Keith Fray Non-farm payrolls explained

For market traders, economists, and data geeks alike, Friday is one of the highlights of the month – non-farm payrolls day.

For the uninitiated this is the release of data on US jobs growth over the previous month – more properly called the Employment Situation report - published by the Bureau of Labor Statistics, usually on the first Friday of the month following the data (i.e Friday’s new data will be for August).

It is undoubtedly the most eagerly awaited monthly data by world markets and has attained a totemic status, perhaps beyond its real importance. Morning trading volumes are slim in European markets on the day of release as they await the afternoon release time (8.30am Eastern Time in the US).

Why do non-US markets care so much? Well if China continues to grow at current levels then the US will surrender its status as the world’s largest economy in the next decade (and probably in the current decade if measured in purchasing power parity terms). For now though, the US remains the bellwether of the world economy, accounting for a fifth of global output.

Should we care as much as the markets seem to? How important are these numbers? What should we be looking for?

The ‘headline’ number in the report gives the number of jobs (obviously excluding farms) that the US economy has gained or lost in the previous month. In July, for example, total US employment was 163,000 higher than the June level.

A more important number for the markets, however, is the difference between the actual change and the consensus of estimates by analysts and economists published in advance (available in Monday’s FT in the table on the back page or at the economic calendar). Market reaction will depend on whether the published figure is above or below the consensus, and by how much.

As well as new data the publication also includes revisions to the two previous months. A recent study found that over the period 1964 to 2011, the first estimate was revised up, on average by 18,000 over the next two months.

However some revisions are much more substantial, in either direction, calling into question the wisdom of any significant market reaction based on a single (preliminary) figure.

Markets used to be concerned about the likely future direction of interest rates – a disappointing figure might presage a cut to boost growth, while if jobs growth was higher than expected analysts might look for the Federal Reserve to hike rates, to cool an overheating economy.

Now, with conventional monetary policy all but exhausted, speculation is more about the prospects for an extension to quantitative easing.

To gauge the true direction of the US labour market it’s necessary to delve deeper than one number – and the monthly report contains a huge amount of detail – employment by industry, geography, age, sex and race, as well as unemployment and hours worked.

Let’s, for example, compare some data over the longer term. The chart below shows the percentage change in non-farm employment from the low point of each recession since 1945.

The latest downturn bottomed out in July 2009 according to the National Bureau of Economic Research. Over a comparable period, this recovery is one of the weakest – only those following the 2001 recession and the huge shake-out of old industries in the early 1980s are worse.

It’s the nature of markets that they demand an instant response to new information. The initial knee-jerk reaction is often reversed as traders take stock of the whole picture. Regarding data, markets tend to respond to ‘waves’. Proper analysis demand that we look at ‘tides’.

That said, the consensus for tomorrow is for a rise of 120,000 during August. Stand by your computers at 8.30 EDT!