Bill Mitchell has responded to my criticism at great length. I’ll try and boil down the points of disagreement, but this is still going to be pretty long in response!
First, Mitchell skips right over my central criticism: that the expectations and opinions of wealth-holders matter even when they are wrong. Even if money works the way Mitchell thinks it does, if money managers expect a fall in the value of the currency, they’re going to speculate against it. Note the period after Hawke’s election when despite policymakers realising that monetary targeting was rubbish, they kept announcing targets to appease the wrong-headed markets. (See Simon Guttmann’s 2006 ‘The Rise and Fall of Monetary Targeting in Australia’.)
2. Tax and the demand for currency
Mitchell implies that the demand for domestic currency to pay domestic taxes underpins its value, such that the movements out of the currency by wealth managers would be foiled by their need to pay tax. A fixation on the fact that tax payments are denominated in national currency is common among cartalist (state) money theorists. I’m not entirely sure why – there will also be a demand for currency arising from its legal tender status, from the need to make all kinds of payments denominated in it – why single out tax payments? At any rate, if the government is running a deficit and not issuing equal amounts of bonds, it is – by definition of the word ‘deficit’ – injecting more money into the economy than it is calling back in taxation. Finally, the fact that wealth-holders are selling the local currency to buy foreign currency doesn’t destroy the local currency, which remains in someone’s hands. Even if at a lower value in terms of other currencies and goods, it still remains available to pay taxes, even if individuals have to borrow back some of the currency they’ve sold to do it. I’m sure people are aware that currencies can and do dive in the foreign exchange markets despite the government that issues the currency requiring tax payments in its own currency.
3. Banks and the supply of currency
My argument is not at all based on a money multiplier conception as Mitchell presents it. My thinking on money is also heavily influenced by post-Keynesian thought, especially the so-called ‘structuralists’ such as Hyman Minsky and Victoria Chick. I note that Mitchell includes Minsky as one of his ‘modern monetary theorists’ but Minsky is miles away from his cartalism.
To summarise very quickly a complex picture: Currency (or ‘high-powered money’) plays a special role within a country’s monetary system, mainly as bank reserves but also the paper cash we carry in wallets. But the bulk of the money supply is not currency but private bank liabilities – bank deposits, and also various other liquid financial instruments – where you draw the line between ‘money’ and just plain ‘debt’ depends on what you’re looking at.
Banks are traditionally constrained in their lending by a couple of factors. First, they need to be able to meet net withdrawals and net transfers to other banks with currency, and since their assets tend to be less liquid (i.e. readily saleable) than their liabilities, a bank needs either an adequate reserve of currency or a way to quickly get its hands on some when it needs it. Second, banks are often constrained by regulations requiring them to hold at least a certain portion of their assets in liquid form, and/or limiting their lending to some multiple of their capital.
Over the last several decades banks evolved various practices and markets to minimise reserve holdings – first ‘asset management’, involving the development of markets in which banks could quickly offload securities; then ‘liability management’, involving markets in which banks could quickly borrow to meet their reserve needs. (Finally, recent engagement in off-balance sheet activities was motivated by the capital requirement regulations, but we can leave that alone here.) Mitchell is quite right that individual banks are no longer really quantitatively constrained by their reserves, and this is why central banks are these days generally concerned instead with the interest rates on short-term borrowing in the money market, representing the cost of reserves to banks.
But for the banking system as a whole, the quantity of reserves available is limited, and the base interest rate represents supply and demand conditions in the money market. Supply is a real constraint on the system as a whole, though not a firm, clear-cut one. Of course, the central bank intervenes in this market in various ways, buying and selling short-term instruments, lending directly to banks, and entering into repurchase agreements. But the fact remains that it is intervening in a market in which it is powerful, but which it does not control. Notice that though the central bank has a good deal of power over short-term rates, its power diminishes more the further you go up the rate spectrum.
The question relevant to the discussion here is: will the central always be able to soak up extra currency that finds its way into the money market as a result of unfunded government deficits? I note that on this question Mitchell actually goes further than I would, claiming that attempts by the central bank to reduce the money supply always fail. I would say, rather, that sometimes it would be able to soak up the currency and sometimes not, depending on the state of demand for credit in the broader economy.
4. Sectoral balances
The fundamental problem here is that Mitchell is taking highly aggregated national accounting identities and trying to turn them into things of causal significance. He implies that given a structural current account deficit, it is better that the government run a deficit than the domestic private sector. Why? Because if the private sector keeps accumulating debt it will eventually have to try to pay it down, whereas the government is not constrained in such a way, and can continue to accumulate indefinitely.
The trouble is, though, that there is absolutely no guarantee that a government deficit run deliberately to offset the structural current account balance would have the intended effect. By increasing domestic demand it is actually likely to have precisely the opposite effect. Factor in the inevitable reactive capital flows and movements in the exchange rate, and who knows what the ultimate effect would be (well, probably an increased current account deficit).
On a side point, Mitchell is wrong to say that “the non-government sector cannot fulfil its tax obligations unless the government has spent first”. Currency also enters private sector balance sheets via central bank activity. For example, in recent years of surplus, and consequence dearth of government debt as a bank ‘position-making instrument’, the Reserve Bank of Australia has operated primarily by entering into repurchase agreements with private banks, temporarily supplying currency when necessary. There’s no reason why this couldn’t go on indefinitely. (Not, of course, that I think it ought to – as noted previously, it is not my position that the government always needs to pay off all its debt.)
5. Finance, the Jobs Guarantee, and inflation
Mitchell points out that his costing of the Jobs Guarantee plan requires a deficit much smaller than the deficit we are presently running. This is a fair point, (though remember that he also implied that the government ought to run deficits matching the structural current account deficit, implying quite a different level of structural government deficit).
As I said in my original post, it’s not the Jobs Guarantee I have a problem with, it’s Mitchell’s idea that the government is completely unconstrained by budgetary concerns. If the Jobs Guarantee policy idea was uncoupled from these monetary eccentricities it might actually find more support, instead of Mitchell’s followers continually finding themselves butting up against a crazy world that doesn’t understand how bloody brilliant – and yet how utterly commonsensical – it all is.
In the right conditions, I would support a government trying out something like the Jobs Guarantee, which is basically a large extension of government employment at the minimum wage to mop up unemployment. But while Mitchell expects it to be a stable remedy for unemployment, I would expect it to be economically destabilising. The reason is that capitalism relies on a certain level of unemployment to discipline wages. The Jobs Guarantee is designed to be non-inflationary by setting its wage at the minimum wage. But I think even then it would remove much of the sting from unemployment, make employed workers feel more secure, and embolden labour.
Great, you might say, and I would agree – which is why I would support the policy. But it would destabilise in an inflationary direction, and present the government with choices it faced at the end of the full employment period in the 1970s: extend its controls in further reforms: price and wage controls, capital controls, public investment, etc., in other words completely politicising the economy; or abandon the policy. In the right political conditions, the Jobs Guarantee could be part of the opening salvo in a much broader social change. In the wrong political conditions it could be an abortive disaster.
I think the whole Mitchell system appeals to a certain kind of person, who is rightly sickened by a permanent pool of unemployment, sees policymakers as irrational, and thinks the whole thing could be put right, with no losers and no conflict, with a single Big Idea. As such, it risks being a Douglas Credit for the 21st century. I would rather see the energy joining a broader project which realises the extent to which capital shapes capitalist political possibilities, and recognises that there is no simple policy switch that would suddenly make everything rational without a serious confrontation with that power. Mitchell says that’s ‘quasi-Marxist’ and so it is!