It might seem a strange thing to say when next week's figures are likely to show that core CPI inflation is at a six-year low, but the US has an inflation problem - and it's one that could hold share prices down for years.
The problem is that inflation is higher than it should be, given the amount of spare capacity in the economy.
My chart shows the point. It shows that since the mid-1990s there has been a very strong relationship between core CPI inflation and the unemployment rate 12 months previously; high unemployment leads to low inflation, and low unemployment to high inflation. In the past few months, however, inflation has been higher than this relationship predicted a year ago. In March 2009 the unemployment rate was 8.6 per cent. The 1998-2009 relationship between unemployment and subsequent inflation predicted that inflation would be 0.7 per cent in March of this year. In fact, it was 0.5 percentage points - 1.8 standard errors - higher than this.
Economists expect this deviation to get worse. Unemployment was 10 per cent in December. This means the 1998-2009 relationship predicts core CPI inflation of just 0.1 per cent in December 2010. However, forecasters surveyed recently by the Philadelphia Federal Reserve expect inflation then to be 1.5 per cent, with a less than 10 per cent chance of it being as low as 0.1 per cent.
What's happening here? We can ignore the possibility that overly loose monetary policy is stoking up inflation. Most stories of how monetary policy causes inflation say that it does so by boosting aggregate demand, which should mean that unemployment falls. Something else, then, is going on.
One possibility is that the Phillips curve - the trade-off between unemployment and inflation - becomes flat at low inflation because there is, in effect, a floor under inflation.
The story here starts from the correct premise that low inflation requires that some prices fall a lot. It then says that many prices are 'sticky"'; sellers are loath to move them. This stops some prices falling, which puts a floor under inflation. The effect of this is that at low inflation rates, inflation stops falling so a drop in aggregate demand leads not to even lower inflation, but to higher unemployment.
This theory, however, runs into a problem - it's not clear that prices are 'sticky'. If they were, all that talk in 2008-09 about the risk of deflation would have been utterly wrong. Back in the mid-90s, Mark Bils and Peter Klenow, two US academics, studied a huge number of individual prices and found that they moved a lot; half of prices, they estimated, changed at least once every 21 weeks. And there's plenty of evidence of unsticky prices in the latest inflation numbers. Prices of curtains, for example, have fallen 11.3 per cent in the past 12 months; motel bills are 5 per cent down, prices of women's dresses are 8 per cent down (there are no data on men's dresses); and prices of TVs have fallen 29.1 per cent.
So, can we dismiss the sticky-price theory as outmoded Keynesian nonsense?
Maybe not. Many other prices are sticky. Jeffrey Campbell at the Chicago Federal Reserve has found that the chances of a price moving decline if it hasn't moved recently. This suggests that firms have a 'suck it and see' approach to pricing; they change prices often while they are testing the market, but once they've found a tolerable price, they stick with it. And other researchers have found that the prices that do change a lot tend to do so not because of macroeconomic factors such as boom and slump, but because of sector-specific reasons. TV prices, for example, have fallen sharply not so much because the economy is weak, but because there's been rapid technical progress in the the industry.
The evidence for a sticky-price explanation for a flattening in the Phillips curve is, then, mixed.
Instead, there might be another, much darker, process at work. Maybe the Phillips curve is shifting rightwards - the trade-off between unemployment and inflation is worsening. If so, the next few years will see higher unemployment for any given inflation rate (or higher inflation at any particular unemployment rate) than we have seen in recent years.
Such a move would reverse the leftward shift in the Phillips curve that occurred in the early 1990s; there's a reason why the data in my chart begin when they do.
Economic forecasters believe this is what is happening. The Philadelphia Fed survey found that the consensus expects core inflation to rise to 2 per cent by the end of 2012, even though it expects unemployment to still be around 9 per cent in 2011.
But why might the Phillips curve shift rightwards? The answer lies in some kind of adverse supply shock - the sort that makes it harder for firms to produce things cheaply with the result that we get higher inflation and lower economic activity.
Casey Mulligan, a professor at the University of Chicago, suggests one supply shock. He believes people have become less willing or able to work, and employers have become less willing to hire. There are several reasons for this: the minimum wage rose in 2008; people who are unable to sell their houses in a depressed market cannot move to where there are jobs; and the mortgage modification scheme gives some people an incentive to work less. The combined effect of these is that unemployment has risen without cutting inflation.
There's another adverse supply shock - the reduced availability of bank credit; loans to commercial and industrial firms have fallen by 17.9 per cent in the past 12 months.
To see how this matters, put yourself in the shoes of a businessman in an ordinary downturn. He sees high unemployment and its correlates - underemployed workers and cheap premises - but also the possibility of economic recovery. He might think: "With so much spare capacity, it's cheap for me to expand my business, so I'll cut prices and try to win more customers." If enough entrepreneurs think like this, at least some prices fall and so inflation falls after a period of unemployment. This is what normally happens.
Now, though, things are different. Our entrepreneur might well think: "I'd like to expand, but I can't get a loan from the bank. And there's no point me cutting prices to win more business if I can't meet those extra orders, so I'll keep prices where they are." Inflation, then, doesn't respond to high unemployment.
What we don't know for now is whether these adverse shocks will persist or not. It's possible they won't. If lower house prices cause some workers to postpone retiring, and others to come out of retirement, there will be a favourable labour supply shock - one that will bid down wages and hence prices. And it's possible that banks will eventually start lending again.
There is, however, one thing we can say: that if the Phillips curve has indeed shifted rightwards, it would be bad news for shares.
History tells us as much. Between January 1998 and January 2009 - when the Phillips curve was favourable - the price-earnings ratio on the S&P 500 averaged 25.4, despite a bear market and financial crisis during that time. However, in the previous 10 years, when the Phillips curve was less favourable, the PE ratio averaged just 18.7.
A bad Phillips curve, then, means low stock market valuations.
The reason for this is straightforward. A bad Phillips curve means the economy cannot grow strongly without generating inflation. And this gives investors a dilemma. If you expect strong aggregate earnings growth you must also expect higher bond yields. But higher bond yields mean that future earnings are discounted more heavily. Yes, we can avoid a high discount rate - but only by suffering lower growth. Either way, share valuations must be low.
Of course, you could interpret the market's current modest valuation - a PE ratio of 17.5 - as a sign that investors are pricing in a worse unemployment-inflation trade-off. And, of course, a low long-term average PE ratio is quite consistent with some nice short-term rallies.
Nevertheless, the fact remains. Latest figures, plus economists' forecasts and market opinion, all point to the same thing: to the possibility that the financial crisis has done lasting damage to the US economy by worsening the trade-off between unemployment and inflation. And this is not good for shares.
MORE FROM CHRIS DILLOW...
Chris blogs at http://stumblingandmumbling.typepad.com