The Continuing Rise of Temporary Work

I made a short presentation on the disturbing rise of temporary work last week. It seems the cutting edge of the new normal is to be found in our schools, colleges and universities.

As most of us would know or strongly suspect, paid work has become more casualized or precarious over time as the standard employment relationship of full time, permanent paid work for an employer has been eroded as the norm.  In 2010, just 64.0% of men and 59.5% of women were still in such jobs.  The others were in part-time jobs, or temporary jobs, or in self-employment.

Temporary work is the most obviously precarious category. By definition, it is employment that comes to an end, either because it is seasonal work, or because it is contract work with a pre-set end date. Many temporary jobs are also part-time (one third of such jobs for men, and half for women.) While some people do seek temporary jobs, they are an inferior choice for most and are less well paid and provide limited if any access to benefits.

The incidence of temporary work for men rose from 1997 to 2007, the last full year before the recession, from 8.8% to 10.0%. For women, the increase was from 10.1% to 12.0%. For both, the increase came before 2005 and then fell slightly.  As one might expect, employers seem to have been obliged to offer more secure jobs when the national unemployment rate fell below 7%.

Between 2007 and 2010 – the years of recession and partial recovery – the incidence of temporary work again rose, from 10.0% to 10.5% for men; and from  12.0% to 12.1% for women.  That does not sound dramatic, but the majority of job creation at the margin has been in temporary jobs.  In fact, two-thirds of the net job increase for men 2007-10 was in temporary jobs, and one sixth of the increase for women.

(I am looking at changes in annual averages since the monthly data are not seasonally adjusted. If I compare December, 2007 to December, 2010, I find that less one third of net paid job growth was in permanent jobs.)

What some might find surprising is that temporary jobs are disproportionately to be found in the public sector, which accounts for 25% of all employees, but 31% of temporary jobs, and 40% of all contract jobs. Within public services, the incidence of  contract work is extraordinarily high in educational services which accounts for 8% of all workers, and 22% of all contract jobs. The incidence of temporary work is just about average in public administration and health and social services, though if one just looks at contract work the incidence is a bit higher than average.  

As one would expect, temporary workers tend to be much younger than average, and, less predictably, are somewhat more likely than average to be unionized.  

If the experience of the 1990s is anything to go by, a lot of new hiring in a weak recovery will be into temporary jobs. Temporary work will likely be the norm for younger workers leaving school and entering the full-time job market, and they have to wait a long time to get into permanent jobs.

The tilt of new hiring towards temporary jobs underlines just how much slack there is in the job market.


  • December may be a bad month to use because of the sigificant number of temporary workers hired for the Christmas month would distort the results, even year over year.

  • Yes Darwin. But looking at the same month year over year is still meaningful sinc the same swasonal factors are at play. All in all though its best to use annual averages to weigh any changes in temp work,

  • The seasonal factors can change over time, too. That’s why it would be better to use a month where the seasonal effect isn’t so large.

  • Nice work Andrew.

    and this is just one component that is feeding into the precarious dynamic of work.

    Given the great recession, I am quite sure the seasonality algorithm may actually be doing a lot of harm to the final numbers.

    There is undoubtedly a large shock of new variability added into the nexus of precarious work through the recessionary period. I am sure some of the temporary is actually adding more temporarily and some could be adding it to the medium and longer run. So it will be difficult to measure, and leaving it to months that have reduced temporary hirings may only capture parts of the component that Andrew is trying to assess, so leaving out the higher months of temporary hirings may lead to more error and not less.

    Personally, given the great recession, I would tend to stick to the annual averages or potentially 12 month moving averages.

    I am sure something more complicated could be devised, but for the work involved I think annual averages are just as good.


  • The reason is simply the employers can avoid paying benefits.

Leave a Reply

Your email address will not be published. Required fields are marked *