In the wake of youth unemployment passing the million threshold there have been a number of claims and counter-claims about what was happening to young people's employment prior to the recession. This is an attempt to address some of these arguments. While there is evidence of a worsening situation for young people between 2004 and 2007, followed by a modest improvement before the onset of recession, neither of these developments is truly comparable with the rise in youth unemployment since the recession, which is clearly driven by lack of jobs (constrained demand for labour in the jargon). The idea that the surge in youth unemployment is due to 'structural' factors pre-dating the recession rather than collapsing demand is simply wrong, as is the notion that it can be reversed through supply-side measures such as increased sanctioning of benefit claimants or work experience programmes. The question of how to help the important minority of young people who face difficulty in establishing themselves in work even under benign labour market conditions is not the same as the question of how to reverse the continuing increase in youth unemployment.

The chart (click to enlarge) shows two measures of youth unemployment from the second quarter of 1984 to the most recent figures for the third quarter of 2011. Note that all data points are quarter 2 apart from the last point. Source is ONS, Labour Market Statistics, tables A05 and A06. http://www.ons.gov.uk/ons/publications/re-reference-tables.html?edition=...

The blue line shows the standard ILO unemployment rate, which is the unemployed as a percentage of those who are economically active (in work or looking for work). The unemployment rate as of the third quarter of 2011 is the highest recorded in this data series.

However the unemployment rate for young people is influenced by educational participation as well as by the numbers of unemployed. If educational participation at age 16-24 increases, as it has done over this period, this can increase the unemployment rate independently of changes in the numbers unemployed, as students tend to be economically inactive (so the denominator for the unemployment rate is lower). For this reason, it is useful to supplement the headline unemployment rate data with the unemployed/population ratio: unemployed 16-24 year olds as a percentage of the total population aged 16-24, the red line on the chart.

Both of these measures have moved together over time, but the gap between the two tends to widen during periods of high unemployment as young people tend to remain longer in education or move into some other form of inactivity, further boosting the headline unemployment rate. Over the entire period there has also been a widening of the gap during periods of low unemployment, comparing 1989-91 with the mid-2000's. This reflects the long term trend towards increased educational participation post-16.

The red line allows us to assess some of the assertions that have been made over recent weeks about longer term trends in youth unemployment. A number of commentators have rightly pointed out that unemployment among 16-24 years olds increased in the mid-2000's, and have suggested that the current very high rate of youth unemployment should be seen as the acceleration of an underlying trend http://blogs.telegraph.co.uk/news/danielknowles/100118217/our-greatest-s.... This argument sounds plausible when based on the headline ILO rates, and even more so when (illegitimately) based on numbers of unemployed young people (ignoring demographic change) but is considerably less plausible when we look at the unemployed/population ratio. The rise from 2004 to 2007 is certainly marked enough to be a cause for concern, but it is far less marked than the rise in the headline ILO rate, indicating that educational participation -more specifically, economically inactive students- played a role in the increase in the ILO rate. The upward trend in unemployment was then partially reversed between mid-2007 and mid-2008, on both measures.

Thus there is no long term trend of rising youth unemployment, and the sharp rise with the onset of recession looks like an unambiguous reflection of constrained labour demand. In other words,much of current youth unemployment, and most of the change in youth unemployment, is unemployment pure and simple. Young people are bearing the brunt of demand-deficient unemployment because they are struggling to establish themselves in the labour market. There is no evidence to support claims that 'structural' changes are behind this growth in youth unemployment. The rise in the unemployed/population ratio was 4.5 percentage points from the pre-recession trough to the third quarter of 2011. Even if we accepted that the pre-recession rise was entirely due to 'structural' changes, this increase was only 1.1 percentage points. That is not something to be relaxed about, but it is a phenomenon of a different scale to what we are seeing now.

If this analysis is broadly right, the 'solution' to youth unemployment, or rather to the rises in unemployment we have seen since 2008, is to increase labour demand. Measures to keep young people engaged in the labour market make sense (as long as they are sensible measures) but they should not be seen as a substitute for more jobs and on their own their impacts on unemployment will be slight, while 'get-tough' measures will have zero effectiveness. Similarly, while there is a continuing need to help the substantial minority of young people who experience great difficulty in the transition from full-time education to the labour market, the stress falls equally on the words 'substantial' and 'minority'. We should not confuse issues, and policies for young people in general should not be based on the experiences of those who would be struggling even under greatly improved labour market conditions. It is an open question as to whether social security policy and after-the-event interventions are the levers we should be concentrating on here: educational policy may well be more important in tackling this continuing disadvantage.