Join our community of smart investors

The US inflation surprise

Investors should worry less about US inflation, and more about the stock market's nervousness
February 20, 2018

Economists were surprised last week by higher-than-expected US inflation. From a longer-term perspective, however, the surprise is not that inflation has risen but that it was so low in the first place.

Since the mid-1990s, you could have forecast annual core US CPI inflation to within around a third of a percentage point, simply by looking at the unemployment rate. High unemployment led to low inflation and low unemployment to high; the main exception to this was higher-than-expected inflation in 2010-11, perhaps because the recession created a mismatch between the unemployed and the skills companies wanted. There was a reliable Phillips curve. 

Last year, though, this relationship wasn’t so useful. For example, in January 2017 the unemployment rate was 4.8 per cent. The 1996-2015 Phillips curve predicted that this would lead to core inflation of 2.25 per cent. Last week's numbers showed it to be 1.85 per cent. Yes, that was more than economists expected a few days ago. But in a longer-term context, the surprise is how low inflation is, not how high. 

You might think that this suggests a twin danger: that the old Phillips curve will reassert itself and the inflation undershoot will disappear; and that current low unemployment will push up inflation further.

These are only risks, however, not certainties. There are good reasons why inflation might be less sensitive to unemployment now than in the past. For example, a more atomized workforce is less able to parlay labour shortages into pay rises.

Even if the risks do materialise, however, inflation won’t rise very much. The current 4.1 per cent jobless rate combined with the 1996-2015 Phillips curve points to an inflation rate of only 2.4 per cent next January. Although there’s a close link between unemployment and inflation, it’s not a sensitive one. The mere fact that inflation has been stable in the last 20 years while unemployment has varied a lot tells us that inflation doesn’t vary much with unemployment – or with anything else for that matter.

So, what’s the problem? Two things.

First, if the Fed is to keep inflation down, it must raise interest rates by more than the rise in inflation. The Taylor rule says that a percentage point higher inflation requires a 1.5 percentage point higher fed funds rate. If the old Phillips curve returns, this points to the rate rising by almost a percentage point this year.

Secondly, there’s the issue of the stock market’s likely response to this prospect. For most of the last 20 years there’s been a close correlation between output growth and equity returns. This tells us that if the economy grows well – and it’s not so sensitive to monetary policy that rate rises in themselves will prevent this – then equities can withstand higher rates. It’s possible, however, that this relationship will break down. If ultra-low rates have boosted share prices because investors have 'reached for yield', the withdrawal of those rates might have disproportionately adverse effects on the market. Even if this proves eventually not to be the case traders might worry that it will be – or worry that others will worry. This raises the danger that the road to higher rates might well be a rocky one for shares.

But let’s be clear. Investors’ problem is not inflation or even interest rates. It is uncertainty about the market’s possible reaction to these that should worry us.