It is well known that Scotland enjoyed a strong growth performance in the devolution years until the Great Recession in 2008. (This is not to assert that devolution was the cause!) As GDP grew strongly the employment rate also rose both absolutely and relative to England and other countries of the UK. In July -September 2006 the rate rose above the rate in England and stayed above that rate until January - March 2010. It fell below the English rate again as the recession began to bite more strongly on jobs in Scotland if not on Scottish GDP.
One less well known outcome is the extent to which there were variations in the degree of labour market inequality across Scottish regions during the period.
The latest Local Area Labour Markets in Scotland data published on 5 July throws some light on this issue.
The data are drawn from the Annual Population Survey 2011 which offers data back to 2004. These data can be combined with the less reliable Labour Force Survey data to take the series back to 2001.
Using these data and focussing specifically on the employment rate we get the following chart:
What these data show is that there is a clear negative relationship between growth and geographical inequality in Scotland. Between 2001 and 2007/08 the rising Scottish employment rate was associated with a narrowing in a measure of inequality across Scotland's 32 local authorities. The measure controls for the scale of the mean or average unemployment and so should not rise or fall as the mean Scottish rate rises simply for arithmetic reasons.
So, Scottish growth appears to bring reduced spatial inequality. This would seem to fit with a "trickle down," or convergence, view of the spread of growth as overall growth rises. There can be several reasons for this including rising prices/costs and reduced availability of resources in core areas as the economy grows. This supports the EU view that growth enhances regional convergence. It would appear to conflict with the cumulative causation, or even Krugman 'new economic geography', theory that growth is associated with divergence. It is also at odds with econometric evidence that regional inequalities are pro-cyclical, that is widen when the economy booms and diminish when the economy deflates over the cycle.
But the recession stopped the convergence.
The degree of inequality began to rise as the recession bit. But, interestingly, as the Scottish employment fell to, and then below, where it had been in 2001 the dispersion of employment rates across local authorities did not rise to, and above, where it had been in 2001. The coefficient of variation certainly rose from 5.5 per cent of the mean local authority employment rate to 7 per cent but not the 8 per cent that had existed in 2001.
Why was this?
I'm not sure. There is a clear asymmetry as discussed by Maria Demertzis & Andrew Hughes Hallett in this 1996 paper. It seems, in part, to fit with their centrifugal model. They found for the UK a negative/positive correlation between GDP/unemployment and regional variance, with both the negative correlations and the positive correlations decreasing with increasing lags in the cycle. This suggests that regional inequality falls as GDP rises above trend and unemployment falls with the effect diminishing over the cycle. However, we can't push this too far on such limited data with no econometric analysis.
What does appear to be the case is that as the severe recession hit local authority area inequality worsened but then slowed down as employment rates in core areas fell faster. As the Local Area Labour Markets in Scotland data release explains:
Since 2009 the gap has been reducing slowly but has not returned to the level seen in 2008. The most recent data shows that the employment rates are more stable and the gap between the top three and bottom three performing local authorities has reduced by 2.4 percentage points to 16.3 percentage points. However, this is generally due to deterioration in the position of the best performing areas, rather than an improvement in the worst performing areas.
This is clearly an interesting area for more detailed research.