US employment tutorial 1: Five stats that give a good picture

28 Oct 2009 by Jim Fickett (revised 6 Mar 2010)

The economy is, by definition, people producing goods and services for each other. It follows that the employment situation is among the most important measures of the economy. The first set of reference pages put up on the site is on US employment and, beginning with this post and interspersed with articles on other topics, there will be a concise tutorial (in a total of seven posts) on

  • why the reference pages cover the statistics they do,
  • what you need to know in order to interpret the numbers, and
  • what the current trends are.

There is an embarrassment of riches in this area, with the government providing hundreds of statistical series from many different surveys, and private organizations also providing alternatives. It is important to narrow it down not only because one cannot follow it all, but because switching to a new data series every time there is an interesting anomaly is a very good way to let your hunches and prejudices get in the way of objective evaluation.

First, I cut out the sentiment surveys. Some of the factual series are rather delayed, or later revised, but several are quite timely, and the trend in the factual data is usually clear enough that sentiment is icing on the cake. Second, there are many series that make the headlines only because they come out a few days ahead of a more reliable statistic. But the employment situation changes on a scale of months; it is a waste of time to try to get ahead by a few days. Third, there are many series with overlapping content. For example, one of the main things one needs to monitor is to what extent workers are employed part-time, full-time, or more than full-time. Average hours worked, overtime hours, and the number of part-time workers, all give different important aspects of this facet of the market, but all move more or less in parallel and, at a macro level, it suffices to watch one.

The core measures are

  • “Non-farm payrolls”, the best proxy for a count of all jobs, based on a survey of employers
  • The “unemployment rate”, based on a survey of households
  • Some measure of partial employment; I use the number of those working “part-time for economic reasons”, i.e. having a only a part-time job while wanting a full-time job

It takes a little time for the core measures to reflect changing conditions. There are two other measures which are very often used as leading indicators:

  • The number of “initial unemployment claims”, which gives a nearly real-time count of new layoffs
  • The number of those in “temporary help services”, which is an early indicator because companies hire and fire temps more readily than they do those in regular, full-time positions

This gives a mixture of rapid-response and more comprehensive statistics, from several sources, gathered by differing methodologies, and covering the main aspects of employment.

Many people would want to add more to this list. But investing is not like reporting on one specialty area or studying one aspect of the economy in academia. To invest well you need to understand the main trend in each of perhaps 50-100 areas, so it is important to keep what you follow in each area as simple as possible (though no simpler). In my view these five statistics give a solid understanding of the employment situation, and it is based on these five that the US employment reference page, and its subsidiary pages, are updated each month.