You are here:

How low can they go?

Created:
18 November 2008
Written by:
Chris Dillow

What is the shape of the Phillips curve? If this sounds like a pointless academic question, let me assure you it's not. It holds the key to how far interest rates can fall.

Advertising

The Phillips curve describes the fact that higher unemployment leads to lower inflation. Our table shows how it has changed since the 1970s; it shows the results from simply regressing retail price inflation upon the claimant count unemployment rate 12 months previously.

The changing inflation-unemployment trade-off
Intercept Coefficient Unemployment required for 2.5% RPI
May 97 - now 4.15 -0.39 4.2
Oct 92 - May 97 4.67 -0.22 9.9
May 79 - Oct 92 17.45 -1.36 11.0
Jan 71 - May 79 20.02 -2.00 8.8
Regression of RPI inflation on unemployment lagged 12 months

Three things stand out here. First, it's always been the case - on average - that higher unemployment leads to lower inflation. The coefficient on unemployment has always been negative.

Second, this trade-off has improved since the 70s. Since 1997 an unemployment rate of 4.2 per cent, on average, has been sufficient to keep RPI inflation at 2.5 per cent. Back in the 80s, it would have required 11 per cent unemployment to do this.

Third, the trade-off has flattened. In recent years, a one percentage point change in unemployment has had less effect upon inflation than it did in the 70s or 80s.

This is a mixed blessing. In good times, it's fantastic. It means unemployment can fall a long way without igniting much inflation. But in bad times, it's a problem. It means we need a big rise in unemployment to bring inflation down.

Which is why the Phillips curve matters for interest rates. The less inflation responds to this recession, the less justification there will be for further cuts in interest rates.

The post-1997 Phillips curve implies that we need another 400,000 rise in unemployment - to a rate of 4.2 per cent - merely to ensure that inflation hits its target*. But target inflation only justifies Bank Rate being around 5 per cent. It's only to the extent that unemployment looks like rising more than this that we can justify the rate cuts we've already had, let alone further ones.

Except for one thing. Our table shows that the Phillips curve relationship is not stable. Which raises the possibility that it could change again. Simon Hayes of Barclays Capital says there is a "significant risk" that the curve will be steeper than recent experience suggests. If so, recession could have a surprisingly big disinflationary effect.

Most of the evidence for a flattish curve lies in the fact that unemployment fell a long way in the 90s without raising inflation much. This might, says, Mr Hayes, have been the result of some one-off improvements in the labour market, such as tax and benefit reform intended to get people into work. It doesn't follow from this, however, that rising unemployment will have little effect upon inflation.

What's more, the flat curve might be just an artefact of the age of stability that has now ended. From the mid-90s until a few months ago we all expected inflation to vary only a little around 2 per cent. So it did; inflation expectations are self-fulfilling. In our new era of greater volatility, however, the possibility has emerged of global deflation. This means it's possible that rising unemployment will be accompanied by sharp falls in inflation expectations and, therefore, in actual inflation.

These, though, are not the only reasons to expect the unemployment-inflation trade-off to steepen.

Another lies in the fact that shipping costs have collapsed. The Baltic dry index has fallen 93 per cent since May, to its lowest level for seven years.

To see why this matters, imagine shipping costs were so astronomically high that imports were impossible. The unemployment-inflation trade-off would then be horrible, because domestic companies could raise prices without fear of losing business to foreign imports. It follows that as transport costs come down, inflation should fall for given unemployment. This is because, at the margin, lower shipping costs mean UK markets are more contestable by foreigners - which should limit the ability of UK firms to raise prices.

Another force for lower inflation was evident in the recent decision by JCB workers to take a pay cut in a (vain) effort to avoid job losses.

A few years ago, many economists thought such a choice unlikely. They thought that workers with skills and seniority would refuse pay cuts in the belief that their jobs were safe. Such "insider" power meant, they said, that unemployment wouldn't force pay down by much.

But this might have changed, and "insiders" are fearful of unemployment. If so, wages - and therefore inflation - might be more responsive to unemployment than previously thought.

All this suggests rising unemployment could have a bigger impact on inflation than we'd expect from post-90's experience. So perhaps the Bank is right to cut rates so much.

Except for two things. First, one contributor to the flat Phillips curve since the 90s has been that there's often been a positive correlation between economic activity and sterling. Falling unemployment has been accompanied by a rise in the pound that has held inflation down. And rising unemployment has been accompanied by a fall in the pound that adds to inflation. There's little sign yet of this changing.

Second, another reason for the flat curve has been the flood of cheap goods from China and other emerging economies. But, with these economies slowing, this supply will slow. That could add to inflation for any given level of economic activity.

The message here is simple. We just cannot be sure precisely what the unemployment-inflation trade-off is. The knowledge policy-makers need is always available only after they've needed it. This means that there is a high risk the Bank will set rates wrong - which in turn adds to the uncertainty about inflation, at least in the longer run.

For the near term at least, though, two things are unusually certain - the directions in which unemployment and inflation are heading.

*In ordinary times, 2.5 per cent RPI inflation is consistent with 2 per cent CPI inflation.

  • Back to top

Win £1,000 or a golf weekend

Enter IC's Fantasy Portfolio competition for the chance of winning £1,000 or a luxury golf weekend, courtesy of ETX Capital and Celtic Manor.

Enter here for your chance to win

Advertorial Feature

Take control of your investing with CFDs

Have you ever watched a move in the markets that you saw coming, but you weren't able to exploit?

by Dominic Picarda

Advertorial Feature

Spread your risks with spread trading

With so many big moves in the world's financial markets, there have seldom been more opportunities around for spread traders. Isn't it time you joined them?

by Dominic Picarda