Note: This post only really makes sense as a continuation of the argument in this one.
It’s a general rule that the internal logic of a system only becomes visible when it breaks down. A system that is smoothly reproducing itself provides no variation to show what forces it responds to. Constraints are invisible if they don’t bind. You don’t know where power lies until a decision is actively contested.
In that sense, the crises of the past seven years — and the responses to them — should have been very illuminating, at least if we can figure out what to learn from them. The current crisis in Greece is an ideal opportunity to learn where power is exercised in the union, and how tightly the single currency really binds national governments. Of course, we will learn more about the contours of the constraints if the Syriza government is more willing to push against them.
The particular case I’m thinking of right now is our conventional language about central banks “printing money,” and the related concept of monetary sovereignty. In periods of smooth reproduction we can think of this as a convenient metaphor without worrying too much about what exactly it is a metaphor for. But if Greece refuses to accept the ECB’s conditions for continued support for its banks, the question will become unavoidable.
We talk about governments “printing money” as if “money” always meant physical currency and banks were just safe-deposit boxes. Even Post Keynesian and MMT people use this language, even as they insist in the next breath that money is endogenously created by the banking system. But to understand concretely what power the ECB does or does not have over Greece, we need to take the idea of credit money seriously.
Money in modern economies means bank liabilities. [1] Bank liabilities constitute money insofar as a claim against one bank can be freely transferred to other units, and freely converted to a claim against another bank; and insofar as final settlement of claims between nonfinancial units normally takes the form of a transfer of bank liabilities.
Money is created by loan transactions, which create two pairs of balance-sheet entries — an asset for the borrowing unit and a liability for the bank (the deposit) and a liability for the borrowing unit and an asset for the bank (the loan). Money is destroyed by loan repayment, and also when the liabilities of a bank cease to be usable to settle claims between third parties. In familiar modern settings this lack of acceptability will be simultaneous with the bank being closed down by a regulatory authority, but historically things are not always so black and white. In the 19th century, it was common for a bank that ran out of reserves to suspend convertibility but continue operating. Deposits in such banks could not be withdrawn in the form of gold or equivalent, but could still be used to make payments, albeit not to all counterparties, and usually at a discount to other means of payment. [2]
To say, therefore, that a government controls the money supply or “prints money” is simply to say that it can control the pace of credit creation by banks, and that it can can maintain the acceptability of bank liabilities by third parties — which in practice means, by other banks. It follows that our conventional division of central bank functions between monetary policy proper (or setting the money supply), on the one hand, and bank regulation, operation of the interbank payments system, and lender of last resort operations, on the other, is meaningless. There is no distinct function of monetary policy, of setting the interest rate, or the money supply. “Monetary policy” simply describes one of the objectives toward which the central bank’s supervisory and lender-of-last-resort functions can be exercised. It appears as a distinct function only when, over an extended period, the central bank is able to achieve its goals for macroeconomic aggregates using only a narrow subset of the regulatory tools available to it.
In short: The ability to conduct monetary policy means the ability to set the pace of new bank lending, ex ante, and to guarantee the transferability of the balances thus created, ex post.
It follows that no country with a private banking system has full monetary sovereignty. The central bank will never be able to exactly control the pace of private credit creation, and to do so even approximately except by committing regulatory tools which then are unavailable to meet other objectives. In particular, it is impossible to shift the overall yield structure without affecting yield spreads between different assets, and it is impossible to change the overall pace of credit creation without also influencing the disposition of credit between different borrowers. In a system of credit money, full monetary sovereignty requires the monetary authority to act as the monopoly lender, with banks in effect serving as just its retail outlets. [3]
Now, some capitalist economies actually approximate to this pretty closely. For example the postwar Japanese system of “window guidance” or similar systems in other Asian developmental states. [4] Something along the same lines is possible with binding reserve requirements, where the central bank has tight operational control over lending volumes. (But this requires strict limits on all kinds of credit transactions, or else financial innovation will soon bypass the requirements.) Short of this, central banks have only indirect, limited influence over the pace of money and credit creation. Such control as they do have is necessarily exercised through specific regulatory authority, and involves choices about the direction as well as the volume of lending. And it is further limited by the existence of quasi-bank substitutes that allow payments to be made outside of the formal banking system, and by capital mobility, which allows loans to be incurred, and payments made, from foreign banks.
On the other hand, a country that does not have its “own” currency still will have some tools to influence the pace of credit creation and to guarantee interbank payments, as long as there is some set of banks over which it has regulatory authority.
My conclusion is that the question of whether a country does or does not have its own currency is not a binary one, as it’s almost always imagined to be. Wealth takes to form of a variety of assets, whose prospective exchange value can be more or less reliably stated in terms of some standard unit; transactions can be settled with a variety of balance-sheet changes, which interchange more or closely to par, and which are more or less responsive to the decisions of various authorities. We all know that there are some payments you can make using physical currency but not a credit or debit card, and other payments you can make with the card but not with currency. And we all know that you cannot always convert $1,000 in a bank account to exactly $1,000 in cash, or to a payment of exactly $1,000 – the various fees within the payment system means that one unit of “money” is not actually always worth one unit. [5]
In normal times, the various forms of payment used within one country are sufficiently close substitutes with each other, exchange sufficiently close to par, and are sufficiently responsive to the national monetary authority, relative to forms of payment used elsewhere, that, for most purposes, we can safely speak of a single imaginary asset “money.” But in the Greek case, it seems to me, this fiction obscures essential features of the situation. In particular, it makes the question of being “in” or “out of” the euro look like a hard binary, when, in my opinion, there are many intermediate cases and no need for a sharp transiton between them.
[1] Lance Taylor, for instance, flatly defines money as bank liabilities in his superb discussion of the history of monetary thought in Reconstructing Macroeconomics.
[2] Friedman and Schwartz discuss this in their Monetary History of the United States, and suggest that if banks had been able to suspend withdrawals when their reserves ran out, rather than closed down by the authorities, that would have been an effective buffer against against the deflationary forces of the Depression.
[3] Woodford’s Interest and Prices explicitly assumes this.
[4] Window guidance is described by Richard Werner in Masters of the Yen. The importance of centralized credit allocation in Korea is discussed by the late Alice Amsden in Asia’s Next Giant.
[5] Goodhart’s fascinating but idiosyncratic History of Central Banking ends with a proposal for money that does not seek to maintain a constant unit value – in effect, using something like mutual fund shares for payment.
To me "full" monetary sovereignty means that the monetary authority pays no attention to exchange rates. By this definition the number of monetary sovereigns is much smaller than the number of currencies, since most central banks try to stabilize exchange rates, at least as a secondary objective. A corollary is that some central banks are monetary superpowers, influencing monetary policy beyond their borders.
A CB can largely sterilize international capital flows if it has the will. No limit on how low it can push the currency, limit on how high it can keep it.
That is a strong claim. What's the response to financial outflows?
Sorry I wasn't very clear. Ability to control financial outflows is a constrained by size of FX reserves, so a CB can always push the price of it's currency down by selling, but is limited in its ability to push it up by its FX reserve.
This Godley Lavoie paper is good on why mainstream mundell-fleming and monetary views on exchange rates and monetary theory are wrong,
"These “findings” are profoundly at odds with much conventional wisdom and with the received view that arises from the standard Mundell-Fleming model. Alan Greeenspan has been saying that Chinese accumulation of US Treasury bills is making it difficult for them to manage their monetary policy; but the above analysis strongly suggests that he is mistaken. Peter Kenen (1985: 669) writes, as many others: “…. Reserve flows alter the money stock, undermining the influence of monetary policy…. The monetary approach to the balance o payments is built on this basic proposition”. But the monetary approach does not have a fully articulated monetary system in which the private sector allocates its wealth between money and other assets. In both countries, the private sector’s accumulation of wealth and its allocation between available assets are not in any way affected by these central bank operations beyond what is implied by the step change in disposable income. "
"Two-country Stock–Flow Consistent Macroeconomics Using a Closed Model Within a Dollar Exchange Regime"
http://www-cfap.jbs.cam.ac.uk/publications/downloads/wp10.pdf
This is clearly in line with what you are saying in this post.
*formatting problem in the above deleted post
Ability to control financial outflows is a constrained by size of FX reserves, so a CB can always push the price of it's currency down by selling, but is limited in its ability to push it up by its FX reserve.
I would say that this is conventional wisdom.
These “findings” are profoundly at odds with much conventional wisdom and with the received view that arises from the standard Mundell-Fleming model.
I don't think these dueling formalisms are very enlightening. Yes, it is silly to say that there is a direct link between foreign exchange intervention and the "money stock," a useless concept in any case. But silly arguments are often made for valid positions. Historical experience shows clearly that central banks' ability to control domestic credit conditions are limited in a setting of free capital mobility. Just look at the experience of the United States in the 1960s, with the development of the eurodollar market.
"Historical experience shows clearly that central banks' ability to control domestic credit conditions are limited in a setting of free capital mobility."
But as you point out in the post, a CB's ability to control domestic credit conditions are limited by *domestic* capital mobility as well as international capital mobility. So the issue here is really one of financial innovation vs regulation.
As far as exchange rates go, In the Godley and Lavoie paper, there are three conditions that have to be met so that an exchange rate peg constrains a CBs interest rate setting power.
1. The country has a current account deficit
2. The country does not have adequate forex reserves
3. The countries currency is not acceptable as an international reserve
Obviously many countries meet these criteria, usually developing countries, which is where we see most currency crisis.
But to say that a CB which stabilizes its currency is giving up monetary sovereignty is just not correct.
I would say that's a necessary condition but not a sufficient one.
As you know, I think it's pretty clear that day in, day out monetary policy these days means influencing interest rates — not only "the" interest rate but also the structure. One dimension of sovereignty is a central bank's ability to do this. It's a problem for Greece that its CB can't just supply liquidity to buy up a whole bunch of public and private debt, although this would be only slightly less the case if it were back in drachmas. But I also think you are lumping all the various financial liabilities together too precipitously; there are liquidity differences between different types of tradeable liabilities, and this matters for "the" money supply. ("The money supply" is quite a misnomer.) Being able to supply more liquid means of payment to support less liquid is an important CB function, and perhaps not only in times of crisis. A different way to put it is that the pyramid of private sector credit creation rests on top of a set up implicit guarantees from the state. This is how the 21st c. is different from the 19th and before. That issue also came up with Bitcoin, didn't it?
I don't think we disagree at all.
A point that is quite OT:
You say: "Money is created by loan transactions […] Money is destroyed by loan repayment"
Then you say: "To say, therefore, that a government controls the money supply or "prints money" is simply to say that it can control the pace of credit creation by banks"
However these two statements are very different IMHO. If we take the first statement, and we take a certain period of time as a reference, say one year, we see that the net money creation might be 0, while the flow of money will be equal to GDP:
1: capitalist borrows 100$ (at 10%interest) and pays for workers (100$ are created), who produce 100 stuff units;
2: capitalist sells each stuff unit for 1.5$;
3: workers buy 67 stuff units, payng with their whole wages (100$ are destroyed as the capitalist pays principal to the banker);
4: The banker adds 10$ of debt to the capitalist (for interest), and 10$ of credit for himself. Then uses his 10$ to buy 7 stuff from the capitalist, who then pays back the debt (10$ are created and immediately destroyed);
5: The capitalist then sells to himself the remaining 36 stuff, which represent his profits (makes sense in terms of a capitalist class);
6: All the actors then consume the stuff, that is composed by consumption goods.
In this cycle there is exactly 0 net money creation, so the government has no reason to "print money", however there are also 0 money savings.
If one of the actors, let's say the banker, wants to save 3$, the capitalist will be stuck with 2 unsold stuff and an unpayable debt of 3$.
In the next iteration of the cycle there will be still a flow of 100$, that are created and destroyed during the year, plus a stock of savings of 3$ for the banker, plus a stock of debt of 3$ for the capitalist.
There isn't a direct relationship between the flow of NDP and the stock of savings, one could have a stock of savings/debt of 50% NDP, or of 150% NDP, however the government can try to influence it e.g. by rising or lowering the interest rate, or by borrowing/saving.
Now when you say that the government "control the pace of credit creation by banks", it seems to me that you mean that it controls the savings/debt stock, not the NDP flow, and the two things are very different.
You are right. To avoid confusion, we simply need to be careful about using a consistent language. If we are thinking in terms of balance sheets, then the change in the outstanding quantity of money is equal to the change in the outstanding quantity of loans. In this case, we are not talking about income and payment flows, but comparing snapshots at different moments at time. On the other hand, if we are interested in income and payment flows, the quantity of money is not something we need to be talking about at all, and the relationship of credit transactions to income and payment flows is indeterminate — or rather, depends on a lot of specific factors that we are ignoring at this level of abstraction. In general, credit is used to take asset positions. Thinking there is some direct relationship between current income and payments, on the one hand, and balance sheet variables, on the other, is usually a mistake.
I should add — I think the example you give is clarifying. But credit is really specific to an expanding, or at least changing, economy. In simple reproduction money and credit are just accounting devices.
Incidentally, in the case you describe, income falls by $3 times a multiplier, which in this case will be 3. So total production falls by $9 — the initial reduction in the consumption out of interest of $3, and then an additional reduction of workers' consumption of $6. (We can assume the capitalists consumption does not change.) Describing the balance-sheet results would require some additional assumptions, which are far from obvious.
That's a really excellent description of the endogenous/horizontal money in a modern economy, cutting through the euphemistic/holdover language. But it seems like a big piece is missing here without giving the same treatment to the vertical/exogenous government money (or maybe 'net financial assets' in MMT language). Cutting through the lazy/holdover language there usually ends up with some description like "all government spending is new money creation" or maybe "government borrowing is just 'printing bonds' ", etc.
I only bring it up because it seems like this is what people are usually meaning when they worry about 'printing money'. I take it some people still want to call QE money printing but they just look foolish lately. But for instance if you were to remove the self-imposed constraints between the central bank directly 'monetizing' the deficit (direct purchase of treasury bonds, or just allow an infinite overdraft on the treasury reserve account); that's when people come out of the woodwork in shock about printing money.
I think the focus on exogenous money is a wrong turn, probably the biggest mistake MMT makes. The statements "all government spending is new money creation" and "government borrowing is just 'printing bonds" are simply false. Or rather, they are true but totally misleading — when government spends in excess of current income, it creates money in exactly the same way any economic unit does, through the creation of new bank liabilities. Currency is a device for transferring deposits from one bank to another; it is in no way different from other paper media used for this purpose, like travelers' checks, cashier's checks, money orders, etc. The fact that physical currency happens to be printed by the government is a historical accident of zero economic significance. (OK, it does prevent the banking system from using paper currency as a another site for rent extraction via fees. But no MACROeconomic significance.) As for reserves, they are a regulatory tool for controlling the pace of credit-money creation by banks.
So, you are right that government spending creates money, but wrong to think that the word "government" has any relevance to the truth of this claim. This side of MMT is isomorphic to the totally mainstream fiscal theory of the price level, and is just a transposed version of monetarist thinking from a decade ago — they said all that mattered was money creation by the monetary authority, MMT says that all that matters is money creation by the fiscal authority, but it's the same fixation on an imaginary quantity "M". As so often, heterodoxy is just the orthodoxy of a generation ago.
MMT: Keep the functional finance, lose the chartalism. That's my slogan.
"Or rather, they are true but totally misleading"
Yeah I somewhat agree, which is why I'm interested in a more articulate description that stays accurate. Seems like usually the big stumbling block is treasury/CB consolidation, but when anyone accepts the 'consolidated government' analysis, then those buzzy taglines get pretty close to telling the right story.
"when government spends in excess of current income, it creates money in exactly the same way any economic unit does, through the creation of new bank liabilities." …. "As for reserves, they are a regulatory tool for controlling the pace of credit-money creation by banks"
Here's where it seems like you're selling the exogenous-side a bit short. The fundamental accounting is absolutely the same, where someone (government/firm/household) spending more than their income is creating 'money' / financial assets for others. But the private credit creation must be underwritten and carries innate default-risk, because it's based on liabilities with promised convertibility. Government spending on the other hand creates non-convertible financial liabilities, which not only don't need to be underwritten, but actually *are* the thing that private liabilities promise to convert into.
It's true that most government deficit spending just ends up as commercial bank demand deposits, rather than any physical currency. But these are simultaneously matched with adding a combination of reserves and treasury securities of the same amount, so this money is more akin to equity for the private sector. That's one framing I'm pretty interested in, from geerussell: http://imgur.com/8Q8V336
I think your analysis of the government side is correct. What I think you're missing is that the same analysis basically applies to the private financial system, or at least it's a difference of degree but not of kind. I think it might be helpful to you to imagine an economy with a banking system, but no government, and specifically about the WIcksellian limit where there is no outside money.
http://imgur.com/8Q8V336
It's an interesting picture. But I think a hard line between government and private liabilities is not helpful. Government issues a variety of more and less liquid liabilities, and so do private units. It is certainly true that government borrowing is a source rather than a use of liquidity for the private sector. But this can be true of some private units too.
The other perhaps more important point that I think is missed in the MMT analysis is that the macroeconomic effects of government spending do not only come from it effects on private balance sheets. In some conditions, it may indeed be an important consequence of increased government spending that it creates additional liquid assets for private units. But it also increases incomes or private units, independent of these balance sheet effects, and these effects are (I think) generally more important. Consider the effects of a public works program paid for by a tax on high incomes — "net wealth" for the private sector does not increase but incomes and output certainly does. Or more generally the existence of a balanced-budget multiplier.
One more attempt. (And thanks for the comments, it's helpful to me in thinking about this.) It is true that, if we think of a consolidated government sector that includes the central bank, it can never face financial constraints (as long as it doesn't emit liabilities in foreign currencies, obviously.) But the exact same thing is true if we think of a "consolidated financial sector" that includes the central banks and the systemically-importabt/too big to fail/etc. financial institutions. And there are other private units that, while they may in principle be subject to financial constraints, are not actually constrained in practice, for various reasons. Now furthermore — and this is really the point here — the problem in Greece, and in many other crisis situations, has to do with the liabilities of the financial sector, not the (ex-CB) government sector. So the fact that state liabilities are in principle backed by the CB is irrelevant.