This has been bugging me (and therein you get a glimpse of how dull my mind is).
A day or two ago, Scott Sumner explained extremely well how you need a permanent increase in the money supply to generate inflation. If the economy knows that it will be withdrawn in the next period, then money will be hoarded until that period. Therefore nothing happens: it is not spent in the way it needs to be to get the price level up. Sumner puts it better:
Now consider the temporary currency injection. We know that the long run price level doesn’t change. Once the money supply returns to the original level (a year later) prices should also return to the original level. At first you might assume that monetarists would claim that the price level would double, and then fall in half. But consider the real interest rate. Monetarists typically assume that real variables like the real interest rate aren’t much affected by monetary shocks. But if prices doubled, and then were expected to fall in half, the expected real return on currency would be 100% in year two. That is, the purchasing power of money would be expected to double in year two. More likely, almost all of the temporary currency injection would be hoarded and prices would rise by just one or two percent—the risk free real rate of return
Many of you will notice that this is similar to what has happened in Japan (I think Krugman here is enlightening on that situation)
This is all very pertinent to QE, and our current situation. In brief, our current authorities are so nervous about being called inflationists that they bend over backwards to insist on the temporary character of QE. Observe the first question that Mervyn King fielded on this before Parliament:
Q7 Sir Peter Viggers: Quantitative easing. Why is buying assets better than printing money and throwing it out of a helicopter?
Mr King: Because we get the asset.
And you get to sell assets back. It may mean, as Duncan argued, that it may be failing.
Now, I am coming round to the idea that Japan, 1997 is such a bad place to be that a little extra inflation may be a least bad option. But I remain sceptical that Tim Leunig’s idea of a five year burst of managed inflation can be actually achieved. His idea is attractive because it retains some democratic accountability for what is, after all, a huge change: a massive redistribution from debtors to creditors (HT Buiter). Sumner’s post helps elucidate why. Tim writes:
Since it takes about two years for central-bank policy fully to influence inflation, a sensible policy would be to target 4 per cent inflation for the five years from 2011, followed by 2 per cent thereafter. In Britain, the government would simply redefine the Bank of England’s inflation target . . . An increase in inflation by an extra 2 percentage points for a period of five years would have many benefits – for governments, companies, households and the banking system.
I have no doubt of the benefits. What I doubt is whether in a world of forward-looking agents, hoardable-money (or investable in risk-free things), and a crucial role for expectations in setting real conditions, such a managed up-and-down can be done. After all, current spending is largely influenced by expectations of future inflation. If in period 1, agents know that in period 2 inflation will drop suddenly, will they not change their behaviour in such a way that damages the simple quantity-price relationship in period 1?
Wolfie Munchau said it too:
Of course, it can be done, but only for as long as the commitment to higher inflation is credible. Inflation is not some lightbulb that a central bank can switch on and off. It works through expectations. If the Fed were to impose a long-term inflation target of, say, 6 per cent, then I am sure it would achieve that target eventually. People and markets might not find the new target credible at first but if the central bank were consistent, expectations would eventually adjust . . . If, however, a central bank were to pre-announce that it was targeting 6 per cent inflation in 2010 and 2011, and 2 per cent thereafter, the plan would probably not succeed. We know that monetary policy affects inflation with long and variable lags. Such a degree of fine-tuning does not work in practice.
And it’s obviously even harder when there’s a Crunch on . . . This graph is “expectations of rates on mortgages, bank loans and savings from the Bank:
Despite the purpose of QE being partially “shock and awe” – “we’re doing everything we can” – only for ONE QUARTER of the past few years has the expectation ahead been of lower rates. Some bazooka.
So, my wonkish question is this: Sumner is clearly correct about the (in)ability of a temporary change in Quantity of Money to change price levels, IMHO. Is the reasoning also applicable for the first derivative of price levels?
UPDATE: When people get more cash in their pockets, do they spend it? Depends on their expectations and animal spirits (H/T, Keynes . . . ). Currently, it looks like they are paying down equity in their houses (BOE figures)
ANOTHER UPDATE: There seem to be few commentators out there willing to brave the complexity and confusion of monetary statistics. Ambrose Evans-Pritchard seems to be one of them, linking to that MoneyMovesMarkets blog too. Where I really agree with him is:
What is clear is that zero rates are playing havoc with the indicators, so nobody really know what is going on — and that includes the central banks. You have to trust your instincts here. We are in the field of psychology and anthropology.
Yes – the only people you should doubt should be the people who seem certain