Figures next Wednesday are likely to show that unemployment is approaching the 7 per cent threshold which the Bank of England adopted last summer. This doesn't mean Bank rate will rise automatically; Bank governor Mark Carney has emphasised that the 7 per cent rate is a threshold, not a trigger. Whether it does so depends upon the Bank's assessment of whether lower unemployment will raise inflation. And this raises a problem.
If we look at data since the start of 2000 - when the ONS's current data on wages begins - there is a negative correlation (of minus 0.55) between the unemployment rate and wage growth in the next four quarters. That suggests that lower joblessness raises wage inflation, just as the Phillips curve tells us.
However, if we look at a scatter plot something curious emerges. We see two clusters of points. There's one - the pre-crisis period - with unemployment and wage growth both generally between 4 and 6 per cent. And there's another - since the crisis - with unemployment over 7 per cent and wage growth under 2.5 per cent. The negative correlation between unemployment and wage growth exists only if we compare these two clusters. If we look at each separate cluster, there's no correlation.
This is an example of Simpson's paradox - a strong relationship in the aggregate data does not always mean there's a strong relationship within subsets of that data.
This poses a puzzle for the MPC. Does the fall in unemployment mean we're moving back to the pre-crisis cluster or not? If it does, then wage inflation will rise and so there might well be a case for higher rates. If not - if we're in a new cluster instead - then there isn't.
So far, there's no evidence for the former view. Although unemployment fell from a peak of 8.4 per cent to 7.8 per cent between late 2011 and late 2012, wage inflation subsequently actually fell. And there are four reasons to suspect it could stay low even if unemployment continues to fall:
■ Memories of the great recession will have a scarring effect, by making workers scared to push for higher wages.
■ Public sector pay restraint will spill over into the private sector, in part by giving private companies a potential supply of disgruntled public sector workers.
■ The unemployment rate understates the true excess supply of labour. There are 2.3m economically inactive people who say they'd like a job - almost as many as there are officially unemployed. And their preference isn't an idle one. In the last three years, 146,000 people have moved from inactivity into work every month on average. This potential supply of labour will bid down wages.
■ The labour market is now globalised, so wages are held down by cheap labour in the Far East.
The question for the MPC is: how powerful or long-lasting will these effects be? Given that there's so far no sign of wage inflation rising, I suspect it will wait several more months before raising rates. Which means savers must tolerate negative real interest rates for a long time yet.
