As we noted in the last section, most capitalist economic theories argue that unemployment is caused by wages being too high. Any economics student will tell you that high wages will reduce the quantity of labour demanded, in other words unemployment is caused by wages being too high — a simple case of "supply and demand." From this theory we would expect that areas with high wages will also be areas with high levels of unemployment. Unfortunately for the theory, this does not seem to be the case.

Empirical evidence does not support the argument the neo-classical argument that unemployment is caused by real wages being too high. The phenomenon that real wages increase during the upward swing of the business cycle (as unemployment falls) and fall during recessions (when unemployment increases) renders the neo-classical interpretation that real wages govern employment difficult to maintain (real wages are "pro-cyclical," to use economic terminology). But this is not the only evidence against the neo-classical theory of unemployment. Will Hutton, the UK based neo-Keynesian economist, summaries research that suggests high wages do not cause unemployment (as claimed by neo-classical economists):

"the British economists David Blanchflower and Andrew Oswald [examined] . . . the data in twelve countries about the actual relation between wages and unemployment - and what they have discovered is another major challenge to the free market account of the labour market. . . [They found] precisely the opposite relationship [than that predicted in neo-classical theory]. The higher the wages, the lower the local unemployment - and the lower the wages, the higher the local unemployment. As they say, this is not a conclusion that can be squared with free market text-book theories of how a competitive labour market should work." [The State We're In, p. 102]

Blanchflower and Oswald state their conclusions from their research that employees "who work in areas of high unemployment earn less, other things constant, than those who are surrounded by low unemployment." [The Wage Curve, p. 360] This relationship, the exact opposite of that predicted by neo-classical economics, was found in many different countries and time periods, with the curve being similar for different countries. Thus, the evidence suggests that high unemployment is associated with low earnings, not high, and vice versa.

Looking at less extensive evidence we find that, taking the example of the USA, if minimum wages and unions cause unemployment, why did the South-eastern states (with a lower minimum wage and weaker unions) have a higher unemployment rate than North-western states during the 1960's and 1970's? Or why, when the (relative) minimum wage declined under Reagan and Bush, did chronic unemployment accompany it? [Allan Engler, The Apostles of Greed, p. 107]

Or the Low Pay Network report "Priced Into Poverty" which discovered that in the 18 months before they were abolished, the British Wages Councils (which set minimum wages for various industries) saw a rise of 18,200 in full-time equivalent jobs compared to a net loss of 39,300 full-time equivalent jobs in the 18 months afterwards. Given that nearly half the vacancies in former Wages Council sectors paid less than the rate which it is estimated Wages Councils would now pay, and nearly 15% paid less than the rate at abolition, there should (by the neo-classical argument) have been rises in employment in these sectors as pay falls. The opposite happened. This research shows clearly that the falls in pay associated with Wages Council abolition have not created more employment. Indeed, employment growth was more buoyant prior to abolition than subsequently. So whilst Wages Council abolition has not resulted in more employment, the erosion of pay rates caused by abolition has resulted in more families having to endure poverty pay.

(This does not mean that anarchists support the imposition of a legal minimum wage. Most anarchists do not because it takes the responsibility for wages from unions and other working class organisations, where it belongs, and places it in the hands of the state. We mention these examples in order to highlight that the neo-classical argument has flaws with it.)

While this evidence may come as a shock to neo-classical economics, it fits well with anarchist and other socialist analysis. For anarchists, unemployment is a means of disciplining labour and maintaining a suitable rate of profit (i.e. unemployment is a key means of ensuring that workers are exploited). As full employment is approached, labour's power increases, so reducing the rate of exploitation and so increasing labour's share of the value it produces (and so higher wages). Thus, from an anarchist point of view, the fact that wages are higher in areas of low unemployment is not a surprise, nor is the phenomenon of pro-cyclical real wages. After all, as we noted in section C.3, the ratio between wages and profits are, to a large degree, a product of bargaining power and so we would expect real wages to grow in the upswing of the business cycle, fall in the slump and be high in areas of low unemployment. And, far more importantly, this evidence suggests that the neo-classical claim that unemployment is caused by unions, "too high" wage rates, and so on, is false. Indeed, by stopping capitalists appropriating more of the income created by workers, high wages maintain aggregate demand and contribute to higher employment (although, of course, high employment cannot be maintained indefinitely under wage slavery due to the rise in workers' power this implies). Rather, unemployment is a key aspect of the capitalist system and cannot be got rid off within it and the neo-classical "blame the workers" approach fails to understand the nature and dynamic of the system.

So, perhaps, high real wages for workers increases aggregate demand and reduces unemployment from the level it would be if the wage rate was cut. Indeed, this seems to supported by research into the "wage curve" of numerous countries. This means that a "free market" capitalism, marked by a fully competitive labour market, no welfare programmes, unemployment benefits, higher inequality and extensive business power to break unions and strikes would see aggregate demand constantly rise and fall, in line with the business cycle, and unemployment would follow suit. Moreover, unemployment would be higher over most of the business cycle (and particularly at the bottom of the slump) than under a capitalism with social programmes, militant unions and legal rights to organise because the real wage would not be able to stay at levels that could support aggregate demand nor could the unemployed use their benefits to stimulate the production of consumer goods.

In other words, a fully competitive labour market would increase the instability of the market, as welfare programmes and union activity maintain aggregate income for working people, who spend most of their income, so stabilising aggregate demand — an analysis which was confirmed in during the 1980s ("the relationship between measured inequality and economic stability. . . was weak but if anything it suggests that the more egalitarian countries showed a more stable pattern of growth after 1979" [Dan Corry and Andrew Glyn, "The Macroeconomics of equality, stability and growth", in Paying for Inequality, Andrew Glyn and David Miliband (Eds.) pp. 212-213]).