Because education helps to unleash the productive potential of the poor, it is good for growth and good for poverty reduction. Linked to such direct benefits are wider equity effects. Greater equity in the distribution of educational opportunities will enable the poor both to capture a larger share of the benefits of economic growth, and to contribute to an overall increase in the rate of growth. By contrast, large-scale exclusion from educational opportunities results in slower economic growth; the benefits of growth are in turn enjoyed mainly by those with access to skills and knowledge.

The relationship between education and income equality derives from the economic returns associated with education. In cases where the demand for educated labour outstrips supply, its price - the wage - will increase. If demand for unskilled labour is growing at a slower rate than the demand for skilled labour, or contracting, then wage inequalities will increase, widening the gap between rich and poor. Where educational opportunities become more equitably distributed, they can act as a great leveller in relation to income distribution. This matters to poverty reduction for an obvious reason: the smaller the share of the poor in national income, the smaller the share of any increment in growth which will 'trickle down' to them. In Brazil, the poorest 10 per cent receive less than 1 cent for every $1 generated by growth.

The weakness of this 'trickle down' effect means that Brazil has to grow at seven times the rate of Indonesia for the poorest to receive the same level of income. The distribution of benefits from any growth process will reflect the distribution of assets and opportunities which different social groups already possess. Where this is highly skewed, any sudden spurt in growth linked to a rising demand for skilled labour is likely to exacerbate income inequality. This is what has happened in industrialised countries during the 1990's (see below), but it has also happened in many developing countries. One of the main driving forces which has increased inequality in the 1990's has been the emergence of growth processes in which demand for skilled labour increases faster than supply. The divergent experiences of South Korea and Brazil illustrate the importance of education to income-distribution patterns.

In the 1950's, Brazil had higher primary-school completion rates - 60 per cent compared to 36 per cent in South Korea. But by the mid-1980's, South Korea's primary completion rates were four times higher than those of Brazil. Because of the rising general level of education in South Korea, wages differentials actually narrowed when rapid economic growth began in the 1960's. The supply of skilled labour was expanding, so that those with education were unable to capture a larger share of the benefits of growth. Over the two decades up to the mid-1980's, the average income of those who had completed higher education declined from 97 per cent above the average wage, to 66 per cent above it. In Brazil, over the same period, the earnings of those with higher education had reached 156 per cent of the average wage.

Education was therefore one of the factors behind increasing inequality in Brazil. There is a two-way relationship between growth and equity. From a perspective of poverty reduction, countries with highly unequal patterns of income distribution must run faster to achieve the same rate of poverty reduction as countries with higher levels of equity. Less widely recognise of is the fact that the inequality linked to exclusion from education is bad for growth. Because lack of education translates into lower incomes, an economy's levels of saving and investment suffer, with attendant consequences for future income flows. Reduced potential for innovation and lower productivity also translate into lower incomes.

In turn, lower incomes mean reduced demand for labour, inputs, and consumer goods, all of which depress activity elsewhere in the economy. Once again, the experience of East Asia is instructive. In countries such as Indonesia, the multiplier effect of rural growth on manufacturing was large, with each percentage point growth in farming matched by output a 1.5 per cent increase in non-agricultural output.