Tuesday, April 06, 2010

You Know You're Old When You Have a Job (The Advantage of Being Over 25 in the Labour Market)

In a previous post I noted that Ireland, with an unemployment rate of 13.4%, is faring almost 4 percentage points (in some cases more) worse than Portugal, France, Greece, Germany, Finland, Poland, Italy, Bulgaria, the Czech Republic, the United Kingdom, Denmark and Romania. At the time I wondered why there was a difference between the unemployment rate on the the graph (generated using Google Public Data) and the official figure from the CSO. For those who are interested, the question is still open. Let's begin by looking at the seasonally adjusted monthly unemployment figure for Ireland; as calculated by Google Public Data (shown below).

We can see that a figure of 13.8% is provided for January 2010; almost 14%. However, how does this stack up with the current official figure for March: 13.4%? (The recent revisions in the calculation have been flagged by Karl Whelan). Google Public Data directs to Eurostat for details about calculation: this is where the link leads. Unfortunately, the Eurostat monthly figures (seasonally adjusted) are different to the Google series; and the official CSO figures on seasonally adjusted standardised unemployment rates are a different series yet again. However, the differences are not massive between Eurostat and CSO; I suspect that Eurostat may be using some sort of harmonisation technique.

Here we are told by Eurostat that "quarterly LFS data are combined with monthly registered unemployment data by using a temporal disaggregation Denton model. For the most recent months (for which the LFS data are not yet available), the monthly benchmark factors are forecasted using seasonal ARIMA regression models. The provisional estimates are calculated by multiplying these factors by the available registered unemployment figures." Here Karl Whelan describes how the CSO's seasonally adjusted standardised unemployment rate extrapolates from the most recent QNHS data using Live Register figures on the number of people claiming benefits. Whelan also addresses the issue of whether the unemployment rate may be stabilising: "The most recent Live Register release reported an unemployment rate of 12.6% in February. A simple extrapolation from the QNHS release would suggest that this would be revised up to 13.3%. Overall, the picture has changed somewhat from one in which the unemployment rate appeared to be flattening to one where it still seems to be rising." He also notes that a decline in male participation may reflect discouraged former construction workers leaving the labour force. Also, by 2009:Q4, one third of the unemployed had been out of work for more than a year.

Nothwithstanding long-term unemployment becoming a more important factor, there is an advantage to being "over 25". It has already been suggested (in the States) that the recession is "causing much more job loss among the less educated than among college graduates... The brunt of the layoffs in this recession is falling on construction workers, hotel workers, retail workers and others without a four-year degree." This indicates that there is a general advantage (or greater chance of being "somewhat immune to the recession") in being more highly educated. There seems to be a similar effect for those who are aged 25 years and over. As I noted before, 1 in 3 men under the age of 25 are currently unemployed in Ireland. It is also helpful to look at the historical picture. First however, a data series must be chosen: based on the table below. Recent declines in the labour force, and broader measures of labour market distress, which would push the number higher than 13.4%, are ignored for now.

The Status Ireland visualisation of the seasonally adjusted standardised unemployment rate corresponds to the table above. However, I am drawn back to the Google visualisation, due to the flexibility of graphing options. By availing of the option to the restrict the age-group to less than 25 years, we get the graph shown below. We can see that the under-25 unemployment rate is currently *approximately* 32.4%, higher than it has ever been since data has been available. (Brendan Walsh shows that the national unemployment rate fluctuated in a band approximately between 5% and 10% from 1961-1981; subscription required for access). In the 1980's recession, the Google graph shows that the under-25 unemployment rate was previously in the region of 26.3%, at peak level. It seems that there is indeed some advantage to being "over 25" in the labour market; and now more than ever.

For completeness, it is shown here: in another Google graph, that the over-25 unemployment rate was previously in the region of 14.3%, at peak level during the 1980's recession. Now, it is in the region of 11.3% (three percentage points lower than its peak level during the 1980's). Finally, it is interesting to note that under-25 males are more affected by unemployment now than they were in the 1980's (compared to females).


Kevin Denny said...

It would be interesting to see what happens to participation at the other end of the age spectrum. An older worker who loses his job may just leave the labour market if hes reckons theres not much chance of getting a job again.

Martin Ryan said...

I agree Kevin. And it's technically a case of leaving the 'labour force', to use the survey-jargon. Unfortunately the (QNHS) micro-data only follows the same households for five quarters, which limits analysis to some extent.

The outstanding item on my agenda is to begin an investigation into the QNHS) micro-data. As you suggest, the cardinality of the age spectrum is important. Linearity is also important. One of my statements was:
"Nothwithstanding long-term unemployment becoming a more important factor, there is an advantage to being 'over 25'."

Is the age of 25 the most important in this debate? What about 24 years of age? Or 26 years of age? The micro-data should provide some guidance. Another issue is how much "long-term unemployment" is a problem across the age-profile; might this be similar for under-25's and some older age-group?

Finally, how does all of this relate to concerns about graduate unemployment? (Bachelor) graduates, typically of the age 20-22, (I assume) have a distinct advantage over their labour market competitors who do not have a four (or in many cases now, a three) year degree. Perhaps more importantly, (bachelor) graduates have a distinct disadvantage in relation to over-25's with more labour market experience. Again, I must make an assumption. Is the extra experience the basis for a premium? Or does it end up being viewed (by employers) as the correlate of higher (labour) costs that they really rather avoid?

Maybe the real advantage that over-25's posess is not that there is a market premium for their labour market experience; rather, maybe they got into steady jobs before more recent graduates?