Saturday, 11 January 2014

A short Saturday morning US jobs post

A quick run down of the US employment data and where it leaves this particular fool...

There are three charts to consider. The first shows falling unemployment rate, which is a function of the on-going decline in the labour force. That's been written about to death. At some point people will stop just leaving the labour force, but it hasn't happened yet. When they do, the pace of decline in the unemployment rate will slow, but the best estimate I have seen of the long-term trend growth rate of the US labour force is around 80,000 per month, which suggests that even yesterday's 74k increase in jobs is enough to keep the unemployment rate steady.

I've plotted it against the hourly earnings series, which has slowed to 1.8% y/y, and to 2% on a 3-month smoothed basis. Unemployment has fallen sharply but there's nothing pointing to any significant pick-up in wage growth yet. That will be yet another factor keeping consumer price inflation at bay.

The second chart shows the two measures of employment, the non-farm payroll series, plotted against the growth rate in employment from the household survey (the one that is used to measure the unemployment rate). Employment growth in the NFP series has slowed to 1.62% from 1.73%, and you'll excuse me I hope if I fail to get over-excited. The December payroll data were bad, albeit possibly excused by the weather, but the underlying trend in employment growth is close to long-term averages and still very consistent with a 'new-normal' economic recovery. Little has changed as a result of yesterday's soft headline and won't unless it is repeated for a couple more months. But the household survey shows employment growth of 0.8%, which is much worse. Even if I smooth it out, the gap is bigger than it has been since the late 1990s. Just looking at the two lines you can see that the (smaller) household survey is more volatile, but even so, the divergence is big enough that I for one will be watching it in the months ahead.

The last chart shows the NFP data, plotted against a 36-month standard deviation, i.e. a measure of the volatility of the series. Relative to recent norms, yesterday's number was a genuine outlier. If you glance across a the last period of low NFP volatility in the late 1990s, you can see that there were spikes, but they were followed by corrections. The lack of volatility in the US employment data is a source of optimism about this year's growth, because weather aside there isn't much which 'ought;' to cause a negative shock. In the 1995-1998 period, low employment volatility was accompanied by equity price gains, a strong dollar, and strengthening growth until the Fed 'blew it' by cutting rates and keeping them too low for too long after LTCM went bust -  re-fuelling the housing and dotcom booms. I take heart from both solid employment gains and the lack of volatility in the series, so a pick-up in vol would alarm me...

The jobs report poses questions about whether this was a one-month weather-induced outlier or something more meaningful. Most people will guess that it's the former, at least for a couple more months,. But the slowdown in wage growth either puts NAIRU even lower or says the wage/unemployment relationship is broken. Either way, the only inflation we should worry about is asset price inflation and the Fed seems happy to ignore that. The household survey meanwhile, needs to show employment growing faster or I'll start to really worry about that, too. And as a final aside, since no-one knows what to make of the jobs data, the big news was the fall in the US trade deficit as energy independence raises the attractive prospect of an economic recovery that is NOT accompanied by a widening deficit. In short, a recovery with much better balance between output and demand growth than we've been used to in the US. Oh, that we in the UK could dream of such balance in our recovery....

No comments:

Post a Comment