This is, of course, anathema to dedicated believers in the omnipotence of central banks. But Krugman is in good company. I recently heard Richard Koo speak on lessons from Japan for the Eurozone. Koo questioned QE's effectiveness when the private sector is refusing to take on more debt because it is determined to deleverage.
Koo's and Krugman's scenarios are actually the same, though they attack the problem from different angles. In both cases, interest rates are zero, though Krugman explains this as an infinite demand for interest-free money (the liquidity trap), and Koo explains it as a lack of demand for borrowing. Both argue that central banks are unable to generate inflation when the private sector does not wish to spend. Both insist - though for different reasons - that when interest rates are zero, fiscal stimulus is needed to get the private sector to spend.
Sam Bowman thinks this is absurd:
If Krugman is right that central banks can't always cause inflation, let's abolish tax and fund govt by printing £££ http://t.co/SJUHrAM4wG
— Sam Bowman (@s8mb) December 17, 2014
Let's test out Sam's theory. Imagine a country in which there is a fiat currency but no central bank. Money is directly created by government and paid to the population as a basic income via commercial bank accounts in a free banking system (Sam will like this!). People spend that money on goods and services, enabling businesses to flourish and creating jobs, which in turn enable people to top up their basic income with earnings. People get richer, and they spend more. The government keeps on printing money to provide the basic income, even though people are earning more and more from their jobs. Soon the place is awash with money. People can buy all they need and save as much as they want, and still have money left over. It's an earthly Paradise.
Except that it isn't. When everyone has everything they could possibly ever need, money is worthless. Money is only valuable to the extent that it is scarce relative to goods and services not only now, but for all time: as long as there is (relative) poverty, and therefore demand for money, money has value. Inequality is necessary if money is to maintain its value. This is why measures to reduce inequality, such as transfer payments from rich to poor, can be inflationary. But when the government prints more and more money, dishes it out as helicopter drops to everyone equally, and never taxes any of it away, goods and services become scarce relative to money, and therefore more valuable. If people start to believe that the supply of money is infinite, goods and services become infinitely valuable. Please note that this is NOT the same as Krugman's liquidity trap. In a liquidity trap, the demand for money is infinite, not the supply. What I am describing here is Zimbabwean hyperinflation. Central banks may not always be able to generate inflation, but governments can.
So to control inflation, the government needs a mechanism to drain money from the private sector. It could do what central banks do, namely sell assets. But what assets does it have? It might have some physical assets - schools, hospitals, roads, that sort of thing. But once it has sold those, what then? Well, it could issue bonds. We aren't used to thinking of government bond issuance as a monetary tool, but there is no reason why it should not be. If a government issues bonds but does not spend the money received - as researchers at the Bank for International Settlements have suggested - then money is drained from the private sector. Government bond issuance is monetary tightening. Private sector saving in the form of government deposits is also monetary tightening for the same reason.
In this scenario, there is no need for taxes. Government spends using newly-created money, then neutralises that spending by issuing bonds to the private sector. Interest rates are high enough to ensure that the private sector voluntarily buys enough of them to sterilise the Government's spending - though if the private sector's propensity to save is high, as in Koo's scenario, this could still mean interest rates at or close to zero. It's entirely circular and not inflationary. In theory, therefore, government spending could be entirely sterilised with bond issuance and/or private sector deposits (are you listening, Japan?). Note that spending comes first, just as bank lending precedes deposits. Governments really can act like banks.
However, it is entirely possible that the private sector might refuse to buy the bonds. After all, they are only a promise that some future basket case government will redeem them by printing worthless money. Without something a bit more definite than "we really will pay you this money, honest", the private sector might balk.
This is where taxes come in. In a fiat currency system, taxes are a means of forcing people to return money to government. They are, if you like, the equivalent of required reserves for banks. Everyone must place a certain proportion of their income on deposit at government: the greater their income, the more they must place on deposit. People who refuse to maintain these "required reserves", or who divert them to other purposes, can be fined or jailed. In this way government ensures that it can always sterilise its spending.
In our mythical fiat-currency, free-banking, money-financed country, taxes - or at least the threat of taxes - would be necessary to control inflation. Just as a credible central bank standing ready to buy assets sets a floor under the price of those assets, maintaining their value, so a government that can credibly impose taxes on its population maintains the value of the currency, both now and in the future (this is why the power to tax also supports the price of government bonds). A defining feature of hyperinflationary episodes is that government loses the power to tax. It can print money, but it can't recall it. Central banks use interest rates as a proxy for taxes, but if the government loses its credibility then all too often the central bank does too - in which case interest rates lose their potency. In 1998, interest rates at 160% failed to restore the Russian currency or prevent government default.
The point of this rather fanciful example is that a credible government can in theory finance its own spending without a central bank. "Borrowing" and taxes are both ways of sterilising money financing of governments. Governments do not need central banks to control inflation. They can control it themselves.
So why is money financing of government feared? It is feared precisely because it is far more likely to be inflationary than money creation by an independent central bank. Krugman is right. Central banks simply are not as effective at creating inflation as governments. If they were, we wouldn't entrust them with management of the money supply. After all, as both Russia and the Eurozone are demonstrating at the moment (though in different ways), central banks are only as independent as politicians allow them to be, and only as credible as the governments that back them. If government can't be trusted to manage the money supply without causing inflation, then neither can a central bank - unless it is simply unable to create significant inflation.
But why are central banks so much less effective at creating inflation than governments? I think it is largely due to the way in which the money they create is distributed.
Government spending and taxation directly affects the behaviour of ordinary citizens. Ordinary citizens are by-and-large financially constrained: they have shortages of money relative to their desire for goods and services, and they have limited borrowing capacity. Therefore, if a government creates money and pays it to ordinary citizens, the likelihood is that a fair number of those will spend it. Equally, if government increases taxes - or people expect that it will - many people will cut their spending. True, people might not adjust their spending if they think that tax and spending changes are temporary: but then banks and investors might not adjust their behaviour either if they think monetary expansion or contraction is temporary. Ricardian equivalence applies as much to central bank operations as it does to taxation.
Clearly, if government money-financed spending is likely to flow through directly into increased consumption by ordinary citizens, it can very easily lead to consumer price rises. Tax changes can be too slow and unwieldy to counteract this inflationary effect - and they are unpopular. The real problem with money financing of government is that politicians depend on voters for their jobs, and giving money away is far more popular with voters than removing it, even if removing it is in those voters' best interests. But the same could be said of central banks. Removing the punchbowl when the party is in full swing is equally unpopular whether the party is on Wall Street or Main Street.
Central bank monetary operations depend for their effectiveness on banks and investors being willing to adjust their behaviour. Unlike ordinary citizens, banks and investors do not generally suffer from shortages of money. Their behaviour is driven by their appetite for risk and their desire for return. Monetary policy operations aim to influence their portfolio choices by adjusting yields on certain classes of asset. The main effect of central banks' activities is therefore seen therefore primarily in asset prices, rather than consumer prices. There may be some impact on consumer prices via interest rate effects, wealth effects and the famous "hot potato effect", but it will be much less than the impact of an equivalent fiscal stimulus. It is, frankly, inefficient. To be sure, a central bank supporting asset prices by means of large-scale purchases can interrupt a deflationary spiral: but then a government supporting house prices by means of guarantees and tax breaks can ward off a property market collapse, which might prevent a deflationary spiral forming in the first place.
Ambrose Evans-Pritchard argues that central banks can always generate inflation if they try hard enough, and cites Friedman's famous "helicopter drop" as evidence. But central bank helicopter drops are fiscal policy, since the money goes directly to ordinary citizens rather than to the financial system. Why not simply unchain the fiscal authority so it can do money-financed deficit spending, rather than getting the central bank to do it and calling it monetary policy? When the economy is in a slump, no-one is spending and no-one wants to take any risk, does it really matter whether the central bank or the government reflates the economy? Where did this absurd idea come from that the only good stimulus is a "monetary" one?
Rather than using the loaded terms "monetary" and "fiscal", we should talk about "indirect" and "direct" stimulus. Put like this, Krugman is once again clearly right - as, incidentally, is Friedman. Direct stimulus is obviously far more effective than indirect. How on earth could we possibly think otherwise?
Fiscal pessimism - Pieria
Some incomplete monetarist arithmetic - Pieria
Structural destruction - Coppola Comment