing widespread branch networks, or from securing large amounts of capital by "going public" (or, in the case of some Scottish and most Canadian banks, by making shareholders liable beyond the par value of their shares, which from creditors' point of view is equivalent to having more capital), bank failures have been relatively less common, and losses to creditors stemming from occasional failures that did occur have been relatively minor. Indeed, even such a spectacular failure as that of Scotland's Ayr Bank did not ultimately prevent the bank's creditors from being paid in full, without need for any sort of bailout.

What has all this to do with the question of 100-percent versus fractional reserves? First of all, the 100-percenters are in the habit of posing a simple choice between "safe" 100-percent reserve banks and "unsafe" fractional reserve banks. But, even putting aside the obvious fact that banks keeping the same reserve fraction may be more-or-less safe depending on the sort of non-cash assets they hold, the comparison overlooks the role of capital in limiting the risk borne by fractional-reserve deposit and noteholders, with banks that are both well-diversified and well-capitalized being capable of exposing their creditors to very little risk even despite maintaining slim reserves. More to the point, an entire range of risk-return options may exist even for banks holding approximately equal reserve ratios. Economists are generally suspicious, and rightly so, of "corner solutions." Yet 100-percenters have latched on to just such a solution in suggesting that it must be superior to the whole array of possible fractional-reserve alternatives.

Second, in suggesting that fractional-reserve banks exist only because their customers don't realize that they are engaged in lending, 100-percenters beg the question: if so, why do fractional-reserve banks bother (or why did they bother, in the good-old pre-insurance days) to hold capital, and often plenty of it? As I explained above, the reason for them having done so is precisely because they needed to assure their creditors that, although their money was at risk, the risk was limited, and perhaps made very small indeed, by the bank's capital cushion. To accumulate such capital cushions, let alone advertise them, would, of course, have made no sense at all to bankers who were determined to convince their customers that their money was all kept in a vault. On the contrary: it could only serve to give the game away.

24 Responses to “Capital and Cash Reserves”

  1. avatar Warren says:

    Sir,

    I have read a lot on free-banking but not nearly as much as I wish to so I have some technical questions on this issue. I asked before but received no response. If it's rude to ask questions over again please let me know.

    The 2% or so reserves that the Scottish banks carried were like their buffer against having the shareholders have to sell off personal property to redeem notes, yes? So a bank need not go out of business if this reserve is used up as long as enough personal property could be sold to make up the difference?

    What would cause a bank to call it's loans? If there was a problem and a bank had to call some or all of it's loans how did that help as it's loans were made in it's own currency, the currency that in this hypothetical example, no one wants anymore?

    How much gold was deposited with these banks? Did the banks add this gold to their reserves? And if the banks did not use it as reserves what did they do with it and/or where did it go?

    Why was branch banking in the U.S. so restricted? And did people get around this by opening multiple but allied banks? Because that's what I would do: The banks might all have different names but their notes would be surprisingly similar and redeemable at any allied bank.

    And a new one where I encapsulate what I think I've learned: Say a bank, a very small and strange bank; makes only one loan of 100 in it's own currency. This being the only amount of currency it creates. The bank will receive 120 at the end of the loan's term meaning that those other 20 come from somewhere else. So if it's currency from other banks this means that when the currency is cleared this tiny bank will be +20 against one or more other banks and will then receive that from them in the form of gold?

    So the idea is make loans to generate interest but not so many loans that you end up in the negative when clearing and have to pay out reserves? And this pair of incentives is working over the entire industry and keeping the greedy in check? So that when deposit insurance is provided the second incentive disappears? And the greedy are left unchecked?

    But if all banks are well run doesn't the money just get shuttled around? Bank A is up this week but down next week enough to cancel it's week one gains and this happens to all the banks B through Z.

    Now if there are less competently run banks then their reserves will be where the better banks draw their profits right? Like a poker game needs weak players to provide the better players with a source of profit. The analogy is not exact, of course, as a poker game is a zero-sum affair.

    Of course I recognize that you are in no way obligated to answer these questions. I ask because as a free-marketeer I'm a de facto free-banking supporter as I support peaceful individual's rights to interact how they wish. Thus people can use gold exclusively if they wish or engage in FRB along free-banking lines if they wish. And if I'm going to explain the concept to people I need as much info as I can get.

    Thank you.

    • avatar George Selgin says:

      Warren, I can't answer all of your questions, as really a money and banking course is called for to answer them in full, but I will at least beging to answer a few. First, bank reserves aren't there to protect a bank's shareholders. It is the shareholders who are residual claimants--that is, who must bear losses before any of its creditors must do so. Reserves are held solely for the sake of meeting a bank's daily liquidity needs for interbank settlement and certain cash withdrawals. A bank can be perfectly solvent yet illiquid if it is short of reserves. Conversely, a bank can go broke without being illiquid, much less running out of reserves. (You confuse a bank's liquidity with it's solvency in making several of the points you make.) Lastly, banks that "call" their loans (or simply don't replace maturing loans) do increase their liquidity, by reducing their outstanding liabilites if not by actually gaining basic money reserves.

  2. avatar Mathieu says:

    Pr Selgin,

    "In a free market, an insolvent bank ceases to be a going concern: it must either find a buyer or go into liquidation."

    Or undergo reorganization? Is there a statement against Chapter 11 bankruptcies in the financial sector in there or am I looking too deep into it?

  3. avatar George Selgin says:

    You raise an important point, Mathieu. Bankruptcy procedures analogous to chapter 11, where an insolvent firm is allowed to remain in business pending a restructuring of its debts, are an option in certain countries, but are not universally available. Whether such procedures would be part of a free-banking arrangement is an open question. My inclination is to believe that they could be, provided that they are contractually based, hence entirely voluntary. However I am no expert on this area, and experts disagree. Hence my (unsuccessful) attempt to skirt the issue!

    • avatar Mike Sproul says:

      Don't forget the common alternative of suspending convertibility. A bank whose assets (including capital) are 1% too little to buy back the money it has issued could be liquidated at a 1% loss to the holders of its money, or it could suspend convertibility, in which case its money would trade at a 1% discount. The loss to money-holders is the same in either case, but suspension could easily be less disruptive than liquidation. That's why suspensions have been so common, when the government didn't intrude and outlaw them.

      The biggest problem this creates for economists is that most of them are blind to the difference between "inconvertible" and "unbacked". They look at a currency like the US dollar and see that ONE form of convertibility (dollars into gold) has been suspended, and they then pronounce the dollar "unbacked" (or "fiat money"), as if the Fed and the government have no assets.

  4. avatar George Selgin says:

    Mike, you and I must agree to disagree on the matter of suspension. As I observed in reply to your comment on a previous post, suspension is not a sensible option for insolvent banks. It is neither in creditors' nor in owners' interest to resort to it, for it means (when contracts are "incentive compatible") paying more interest than usual on suspended liabilities, which, other things equal, further reduces a suspended bank's net worth. It can make sense for a bank that's merely illiquid, for such a bank can afford to endure a loss on its income account until it is able to restore its liquidity and get back to business as usual.

    • avatar Mike Sproul says:

      Once a bank has suspended, its money becomes an equity claim against the bank, so the solvent/insolvent distinction loses its meaning. Also, there are many different forms of convertibility, and suspending only one kind (dollars into silver) leaves all the other forms intact (e.g., dollars into tax payments). Now, the suspension of ALL forms of convertibility is an effective total default by the bank, so we have to keep in mind that even when one form of convertibility has been suspended, there will always be other forms of convertibility still in effect.

      There is of course an incentive problem when money becomes a pure equity claim against the issuing bank, but various monitoring arrangements can get around that. Making money into an equity claim against the issuer can have practical advantages, which is one reason it has been so common historically.

      • avatar George Selgin says:

        Mike, I don't know what the basis is for your frst sentence. I can only telll you that this wasn't the case w.r.t either the option-clause terms in Scottish banking or actual U.S. suspensions of payment. It seems to me that you have in mind some suspension arrangement of your own devizing, rather than the sort that draws on historical experience. What you have described does happen in some bankruptcy proceedings, which presumably are what you have in mind by saying that it is "common." But bankruptcy means failure, not temporary suspension. Moreover it isn't the way most bank failures are handled, or have been handled in the past.

        Your talk of suspending of convertibility of bank IOUs into metal reserve media being only a "kind" of suspension with "other forms" ("dollars into tax payments") is frankly off the wall. Bank money may be useful or not in tax payments, but that has nothing to do with convertibility as normally understood. A bank's notes are convertible or not depending on whether they can be converted into the system reserve media. There is no other "kind" of convertibulity that's relevant.

        • avatar Mike Sproul says:

          George:

          When I say that money becomes an equity claim I mean that it becomes a claim to whatever assets the firm has. John Cochrane made the same general point in "Money as Stock". It doesn't mean that money holders are legally changed into the firm's stockholders. So if the firm's total assets including capital add up to 99 oz, while there are $100 outstanding, each dollar is worth .99 oz. If (assets+capital) rises to 100 or above, the money goes back to being priced at 1 oz., just like a bond or other fixed claim.

          It wouldn't take long to rattle off a list of moneys that got suspended for extended periods without the issuing bank being liquidated. You probably know that list at least as well as I do.

          There is nothing off the wall about multiple forms of convertibility. You are just trying to draw clear lines where nature has not put any, declaring one currency convertible and another inconvertible, when the actual differences between the two might not matter to anyone. A banker might normally convert his dollars to 1 oz. of silver, but in hard times he might offer his customers 1 oz. worth of groceries instead. A bank that has 100 oz of silver and 200 oz. worth of loans might never be asked to hand out a single oz. of silver, but might be constantly receiving his own dollars in payment of his loans at the rate of 1 oz./$. After a few years of this he might declare that he will offer silver convertibility only one day per century, but because he is constantly making loans and receiving payments, customers won't care and still value his notes just the same. Convertibility can clearly take many forms without customers caring much. The one thing that all customers do care about is backing, and that's exactly the thing that the quantity theory says is irrelevant to the value of money.

  5. avatar RickDiMare says:

    George, it's not an "all or nothing" thing, so please keep talking about full-reserve vs. fractional-reserve all you want.

    In the U.S. we have 2 kinds of currencies that should be allowed to freely circulate with each other, even among a prospective system of free banks, and even if it means one currency will eventually drive out the other, so in other words, all banks must internally learn to deal with both currencies in a judicially supervised way.

    For lack of better terms, I call one currency "sovereign," which is Treasury-Direct money that banks have no right to subject to fractional reserve practices, and the other I would call "non-sovereign fiduciary media" (NSFM) which should continue to be subject to fractional reserve practices. But, it gets more complicated -- because my research shows that NSFM is part of some kind of negative income tax system which, for better or worse, the U.S. seems to be using in an attempt to fight poverty, so eventually free banking will need to deal with this can of worms, too.

  6. avatar BillWoolsey says:

    George:

    What do the owners of an insolvent bank have to lose by suspending?

    I think it is likely that the normal situation will--depositors fear insolvency, run, suspension, and the management claims that everything is just fine and the whole problem was a bunch of paranoid depositors.

    Now, if it really was just a liquidity issue, so that the reserve drain just happened because the depositors were going about their ordinary business, then presumably the depositors wouldn't complain.

    But in the insolvency (or possible insolvency scenario,) the depositors will be unhappy about the suspension.

    On the other hand, it might be better than continuing to strip the bank of any half-way liquid assets in a run. But, as you say, the depositors will want to see the bank move towards reorganization or liquidation. Letting the insolvent bank run as a kind of closed end mutual fund is not in their interests.

    It seems to me that some kind of third party review at the request of a depositor would be normal for a bank that suspends. If they are insolvent, then they don't resume payments, but rather move into bankruptcy.

    While closing down one bank in an ocean of banks wouldn't be a problem, a system of large, branched banking does create a greater threat of breaking down the payments system if a significant portion of the banking system becomes insolvent at once.

    I think Kevin Dowd's analogy of a hospital going bankrupt was appropriate. You don't put all the patients out to the curb, but you do take steps to reorganize or wind things up.

    Some kind of fractional debt-equity swap seems like the way to go to me. With hand-to-hand currency (which I think gets too much attention) that is hard, but not with deposits. The bank suspends. Third party check. Everyone's deposit is written down by some fraction, say 25% and they get shares in the bank. Institution is solvent. If the "liquidity" problem was about solvency, then problem solved and the suspension is over. If there is a real liquidity problem, then the suspension continues, but the institution is solvent.

    • avatar George Selgin says:

      Bill, the argument is that properly designed suspension contracts will include interest penalties that make suspension unworthwhile for an insolvent bank. That's what Gorton means by "incentive compatible" contracts. If the contracts would allow banks to profit by suspending when it isn't in depositors interest, then the depositors won't agree to the suspension option in the first place. (My argument doesn't refer to the possibility of banks suspending unilaterally, without customers' prior consent.)

  7. avatar David Stinson says:

    My comment on your previous post languished for a long (long) time in moderation so, forgive the repetition, but I'll ask a question that I asked previously: given the role of bank capital in protecting depositors and thus in addressing depositor concerns about safety of deposits, is it fair to argue that the level of bank A's reserves actually has more to do with satisfying other banks that they can expect to get paid by bank A in base money as part of the net settlement process and that therefore they can accept deposits of bank A's notes and cheques? In other words, it is necessary to permit convertibility between the notes issued by various free banks.

    Has it ever been argued that provision of safety deposit boxes already amounts to providing customers with the option of 100% reserve banking given that the latter really amounts just to safe storage?

    • avatar George Selgin says:

      Generally speaking, David, banks don't "show off" their reserves the way they might show off their capital. All other banks care about is whether their rivals actually settle or not. They don't give a bank brownie points for holding more reserves than are needed for daily requirements, and especially not when a good market for interbank loans is present. In that case, banks can readily cover shortfalls so long as they have enough decent collateral. Very slim reserves are often all that banks need in such cases.

      And yes, all discussion of free banking assume banknotes notes are convertible. The notion of an inconvertible note issued by a "free" bank is hard to fathom: who can have given the bank the right to dishonor it's own IOUs, unless the government has interfered?

      Finally, yes, the point about safety deposit boxes has been raised. It hasn't served to convert hard-core 100-percenters. But then again, nothing has--and nothing ever will. The point of making such arguments is to prevent the merely unwary from joining the ranks of the incorrigible.

      • avatar David Stinson says:

        Thanks George for your reply.

        I was thinking that some banks might refuse to accept notes issued by other banks if they had insufficient confidence in the first bank's ability to meet their obligations. Obviously, in that case, merchants' willingness to accept notes of the bank in question would also be affected.

        But, if I understand you correctly, the confidence of other banks in a particular bank's notes is determined in the first instance by access to the interbank market which, in turn, depends on solvency. Consequently, the confidence of both depositors and other banks depends more on solvency than on the level of reserves. Which is kind of interesting given that the soundness of the free banks' money, at least in the popular imagination, would be based on its convertibility into the base money (whatever that might be). Excess money demand would be the great enemy of such a system because of its impact on the solvency of all banks. As long as there was no central bank, monetary disequilibrium would presumably not be a problem. However, if one introduces a central bank or unhelpful regulation and thus the possibility of monetary disequilibrium, banks would presumably need more reserves to in effect protect them from the central-bank-induced insolvency (in the absence I guess of bailouts).

  8. avatar StoneGlasgow says:

    1. When modern banks issue new loans, they create new currency to pay for the home or car that is the catalyst for the loan. They are permitted to issue new loans (new currency) in keeping with the current reserve requirements. Would it not be possible to have a fractional reserve bank with no reserve requirements? What is the purpose of the reserve requirement other than to restrict the lending of smaller banks, and allow more lending by larger, richer banks? Would it not be theoretically possible for a zero-percent reserve bank to operate without any problems at all?

    2. If a 100% bank invests its assets, it is risking default in the same way that a fractional bank risks default by investing the deposits of currency held in its checking and savings accounts. If both banks avoid investing the money held in checking/savings accounts, both banks would avoid any possibility of default.

    3. The 100% bank, because it would not create new money, would have to charge a fee for its checking and savings account services, or invest its deposits to be profitable, but the zero reserve bank would have no such requirements; it would make money when it loans new currency into existence, and receives interest on that loan.

    • avatar George Selgin says:

      Let's be perfectly clear, StoneG.: strictly speaking there are no reserve "requirements" in a free banking system: banks hold whatever reserves they deem adequate to cover their needs, which include (1) the need to meet occasional over-the-counter requests for reserve ("base") money and (2) the need to pay routine interbank settlement dues. These needs make it impossible for any bank to survive for very long with zero reserves: it could perhaps manage for a while by talking out last-minute "overnight" loans from other banks, but eventually it would find this strategy unprofitable owing to the relatively high interest charges it would incur.

      Second, I don't understand what you mean in speaking of a 100-percent bank "investing" its assets. Such a bank holds only cash; it has no investment opportunities in the usually understood sense. The money it takes in is for storage purposes only.

      Finally, you mustn't imagine that a free fractional reserve bank just loans new currency into existence without first having to attract deposits of base money. If a free bank finds it must keep $2 for every $100 of its IOUs outstanding, that means that for every new $100 net increase in deposit it manages to secure, it can lend $98. Not the toughest job around, but not a simple matter of "printing up" loans, either. (Just to be clear: if someone suggests that a bank that seeks to maintain a .02 reserve ratio can, upon receiving $100 in reserves, turn around and lend $5000, that person doesn't know their banking theory. Again, the banking business isn't hard, but it isn't quite this easy!)

  9. avatar StoneGlasgow says:

    I agree that in a free-banking system, banks would choose their own reserve requirements. I was speaking about our current system, which demands a 10% reserve.

    Is it your view that the only reason US dollars hold value today is that government demands tax payments in dollars, and has created legal tender laws?

Leave a Reply

You must be logged in to post a comment. Login now.