Blanchard's chapter 9 argues that a fall in the relative demand for unskilled labour is partly to blame for the moderate rise in equilibrium unemployment in the USA and the much larger increase in most of the European countries. It finds proposals of wage subsidies and graduated payroll taxes too costly to be practical, and concludes that governments should limit themselves to reducing credit constraints to help unskilled workers to invest in skills and education. In light of the large changes in relative demand and supply of skilled and unskilled labour, large changes in the skill premium can be expected and have to be tolerated, at least in the short run.

I would like to offer three glimmers of hope in this otherwise gloomy picture. The first one is that some countries appear to have escaped the dreaded trade-off between rising earnings inequality and an increase in relative unemployment among unskilled workers. An explanation of this puzzle could provide a way out for the USA, the UK and other countries which apparently face the dilemma. Second, the proposed wage subsidies, although very costly themselves, may allow a reduction in many other spending programmes; the gross estimates exaggerate the net social cost. Finally, some recent proposals are not as costly as the general wage subsidy; some could even be implemented at close to zero cost.

Chapter 9 argues that a shift in the relative demand of skilled versus unskilled labour affects aggregate unemployment because of the convexity of the wage curve, representing real wage rigidity.