Saturday, March 10, 2012

Graeber Cycles and the Wicksellian Judgment Day

So it's halfway through the semester, and I'm looking over the midterms. Good news: Learning has taken place.

One of the things you hope students learn in a course like this is that money consists of three things: demand deposits (checking accounts and the like), currency and bank reserves. The first is a liability of private banks, the latter two are liabilities of the central bank. That money is always someone's liability -- a debt -- is often a hard thing for students to get their heads around, so one can end up teaching it a bit catechistically. Balance sheets, with their absolute (except for the exceptions) and seemingly arbitrary rules, can feel a bit like religious formula. On this test, the question about the definition of money was one of the few that didn't require students to think.

But when you do think about it, it's a very strange thing. What we teach as just a fact about the world, is really the product of -- or rather, a moment in -- a very specific historical evolution. We are lumping together two very different kinds of "money." Currency looks like classical money, like gold; but demand deposits do not. The most obvious difference, at least in the context of macroeconomics, is that one is exogenous (or set by policy) and the other endogenous. We paper this over by talking about reserve requirements, which allow the central bank to set "the" money supply to determine "the" interest rate. But everyone knows that reserve requirements are a dead letter and have been for decades, probably. While monetarists like Nick Rowe insist that there's something special about currency -- they have to, given the logic of their theories -- in the real world the link between the "money" issued by central banks and the "money" that matters for the economy has attenuated to imperceptible gossamer, if it hasn't been severed entirely. The best explanation for how conventional monetary policy works today is pure convention: With the supply of money entirely in the hands of private banks, policy is effective only because market participants expect it to be effective.

In other words, central banks today are like the Chinese emperor Wang Wei-Shao in the mid-1960s film Genghis Khan:
One of the film's early scenes shows the exquisitely attired emperor, calligraphy brush in hand, elegantly composing a poem. With an ethereal self-assurace born of unquestioning confidence in the divinely ordained course of worldly affairs, he explains that the poem's purpose is to express his displeasure at the Mongol barbarians who have lately been creating a disturbance on the empire's western frontier, and, by so doing, cause them to desist.  
Today expressions of intentions by leaders of the world's major central banks typically have immediate repercussions in financial markets... Central bankers' public utterances ... regularly move prices and yields in the financial markets, and these financial variables in turn affect non-financial economic activity... Indeed, a widely shared opinion today is that central bank need not actually do anything. ... 
In truth the ability of central banks to affect the evolution of prices and output ... [is] something of a mystery. ... Each [explanation of their influence] ... turns out to depend on one or another of a series of by now familiar fictions: households and firms need currency to purchase goods and services; banks can issue only reserve-bearing liabilities; no non-bank financial institutions create credit; and so on. 
... at a practical level, there is today [1999] little doubt that a country's monetary policy not only can but does largely determine the evolution of its price level..., and almost as little doubt that monetary policy exerts significant influence over ... employment and output... Circumstances change over time, however, and when they do the fictions that once described matters adequately may no longer do so. ... There may well have been a time when the might of the Chinese empire was such that the mere suggestion of willingness to use it was sufficient to make potential invaders withdraw.
What looked potential a dozen years ago is now actual, if it wasn't already then. It's impossible to tell any sensible macroeconomic story that hinges on the quantity of outside money. The shift in our language from  money, which can be measured -- that one could formulate a "quantity theory" of  -- to discussions of liquidity, still a noun but now not a tangible thing but a property that adheres in different assets to different degrees, is a key diagnostic. And liquidity is a result of the operations of the financial system, not a feature of the natural world or a dial that can be set by the central bank. In 1820 or 1960 or arguably even in 1990 you could tell a kind of monetarist story that had some purchase on reality. Not today. But, and this is my point! it's not a simple before and after story. Because, not in 1890 either.

David Graeber, in his magisterial Debt: The First 5,000 Years [1], describes a very long alternation between world economies based on commodity money and world economies based on credit money. (Graeber's idiolect is money and debt; let's use here the standard terms.) The former is anonymous, universal and disembedded, corresponds to centralized states and extensive warfare, and develops alongside those other great institutions for separating people from their social contexts, slavery and bureaucracy. [2] Credit, by contrast, is personal, particular, and unavoidably connected with specific relationships and obligations; it corresponds to decentralized, heterogeneous forms of authority. The alternations between commodity-money systems,with their transcendental, monotheistic religious-philosophical superstructures; and credit systems, with their eclectic, immanent, pantheistic superstructures, is, in my opinion, the heart of Debt. (The contrast between medieval Christianity, with its endless mediations by saints and relics and the letters of Christ's name, and modern Christianity, with just you and the unknowable Divine, is paradigmatic.) Alternations not cycles, since there is no theory of the transition; probably just as well.

For Graeber, the whole half-millenium from the 16th through the 20th centuries is a period of the dominion of money, a dominion only now -- maybe -- coming to an end. But closer to ground level, there are shorter cycles. This comes through clearly in Axel Leijonhufvud's brilliant short essay on Wicksell's monetary theory, which is really the reason this post exists. (h/t David Glasner, I think Ashwin at Macroeconomic Resilience.) Among a whole series of sharp observations, Leijonhufvud makes the point that the past two centuries have seen several swings between commodity (or quasi-commodity) money and credit money. In the early modern period, the age of Adam Smith, there really was a (commodity) money economy, you could talk about a quantity of money. But even by the time of Ricardo, who first properly formalized the corresponding theory, this was ceasing to be true (as Wicksell also recognized), and by the later 19th century it wasn't true at all. The high gold standard era (1870-1914, roughly) really used gold only for settling international balances between central banks; for private transactions, it was an age not of gold but of bank-issued paper money. [3]

If I somehow found myself teaching this course in the 18th century, I'd explain that money means gold, or gold and silver. But by the mid 19th century, if you asked people about the money in their pocket, they would have pulled out paper bills, not so unlike bills of today -- except they very likely would have been bills issued by private banks.

The new world of bank-created money worried classical economists like Wicksell, who, like later monetarists, were strongly committed to the idea that the overall price level depends on the amount of money in circulation. The problem is that in a world of pure credit money, it's impossible to base a theory of the price level on the relationship between the quantity of money and the level of output, since the former is determined by the latter. Today we've resolved this problem by just giving up on a theory of the price level, and focusing on inflation instead. But this didn't look like an acceptable solution before World War II. For economists then -- for any reasonable person -- a trajectory of the price level toward infinity was an obvious absurdity that would inevitably come to a halt, disastrously if followed too far. Whereas today, that trajectory is the precise definition of price stability, that is, stable inflation. [4] Wicksell was part of an economics profession that saw explaining the price level as a, maybe the, key task; but he had no doubt that the trend was toward an ever-diminishing role for gold, at least domestically, leaving the money supply in the hands of the banks and the price level frighteningly unmoored.

Wicksell was right. Or at least, he was right when he wrote, a bit before 1900. But a funny thing happened on the way to the world of pure credit money. Thanks to new government controls on the banking system, the trend stopped and even reversed. Leijonhufvud:
Wicksell's "Day of Judgment" when the real demand for the reserve medium would shrink to epsilon was greatly postponed by regime changes already introduced before or shortly after his death [in 1926]. In particular, governments moved to monopolize the note issue and to impose reserve requirements on banks. The control over the banking system's total liabilities that the monetary authorities gained in this way greatly reduced the potential for the kind of instability that preoccupied Wicksell. It also gave the Quantity Theory a new lease of life, particularly in the United States.
But although Judgment Day was postponed it was not cancelled. ... The monetary anchors on which 20th century central bank operating doctrines have relied are giving way. Technical developments are driving the process on two fronts. First, "smart cards" are circumventing the governmental note monopoly; the private sector is reentering the business of supplying currency. Second, banks are under increasing competitive pressure from nonbank financial institutions providing innovative payment or liquidity services; reserve requirements have become a discriminatory tax on banks that handicap them in this competition. The pressure to eliminate reserve requirements is consequently mounting. "Reserve requirements already are becoming a dead issue."
The second bolded sentence makes a nice point. Milton Friedman and his followers are regarded as opponents of regulation, supporters of laissez-faire, etc. But to the extent that the theory behind monetarism ever had any validity (or still has any validity in its present guises) it is precisely because of strict government control over credit creation. It's an irony that textbooks gloss over when they treat binding reserve requirements and the money multiplier as if they were facts of nature.

(That's more traditional textbooks. Newer textbooks replace the obsolete story that the central bank controls interest rates by setting the money supply with a new story that the central bank sets the interest rate by ... look, it just does, ok? Formally this is represented by replacing the old upward sloping LM curve with a horizontal MP (for monetary policy) line at the interest rate chosen by the central bank. The old story was artificial and, with respect to recent decades, basically wrong, but it did have the virtue of recognizing that the interest rate is determined in financial markets, and that monetary policy has to operate by changing the supply of liquidity. In the up-to-date modern version, policy might just as well operate by calligraphy.)

So, in the two centuries since Heinrich van Storch lectured the young Grand Dukes of Russia on the economic importance of "precious metals and fine jewels," capitalism has gone through two full Graeber cycles, from commodity money to credit money, back to (pseudo-)commodity money and now to credit money again. It's a process that proceeds unevenly; both the reality and the theory of money are uncomfortable hybrids of the two. But reality has advanced further toward the pure credit pole than theory has.

This time, will it make it all the way? Is Leijonhufvud right to suggest that Wicksell's Day of Judgment was deferred but not canceled, and now is at hand?

Certainly the impotence of conventional monetary policy even before the crisis is a serious omen. And it's hard to imagine a breakdown of the credit system that would force a return to commodity money, as in, say, medieval China. But on the other hand, it is not hard to imagine a reassertion of the public monopoly on means of payment. Indeed, when you think about it, it's hard to understand why this monopoly was ever abandoned. The practical advantages of smart cards over paper tokens are undeniable, but there's no reason that the cards shouldn't have been public goods just like the tokens were. (For Graeber's spiritual forefather Karl Polanyi, money, along with land and labor, was one of the core social institutions that could not be treated as commodities without destroying the social fabric.) The evolution of electronic money from credit cards looks contingent, not foreordained. Credit cards are only one of several widely-used electronic means of payment, and there's no obvious reason why they and not one of the ones issued by public entities should have been adopted universally. This is, after all, an area with extremely strong network externalities, where lock-in is likely. Indeed, in the Benjamin Friedman article quoted above, he explicitly suggests that subway cards issued by the MTA could just as easily have developed into the universal means of payment. After all, the "pay community" of subway riders in New York is even more extensive than the pay community of taxpayers, and there was probably a period in the 1990s when more people had subway cards in their wallets than had credit or debit cards. What's more, the MTA actually experimented with distributing subway card-reading machines to retailers to allow the cards to be used like, well, money. The experiment was eventually abandoned, but there doesn't seem to be any reason why it couldn't have succeeded; even today, with debit/credit cards much more widespread than two decades ago, many campuses find it advantageous to use college-issued smart cards as a kind of local currency.

These issues were touched on in the debate around interchange fees that rocked the econosphere a while back. (Why do checks settle at par -- what I pay is exactly what you get -- but debit and credit card transactions do not? Should we care?) But that discussion, while useful, could hardly resolve the deeper question: Why have we allowed means of payment to move from being a public good to a private oligopoly? In the not too distant past, if I wanted to give you some money and you wanted to give me a good or service, we didn't have to pay any third party for permission to make the trade. Now, most of the time, we do. And the payments are not small; monetarists used to (still do?) go on about the "shoe leather costs" of holding more cash as a serious reason to worry about inflation, but no sane person could imagine those costs could come close to five percent of retail spending. And that's not counting the inefficiencies. This is a private sales tax that we allow to be levied on almost every transaction,  just as distortionary and just as regressive as other sales taxes but without the benefit of, you know, funding public services. The more one thinks about it, the stranger it seems. Why, of all the expansions of public goods and collective provision won over the past 100 or 200 years, is this the one big one that has been rolled back? Why has this act of enclosure apparently not even been noticed, let alone debated? Why has the modern equivalent of minting coinage -- the prerogative of sovereigns for as long as there've been any -- been allowed to pass into the hands of Visa and MasterCard, with neoliberal regimes not just allowing but actively encouraging it?

The view of the mainstream -- which in this case stretches well to the left of Krugman and DeLong, and on the right to everyone this side of Ron Paul -- is that, whatever the causes of the crisis and however the authorities should or do respond, eventually we will return to the status quo ante. Conventional monetary policy may not be effective now, but there's no reason to doubt that it will one day get back to so being. I'm not so sure. I think people underestimate the extent to which modern central banking depended on a public monopoly on means of payment, a monopoly that arose -- was established -- historically, and has now been allowed to lapse. Christina Romer's Berkeley speech on the glorious counterrevolution in macroeconomic policy may not have been anti-perfectly timed just because it was given months before the beginning of the worst recession in 70 years, but because it marked the end of the period in which the body of theory and policy that she was extolling applied.

[1] Information wants to be free. If there's a free downloadable version of a book out there, that's what I'm going to link to. But assuming some bank has demand deposits payable to you on the liability side of its balance sheet (i.e. you've got the money), this is a book you ought to buy.

[2] In pre-modern societies a slave is simply someone all of whose kinship ties have been extinguished, and is therefore attached only to the household of his/her master. They were not necessarily low in status or living standards, and they weren't distinguished by being personally subordinated to somebody, since everyone was. And slavery certainly cannot be defined as a person being property, since, as Graeber shows, private property as we know it is simply a generalization of the law of slavery.

[3] A point also emphasized by Robert Triffin in his essential paper Myths and Realities of the So-Called Gold Standard.

[4] Which is a cautionary tale for anyone who thinks the fact that an economic process that involves some ratio diverging to infinity is by defintion unsustainable. Physiocrats thought a trajectory of the farming share of the population toward zeo was an absolute absurdity and that in practice it could certaily not fall below half. They were wrong; and more generally, capitalism is not an equilibrium process. There may be seven unsustainable processes out there, or even more, but you cannot show it simply by noting that the trend of some ratio will take it outside its historic range.

UPDATE: Nick Rowe has a kind of response which, while I don't agree with it, lays out the case against regarding money as a liability very clearly. I have a long comment there, of which the tl;dr is that we should be thinking -- both logically and chronologically -- of central bank money evolving from private debt contracts, not from gold currency. I don't know if Nick read the Leijonhufvud piece I quote here, but the point that it makes is that writing 100-odd years ago, Wicksell started from exactly the position Nick takes now, and then observed how it breaks down with modern (even 1900-era modern) financial systems.

Also, the comments below are exceptionally good; anyone who read this post should definitely read the comments as well.


  1. Interesting post. Thanks for this. Are you saying then that the Austrian economists and some post Keynesians are wrong to say that the government has a monopoly on money?

  2. "Conventional monetary policy may not be effective now, but there's no reason to doubt that it will one day get back to so being. I'm not so sure." Exactly - in a credit economy that has reached peak debt/peak elasticity, monetary policy can stabilise and prevent deflation but it cannot create inflation. In the current scenario, inflation in the UK/US can only come about through new fiscal spending.

    I wrote a post a while ago on this topic referring to the Leijonhufvud paper and also work by Claudio Borio which is criminally underread

  3. JW,
    Writing styles vary, and I much prefer yours to David Graeber's. I read only far enough in his book to discover his "anthropologist" approach to what I see as the mathematical problem of money.

    "Why do checks settle at par -- what I pay is exactly what you get -- but debit and credit card transactions do not? Should we care?"

    Wrong questions, perhaps?

    The "obvious reason" for the "evolution of electronic money from credit cards" is that "debit and credit card transactions do not" settle at par, but rather at profit to the issuer. With the support of policy.

    "Checks settle at par". This (thank you very much!) is the reason I look at M1 money rather than either base or mzm, for my Debt-per-Dollar graphs.

    re: [4]
    The price level is not a "ratio" to my way of thinking. As opposed to, say, the Emmanuel Saez graph showing the "Share of total income going to Top 10%" which -- with its straight-line increase to 50% -- clearly CANNOT CONTINUE BEYOND 100%. Thus, this increase is unsustainable. (The red lines are my markup.)

  4. Yep. Interesting post. I think you need to bring in the fact that private monies are (currently) redeemable at fixed exchange rates into central bank money, and it is the issuer of that private money who is responsible to ensure redeemability, not vice versa. So it's analagous to an asymmetric fixed exchange rate system. (If all other countries pegged their exchange rates to the US dollar, then the Fed controls world monetary policy).

    Because private monies are redeemable in central bank money, it makes sense to talk of them as liabilities. But if the central bank money is not redeemable in anything, it's not really a liability in the same sense.

    1. "and it is the issuer of that private money who is responsible to ensure redeemability, not vice versa"

      It is exactly the opposite. Redeemability, or what officially is called deposit insurance, is the primary task of the central bank. Even for purely political purposes.

      "if the central bank money is not redeemable in anything, it's not really a liability in the same sense"

      On the same logic corporate shares are not really a liability because they are not redeemable in anything.

    2. This comment has been removed by the author.

    3. Central Bank money is redeemable to the Treasury against tax obligations

  5. Ashwin,

    I'd forgotten that (very interesting) post of yours. Come to think of it, that may be where I first encountered the Leijonhufvud paper. Yes, your take is similar. I think, though, where we might differ is the logical response. I thin your idea is that we can restore "natural" limits on credit creation by allowing more financial institutions to fail? (And buffering the real economy against those failures by a system of countercyclical transfer to individuals.) Is that right? Whereas I think it makes more sense for the state to not just reclaim its monopoly on means of payment, but to more actively direct investment finance in general. (This was the heart of Keynes' program, by the way -- he had very little interest in countercyclical demand management.)

  6. Arthur,

    Glad you liked the post. Wish you liked Graeber. I find reaction to that book is bimodal and I've been surprised how many people seem to dislike it as much as I liked it.

    Of course you are right that the rents available are why private entities want to be in the means of payment businesses. But that doesn't explain why they succeeded. There are lots of commons that could be profitably enclosed, but in general the public domain has held up surprisingly well. I can't think of ay other core public function that has been wholesale replaced the way this one has. (Though the postal service seems to be dying...) And what's also remarkable is that despite this transformation, the bulk of the economics profession still talks as though means of payment are a government monopoly -- see Nick Rowe's comment (which I'm about to reply to) just below yours.

    Re ratios, my larger point was just that we don't know which economic variables should be regarded as stationary but that implicit claims that we do seem to underlie most statements about things being "unsustainable." Before WWII, people thought that the price level was stationary -- and they were mostly right. Now we don't think so, and we are right too.

  7. Josh - On the need for the state to reclaim or at least tightly regulate Visa-like networks, I'm completely with you. But I don't think this has much impact on the dynamics of the credit economy or the banks' control over money supply. The key point is the banks' ability to originate credit ex nihilo and this is where the commitment to prevent any bank failure or creditor losses means that the system gets overextended to the point where any snap-back triggers a collapse.

    My take on it is still evolving but currently is as follows:
    1. Public deposit option - some sort of pass-through or direct deposit account for all citizens with the central bank.
    2. Open access repo - if you have the right collateral, then you can repo your asset into cash at the central bank discount/repo window. Central bank manages collateral requirements and rates.
    3. Public or tightly regulated + free means of payment infrastructure.

    Essentially disintermediate and make the banks irrelevant - take away their privileged position. So when the ECB did its recent LTRO operations, all EU citizens should have access to it if they have the right collateral and need cash.

    On the investment finance, I agree with the Post-Keynesian arguments on the importance of investment but prefer stimulating private investment via reducing barriers to entry for small businesses and reducing the corporate savings glut but competing it out of incumbent large businesses - so a somewhat Schumpeterian solution for the problem.

  8. "Indeed, in the Benjamin Friedman article quoted above, he explicitly suggests that subway cards issued by the MTA could just as easily have developed into the universal means of payment."

    Hong Kong has a situation like this where the Octopus Card, which was originally developed as a single touchless smart-card for most of HKs overlapping public transit networks, has become a widespread payment method. Octopus cards can be used for parking meters, vending machines, convenience stores, grocery stores, restuerants, movie theaters and just about anywhere else that makes small-dollar transactions. While it has an HK$1000 limit that prevents its use for large-ticket items, its use has become almost ubiquitous. If I recall correctly, there is no service fee or interest paid or charged, and transaction clear at par (above par for some transit, where Octopus cars get a lower rate than cash transactions).

    (Technically the Octopus Card is operated by a private concern, but HK government is the largest shareholder of the company that owns the card, so it is not purely public but a public-private partnership. I am not sure if there are merchant fees that pay for it, or if it relys on the cost-savings from an efficient universal payment system.)

  9. Because private monies are redeemable in central bank money, it makes sense to talk of them as liabilities. But if the central bank money is not redeemable in anything, it's not really a liability in the same sense.

    Yes and no, but mostly no. First of all, central bank liabilities definitely are liabilities for accounting purposes. Any change on the asset side of the central bank's balance sheet has to be balanced by a change in reserves + currency, as we saw so dramatically in 2008. Second, while there is some formal legal sense in which private moneys are redeemable for c.b. money but c.b. money is not redeemable for private money, I'm not sure it is true in any economically relevant sense. Yes, a bank is legally obliged to give me cash out of my checking account, while it is not obliged to let me deposit cash into the account. But in practice the two transactions are symmetrical. There is no point "at the end of the day" when all bank accounts are cashed out. And more importantly for our purposes, it is not at all clear that demand for private money depends on the ability to convert it into c.b. money. If anything, the opposite -- large transactions settle preferentially or exclusively with private money, so demand for c.b. liabilities (currency) is much more likely to be determined by the ease with which they can be converted to banking system liabilities.

    (This is related to the claim that underwrites the idea of the zero lower bound, that money is the asset with the lowest carrying costs. If by money we mean c.b. money, definitely not true today.)

    More broadly, why does anyone want central bank liabilities? Only because they can be transferred to a third party in return for goods and services, or to extinguish your own liabilities. What defines money, as you've eloquently written yourself, is that it is used for transactions. So there is nothing special about central bank liabilities. Sure, there are payments that can only be made in cash. But there are many more payments for which cash cannot be used. What used to make c.b. money special was the close link between reserves and bank money. This was partly due to formal reserve requirements, and partly because there were no good substitutes for reserves for settling interbank transactions. But neither of those conditions has held for a while now.

    (This is where MMT people would say, "Central bank money is special because it settles tax liabilities!" But this isn't right either. First of all, it's perfectly possible to settle tax liabilities in bank money. More fundamentally, there is nothing special about tax liabilities. Yes, we are legally obliged to pay them, but we are legally obliged to pay all our liabilities. What is unusual about the tax authority is the high share (but well short of 100%) of economic units that have them. But this is a difference of degree, not kind. Any liability that is shared widely enough to define a "pay community" can become money. Thus the MTA example.)

    1. Not correct. If you pay your taxes with bank money, the settlement will ultimately be between your bank and government via reserves, and these are CB Money.

  10. I think you need to bring in the fact that private monies are (currently) redeemable at fixed exchange rates into central bank money, and it is the issuer of that private money who is responsible to ensure redeemability, not vice versa. So it's analagous to an asymmetric fixed exchange rate system. (If all other countries pegged their exchange rates to the US dollar, then the Fed controls world monetary policy).

    You know, there have been times when most currencies were pegged to the dollar, and while this certainly imposed constraints on monetary policy elsewhere, sometimes quite severe ones, it was never the case that "the Fed controls world monetary policy." Unless we are talking about very small, very open economies, the elasticity of local banking systems loosens the link between the world dollar supply and local interest rates, just as (but even more than) the elasticity of the domestic banking system weakens the link between monetary policy and market rates. It may sometimes be a useful simplification to talk about a world of perfect capital mobility in which there is "the" global interest rate, but we shouldn't fool ourselves into thinking that world is real.

  11. JW: Thanks. I did a post, inspired by, and partly in response to, this one:

    "First of all, central bank liabilities definitely are liabilities for accounting purposes. Any change on the asset side of the central bank's balance sheet has to be balanced by a change in reserves + currency, as we saw so dramatically in 2008."

    I disagree. Suppose the Bank of Canada gives $1 billion to charity. Is that logically impossible, because the accounts would be unbalanced? (Actually, it gives about $2 billion a year profit to the government.) Or, suppose the BoC bought a load of assets that suddenly all went bad. Or all suddenly doubled in value. Is that logically impossible?

    1. JW:

      The resolution to all the paradoxes surrounding monetarism is that government-issued money is like GM stock, and privately-issued money is like stock market derivatives (calls, warrants, hypothecated shares, etc.). Just as the share price of GM depends on the value of the assets that GM has to back the shares, the price of a central bank's dollars depends on the the value of the assets that the central bank has to back the dollars. Thus the quantity of central-bank-issued dollars can double, and as long the central bank's assets rise in step, the price of a dollar will not change. If the central bank's assets fell, or if its issuance of currency outran its assets, the price of the dollar would fall.

      If Wells Fargo then issued derivative dollars (like checking account dollars, credit card dollars, etc), there would be no effect on either side of the central bank's balance sheet, and the price of a central-bank-issued dollar would therefore not change.

    2. Sorry, I don't accept this. I think we have to be able to theorize a world of pure credit money. No central bank, no gold standard. (That's Wicksell.) I also think as a historical matter that's how things actually evolved -- credit first, then money, and finally central banks. (That's Graeber.)

    3. Nick-

      Your post is extremely interesting. I still don't agree, but I suspect I'm going to spend a fair chunk of the next few days when I really should be working, articulating exactly why not.

      (Reply above was to Mike Sproul.)

    4. JW:

      Pure credit money creates a free lunch for its issuer, not to mention raising puzzles like Wicksell's judgement day. One nice feature of the money-is-like-stock idea (aka the backing theory of money) is that it doesn't require any special theory of money. Money is valued just like any other liability. Clearly, backed money exists. When money is backed and convertible, everyone can see the backing. When convertibility is suspended, people suddenly suppose that the backing is no longer there, even when it's still in the issuer's vault, and still publicly displayed on the issuer's balance sheet.
      Every so-called pure credit money started out convertible into something, and then convertibility was suspended for a weekend, for 30 days, for 30 years, etc. At what point do you say that the money lost its backing and became a pure credit money? The right answer is: You don't. It was backed all along.

    5. Mike,

      Again, I don't agree. I also am unclear if you are describing the way the monetary system does operate, or the way you think it ought to operate. Your point seems to be that an entity whose liabilities are treated as money should have assets equal to its liabilities. But whether this is meant as a factual or a normative claim, it is trivially true, since all economic units have assets equal to their liabilities.

      Also, I think we should limit the term "real bills doctrine" to its actual historical use. This was to describe a monetary policy in which only trade finance (i.e. accounts payable or bills of exchange, etc.) were accepted for discount by the central bank. The idea was that this would automatically cause the money supply to rise sufficiently to with the requirements of real activity, but would prevent monetary finance of asset positions. We can debate the merits of this view but it has nothing to do with money being "backed".

    6. JW:
      My point was that money is a true liability of its issuer, and the issuer's assets matter to the value of that money. This is usually denied by quantity theorists.

      "...nothing to do with money being "backed"..."

      If a banker prints 100 new paper dollars and issues them to a farmer in exchange for the farmer's promise to pay $101 next month when he sells the wheat, that $100 of new paper is clearly backed by the $101 IOU (assume PV of $101 in 1 month=$100). The IOU is in turn backed by the farmer's wheat, and often with some further claim to the farmer's land. As long as the farmer really pays his IOU, the banker has adequate assets to buy back the paper dollars he issued. If the farmer fails to repay, and the IOU loses value, then the banker's dollars will lose value from the loss of backing.

    7. money is a true liability of its issuer, and the issuer's assets matter to the value of that money.

      I think the second part of this statement is just wrong. (Of course I agree with first.) money trades at par -- that's pretty close to the definitional. For nonfinancial units, money in practice means bank liabilities. The assets held by my bank have nothing to do with the terms on which you will accept my check or my debit-card payment, i.e. a transfer of claims on a bank. (Well, in exceptional situations they would, but that precisely would be a sign of those bank liabilities ceasing to be money.)

      I don't want to debate this much more but I don't think your example works. Whenever a bank creates money ( a demand deposit) it is because it has simultaneously created an asset (a loan.) To borrow a phrase from Martin Wolf, "money is just the liability counterpart to private credit decisions." It doesn't make any sense to say that money has more or less value depending on whether there are more or less corresponding assets on the bank's balance sheets. There are always corresponding assets, just as a matter of accounting.

      On the other hand, central governments are the one exception to the rule that financial transactions must produce equal changes on both sides of the balance sheet. Central governments can increase their liabilities by issuing debt without changing their asset position. (To a first approximation, they don't have any assets.) This doesn't stop government debt from behaving in very moneylike ways.

      The test of moneyness is acceptability as payment by third parties. Currency trades at a discount in a foreign country not because there are any doubts about the assets of the issuer, but because different countries constitute separate "pay communities."

    8. JW:

      The bank hasn't "created an asset (a loan.)". The farmer created the asset (for the bank) when he wrote "IOU $101 next month". Assume for simplicity that the IOU is the bank's only asset, and the $100 in newly printed notes its only liability. Clearly, if the value of the farmer's IOU dropped from $100 to $40, then each of the banker's dollars must fall to $.40 (measured in base money). If the banker's dollars were priced at a level above or below their backing value, there would be arbitrage opportunities.

      The main asset of any government could be called "Taxes receivable". Governments clearly have assets.

    9. The bank hasn't "created an asset (a loan.)".

      Yes it has. The IOU wasn't an asset before, just a piece of paper. It only became an asset when the farmer gave it to the bank. (Or as we would normally say, when the bank made a loan to the farmer.)

      Assume for simplicity that the IOU is the bank's only asset, and the $100 in newly printed notes its only liability. Clearly, if the value of the farmer's IOU dropped from $100 to $40, then each of the banker's dollars must fall to $.40 (measured in base money).

      Not a valid simplifying assumption. Real world banks have reserves (on the asset side) and capital (on the liability side) precisely in order to prevent this from happening.

      The main asset of any government could be called "Taxes receivable".

      Some textbooks say this but I think it's silly. But even if one did accept it, it wouldn't do the work you want it to. Many people think that future US tax collections depend on the current value of outstanding debt. Nobody believes that the value of debt outstanding depends on the future path of taxes. Now for some highly indebted developing countries, this *would* be true -- their debt stock is at the maximum level that their tax capacity can sustain. But in those cases, the debt will not function as money -- it will not be accepted as payment by third parties.

    10. Then suppose a government has issued $100 in paper currency, against which it holds assets consisting of $20 of coin reserves, an IOU worth $80, a building worth $30, and taxes receivable worth $40. The government's net worth is thus $70. If the IOU then dropped in value from $80 to $40, net worth would still be positive ($30), so the currency would hold its value, being adequately backed by assets. But if the IOU fell to $0, the remaining assets would be worth only $90 (=20+30+40), so the bank's dollars would fall to $.90, measured against base money.

  12. What defines and distinguishes a central bank in the first instance is its responsibility for managing the national debt and keeping that debt liquid. Managing currency is just an extension of that responsibility; the currency is just another form of the sovereign's debt.

    In the case of the U.S. Federal Reserve, the currency outstanding is not only a liability, but a collateralized (!) liability, backed by specific holdings of longer-term securities. hee-hee

    I don't know if the money illusion is any more scary, when you realize that the money is just ephemeral electronic blips than it was when previous generations realized it was just paper, or generations before that realized that you couldn't eat coins.

    Money is an institution, and the value of money has to be managed and regulated.

    As far as regulation of the payments system is concerned, the next wave of reform will entail resolving many looming problems of privacy and the authentication of identity. Your personal identity is now part of the transaction -- whether it is your membership in a supermarket loyalty program (which entails more than 5%) or the location of your cellphone. Money is more than tokens, now.

  13. I'm still puttering around the net of recent microfoundations blogposts, and the Axel Leijonhufvud essay was like a thunderclap amid the ineloquent mumbling and muttering.

    "New Classical economics originated as a rational expectations refinement of this Monetarist theory. In its further development, this school adopted an intertemporal (complete markets)
    monetary general equilibrium framework. This has two consequences. First, the theory produced anti-Quantity Theory results, demonstrating that the price level was not necessarily proportional to the contemporaneous stock of outside money in equilibrium. Second, the intertemporal GE model generalizes the Modigliani-Miller theorem and extends it to the entire economy. The theorem demonstrates that, given a system of consistent pricing, the valuation of capital resources should be independent of how they are financed. The macroeconomic generalization of Modigliani-Miller implies among other things Ricardian Equivalence, the Ineffectiveness of Open Market
    Operations and more generally the irrelevance of inside money and credit."

    What a paragraph that was, eh?

    It does leave me wondering why infusing Modigliani-Miller would not lead to a more jaundiced view of mortgage-backed securities and, particularly, collateralized debt obligations, and a less sanguine view of bubbles financed by the same?

    And, perhaps, a related question to Ashwin: what's the role of small-bore regulation in protecting the integrity of the securities used for the national debt and banking?

    1. the Axel Leijonhufvud essay was like a thunderclap amid the ineloquent mumbling and muttering.

      Wasn't it just?

      I think that whole immediate post Keynes generation of American macroeconomists (of which he would be sort of the tail end) is really underrated. Wait til you read his piece on inflation, which I will be blogging about soon, ojala.

    2. And yes, the density of insight in that paragraph (and the whole article, really) is kind of astonishing.

    3. Bruce - not sure I understand your question. Do you mean the regulations regarding the assets than can serve as collateral at the central bank?

    4. No, I mean the administrative processes by which banks and other financial intermediaries control credit (loan and financial security) creation, plus the regulations of the central bank and other public and private regulators, which seek to constrain those administrative control processes.

      I am thinking, in particular, of mortgage underwriting procedures that drove the housing bubble, and related security underwriting (e.g. Collateralized Debt Obligations) practices.

      You seem to think that firewalling the payments system away from Wild West intermediation is necessary, which I don't disagree with, and you argue that intermediaries have to be exposed to self-destructive loss, which I also do not disagree with. The credit/payments system still requires some form of intermediation; if collateral can be exchanged for cash, someone has to apply some set of procedures to judging the collateral and borrower, assigning contingent property rights, etc. It is a technical problem, and must be subject to technical regulation/constraint/prescription, I would think, of some sort. What sort?

      The firewall and failure prescriptions cover the aftermath, but what about the prequel? Don't we want the financial system meeting the Modigliani-Miller condition in creating credit and marketable financial on an on-going basis? No billion-dollar Ponzi schemes. No trillion-dollar housing bubbles. Etc. Certainly, none that seek to insulate the bad guys from the consequences, by holding the payments system hostage, so, at the very least, we need some scheme for regulating the creation of credit, within whatever firewall surrounds the payments system.

    5. Got it - it's an important issue.

      What I'm proposing with this public deposit + open repo window is to separate the depository/transactions/medium of account function from the credit creation function. On the credit creation function, we need collateral standards i.e what is acceptable at the repo/discount window? My first point is that we need to open both functions up to a much wider section of the population rather than just the banks.

      Another crucial point is that we need to change the current criteria for what the CB likes to hold or lend against. I wrote a post a while ago in which I argued that the central banks' aversion to losses makes it intervene only in macro, bubble-prone asset markets. People think that CB stimulus promotes business lending but unrated small-business loans are not eligible collateral for any CB intervention! Too much idiosyncratic risk and risk of losses to CB. When it started out, the Fed discounted agricultural and industrial bills of exchange and nothing else - now they intervene in everything else but business loans. It is only recently that the ECB has included unrated business loans in its list of eligible collateral - a move that is far overdue.

      I want the procedures to judge the collateral, not the borrower. I'd personally prefer that these collateral standards be limited to financial/real assets with significant haircuts and short-term bills of exchange/working capital loans a la the old BoE and Fed. If you believe we still need significant maturity transformation, then we also need to include longer-maturity loans - my opinion is that we no longer need any meaningful amount of maturity transformation given the assets held by pension funds and life insurers that wants to invest for the long-term. Moreover, my reading of history is that in the Anglo-Saxon economies we have never needed much maturity transformation for real economic lending. The UK and the US financed most of their capital spending during the 19th century in the public markets - it is the European economies, especially Germany, that did finance their capital spending through banks in the 19th century but this worked because this was catch-up growth. This is the reason why Schumpeter was so obsessed with banking's importance in economic growth - in Germany, bankers were indeed the capitalists.

    6. One quibble.

      The UK and the US financed most of their capital spending during the 19th century in the public markets - it is the European economies, especially Germany, that did finance their capital spending through banks in the 19th century but this worked because this was catch-up growth.

      Banks were certainly central in Germany (and Japan), just as the state was in later industrializers. (Gerschenkron.) But I don't think it's true that bond markets were the main source of investment finance in the US and UK. Except for railroads -- admittedly a major exception -- the great majority of investment was financed internally, especially in the UK, all the way through the 19th century.

    7. My mistake - I should have said that the "external" financing for investment came from markets rather than maturity transformation via the banks. Of course the majority of the investment was financed internally as you point out.

  14. "Currency looks like classical money, like gold; but demand deposits do not. The most obvious difference, at least in the context of macroeconomics, is that one is exogenous (or set by policy) and the other endogenous."

    Another big difference is that bank accounts or debt require interest payment, and as a consequence are inherently a form of capital, while commodity money insn't necessariously a form of capital.

    From this point of view, treasury bonds are a form of capital (people who whant a capital investment could buy a lot of treasuries instead than build a car factory and run it, for example), while banknotes aren't (one can't sit on a pile of banknotes and hope that it grows). The Hong Kong Octopus Card might be different fron credit cards from this point of view, as I suppose that people just "charge" it with money without expecting to pay or to be payd interest.

    This is important in my opinion because I believe that people who buy capital assets follow a different "rationality" than people who buy consumer goods:
    people who buy consumer goods have to weight the different "utility" they perceive they would get from different consumer goods (or from consumption at a different time), while people who buy capital goods don't get "utility" from the capital goods, they just (hope to) get more money from them that they put in the first place (I think this is what Marx meant when he wrote that workers have a C-M-C cycle while "capitalists" have a M-C-M cycle), and they weight the different profit rates that they can get from different capital goods.

    For example, suppose that A is a saver that thinks as a consumer, while B is a saver who tinks as a "capitalist". A puts 100$ in treasuries, and B also puts his 100$ in treasuries.
    After some time, A gets back 101$, and spends that money in consumption (that is, A just delayed his consumption from t0 to t1); B also gets the money, but since he thinks as a "capitalist" he doesn't spend the money, he just buys another 100$ of treasuries and spends 1$ for consumption (or maybe 100.5$ in treasuries and 0.5$ in consumption).

  15. If the profit rate that B can expect from other investiments is low, he will often "invest" in treasuries instead than in actual economic activity. The government then uses the money to pay for government workers, pensions, whatever, and "recicles" the money back in the demand for consumer goods (thus rising the profit rate of other kinds of investiment). Something similar happens when B lends to private entities (who either invest the money in phisical economic activities or are consumers).
    So in pratice when money is created through credit, the "credit" money becames capital while the "debt" money becames consumption.
    But if the profit rate of non-financial economic capital is very low, and thus many people act like B and invest in "debt", economic growth becames dependent on the growth of total "debt", that has to grow faster than the economy (while the interest payd on debt falls, because it competes with always lower rates from non financial capital).

    In my opinion, USA economy has been "in trouble" since a lot of years, because as a net importer the USA is very likeli to suffer from crises of demand (since part of the demand is leaked outside). The consequence of a demand crisis is that the rate of profit of capital goods falls; the USA government then fought the crises first by "Keynesian" ways (by issuing debt directly), later by "monetarist" ways (by facilitating in various ways private debt). In the meanwhile most other countries doged their demand crises mostly by exporting, in a system that is (still) based on the USA running huge deficits.
    In my opinion, the reason that the dollar is the ultimate safe heaven is that, since all countries ultimately export to the USA (as in: Italy exports to Germany which exports to China which exports to the USA), when the dollar falls all the others face economic crises and also fall (for example the EU would solve most of its problems by devaluing the €).

    Bottom line: I didn't read Graeber and do not know what does he think about debt Vs currency, but I think that in our time the reason whe have mostly credit money is that the increase in total leverage helped, I think from the 70s till now, to hide under the rug a really big "crisis of demand".

  16. Cool - looking forward to more about Leijonhufvud (I'm sure people with hard-to-spell/pronounce names tend to get neglected because people are scared of getting it wrong. Maybe call it Smith's Law, or Bob's, or something).

    There's an interview where he describes that inflation paper as "one of my worst papers because it was written in a bad temper." And "[a]fter giving that paper I became known as an anti-inflation hawk when it was not at all fashionable to be one. I even heard some younger people say... ‘oh, it makes sense, he comes from UCLA and they are worse than Chicago’!"

    1. Wow, thanks for that interview. That the inflation paper was written in a bad mood comes through pretty clearly, but it's still the best case for inflation hawkery that I have seen.

  17. Ashwin,

    I agree, we should be talking about disintermediating the banks, meaning (a) a public option for savings -- postal savings, etc. (It's really unfortunate for this reason also that we are in the process of destroying the postal service, that infrastructure would be ideal for a public saving system.); (b) a public payments system; and (c) access to a discount window type facility for nonbanks. Just what you said, in other words.

    You also say:

    On the need for the state to reclaim or at least tightly regulate Visa-like networks, I'm completely with you. But I don't think this has much impact on the dynamics of the credit economy or the banks' control over money supply. The key point is the banks' ability to originate credit ex nihilo and this is where the commitment to prevent any bank failure or creditor losses means that the system gets overextended

    You may be right, the privatization of means of payments and the lack of purchase of monetary policy over credit creation may be to separate issues. (Though I feel there's a link between them.) Where I continue to disagree with you is the idea that allowing more bank failures is a feasible solution.

    1. Fair enough - let me just clarify a couple of things.

      By failure, I really mean that we need to enforce losses on creditors. So we don't need the TBTF banks to go up in flames necessarily.

      The aim of those three ideas is to suck the oxygen out of the current big banks and put in place a firewall between some core functions and our current dysfunctional system. This should allow us to put in place conditions where banks can then be allowed to fail without triggering systemic collapse as they probably would now.

      On links between the means of payment infrastructure and bank credit creation, you may be right. If you have any speculative ideas on what the links may be, I'd love to hear them.

      On a similar topic, I highly recommend a new book called "Casualties of Credit: The English Financial Revolution, 1620-1720" by Carl Wennerlind. The topic is how creating a credit money system allowed Britain to overcome a chronic shortage of metallic currency in the 17th century. Fascinating - and I discovered that my idea for a open repo window was even advocated in the 17th century!

    2. I guess the issue here is, does the state want to be able to control the scale and direction of private credit? (I phrase this as "does the state want" rather than "do we want the state" because I don't think normative policy discussions make much sense without concrete political contexts.) If the state wants to do that, it needs tools to control private lending/money creation, of which a monopoly in the means of payment (desirable anyway as its basic infrastructure) could be one -- tho I admit i'm not sure how far it goes in that direction and you may be right it doesn't go very far.

      Now one could say -- as I've understood you to be saying -- that if we had an appropriate firewall between the banking sector and the real economy, by removing some of the core infrastructure functions from private banks as discussed above, then active control over private credit creation would be less necessary, because the discipline of bank failure would prevent such wild swings in private lending. That's what I'm skeptical of -- it seems to me we've tried that experiment in the pre-WWII period and it just resulted in bigger downward deviations without reducing the upward ones. (Of course allowing bank failures might be desirable for other reasons, like limiting the political power of banks.) One could also say, as some MTT people do, control over private credit is less necessary if government is willing to use other tools of demand management. In particular, one could argue that because expansionary monetary policy is so unreliable (and tends to create asset bubbles when it does work) that the arrangement of the 1950s-1970s is preferable, where fiscal policy is tuned to overfull employment and monetary policy is contractionary as necessary. Whereas today, we have fiscal policy normally tuned to barely or less than full employment (not so much because fiscal policy itself as changed, as because there's been a downward shift in private investment demand) and we've relied on monetary policy for both stimulus and contraction. I'm much more sympathetic to this argument. But I still think it's hard to imagine macroeconomic stability without a central bank that is able to effectively control private credit creation.

      On the other hand, I imagine you'd say that that is a mixed blessing, since excessive stability can come at the cost of resilience. Which is a position I respect, though I don't share it in this particular case.

    3. Also, thanks for the reference -- look forward to reading it. I think it's very important to approach this stuff in a concrete historical way.

  18. If banks weren't allowed to create money, the Fed or Treasury expanded the money supply by putting money on everybody's postal bank debit cards, how much money are we talking about? Playing with wolfram alpha it looks like about $1000/year.

  19. "I think people underestimate the extent to which modern central banking depended on a public monopoly on means of payment..."

    I would argue central banks matter when they act like very large private banks.

    Assuming that Leijonhufvud is correct, we should dismiss the distinction between "central" and "private" banks. Except for its institutional ties, a central bank is a bank like any other: it can create deposits at will. To make it seem more like a regular bank, let's call it "C*Bank" instead of "central bank". Sure, C*Bank deposits are created for the benefit of other private banks -- it is a correspondent bank, after all -- but this distinction is not important.

    Unlike other banks, the C*Bank has an institutional interest in financing only one specific type of current spending: net fiscal spending. Following a financial crisis, net fiscal spending can reach upwards of 10% of gdp, placing large demands for deposit creation on the part of undercapitalized private banks. In this case, C*Bank is likely to step in to dominate financing at very attractive (negative real) rates. If net fiscal spending is large enough, C*Bank deposits can quickly grow to equal 20-50% of the stock of total bank deposits. In other words, C*Bank becomes a "mega bank" that rapidly grows the deposit base. If the perception takes hold that C*Bank will forever finance public credit growth at negative real rates, then its depositors will likely choose to redeem funds in search of positive real returns (in real goods/assets or non-C*Bank currencies).

    So it seems that private bank deposit creation will drive the price level during most times. Following a financial crisis, however, central banks, acting like private banks to finance current spending growth, can and will create the conditions for high and variable inflation.

  20. Assuming that Leijonhufvud is correct, we should dismiss the distinction between "central" and "private" banks.

    I think this is right. We need to think of central banks as evolving organically out of private banks.

    I'm not sure about the rest of your comment. Your concern with the price level as the key outcome determined by the pace of deposit creation is properly Wicksellian, but I'm not sure it's the best way to frame the issue for modern economies, since it implicitly assumes that total output is fixed.

  21. Why do checks settle at par -- what I pay is exactly what you get -- but debit and credit card transactions do not? Should we care?

    Checks settle at par because they are written from checking accounts (try writing more than the federal limit per month from your savings account - I think that limit is 6 or 7 - and your savings account will be converted to a checking one). Checking accounts don't pay the same interest rate as CDs or savings accounts (I know, I know... at least that was the case till 2008!). So checks settle at par because you give up the interest you could have earned on the money kept in the accounts you write your checks from. There is a cost... just as there is to using credit/debit cards and their interchange fees.

    Of course, the difference between those two price levels - especially in today's ZLB world can be debated.

  22. Our Persian felines are fed a high-quality diet, necessary feline formulated vitamins, mineral supplements and provide plenty of daily attention, kittens care and grooming Here at Marko Persians our cats and kittens are happy, sweet love bunnies, therefore, well socialized Healthy and exquisitely beautiful and unmatched in the Silver Doll Face Industry. Specializing in CFA Persian kittens meeting the breed standard with several coat colors and coat pattern – shades

    Persian Kittens for sale near me
    Persian Kittens for sale
    Persian Kittens for sale