Poverty and social exclusion are rising. The widening spread of incomes between the working rich and poor and the high number of families without any earner are major causes.

Without education no-one can be a full member of society. This too is social exclusion. The quality of education we require to be members of society grows relentlessly, and the exclusionary forces are also growing.

What can we do? Benefit strategies alone will never be enough. Could education play any part in mitigating these trends? Let us look at what we do and do not know about education and social exclusion.

We know that poor educational performance has been increasingly associated with low earnings and unemployment. In the mid-1970s unqualified school-leavers were one third more likely to be unemployed than their better qualified schoolmates. By 1990 they were two thirds more likely to be unemployed.

We know that technological change is producing a wider inequality of rewards for those in work. Rising demands for more highly skilled people have pushed up rewards at the top. But the relatively reduced demand for low-skilled people has reduced rewards at the bottom of the labour market.

We know that world trade has increased the supply of less educated labour “available” relative to the supply of more educated workers. This too has pushed down the price of poorly educated labour.

Here at home, we know that England has one of the widest inequalities in educational attainment in the advanced world. That is true not only of the inequalities in maths scores, where we do relatively poorly in international rankings, but also for science, where our average levels of achievement compare well.

There is an interaction effect at work here - low wage incentives breed low attainment and vice versa. But a wide spread of low attainments can only attract low paid, low technology firms.

And finally, we know that poverty and deprivation in families and neighbourhoods is associated with children’s performance at school. The striking analysis OFSTED produced for the Catholic Bishops, based on a sample of 2,500 schools, last year illustrates the point (see table).

But there are things we do not know. Economists dispute how far more investment in education improves earnings capacity, as opposed to giving signals to the employer that the educated person is an able and motivated employee.

Virtually all the evidence suggests that spending more money on schools, or on reducing class size beyond the Government’s present targets, would be a waste of money - at least if the objective is to raise the eventual incomes of the children.

It is true that training and education programmes for those out of work have produced some gains. Evaluations in the US have shown that basic training and job search help have a positive effect in getting people back into work and off benefit. But people coming off welfare in the US earn very little. So even the most successful training programmes fail to make a large difference in poverty status.

So given what we know, how can we make a difference? Skill levels are important. Education and training may not transform the UK economy as a whole, but they could do much to modify the distribution of earnings at the very bottom. Such a strategy cannot just be targeted on the few who are on benefit. Nor can it simply be done through the schools, because adult literacy and numeracy must be tackled too.

But if the general skill levels of the lowest qualified were raised, it would reduce the supply of low-skilled people competing for the supply of jobs. This would tend to raise the wages of the unskilled and improve the prospects of those who have gained the extra skills.

In the absence of such a strategy, boosting the supply of low skilled people on the labour market through Welfare to Work incentives will only drive down wages at the lower end even more.

Where you live makes a difference. The local skill mix affects where firms of particular kinds locate, and we are currently in a cycle in which they are looking for cheap unskilled labour gravitate to particular areas. Who you are at school with matters. A recent study has suggested that being educated with more able and motivated children translates into higher earnings in later life. Any policy of selection helps winners, but worsens the chances of losers. Able children gain from being educated with more able children, but the less able would gain far more.

We have a hard choice here. If we care about low income and low attainment, we must do what we can to ensure that the poorest pull up. If selective and cream skimming schools re-emerge, children will find it very hard to climb out of poverty.

And when we set performance indicators for schools, we should set a goal of reducing the unequal spread of performance around the mean, not just the overall results.

Achievement of basic mathematical skills at school is also an increasingly important predictor of later wages. In the US, high-school graduate earnings were measured at the age of 24. Basic maths proved important across the board.

English school children perform relatively badly at maths. This has something to do with the sheer time they spend on maths compared to children in other countries. For example, French students doing the maths, physics and science baccalaureate spend nine hours a week on maths in their final year. In England most maths students do four to five hours a week.

The quality of teaching matters and getting more good teachers into schools costs money. One of the best things we could do to help children in poor areas would be to make sure that each primary school had at least one first rate maths teacher who would make maths fun, and help other teachers to do so too.

There is some evidence that parents working full-time has an impact on educational attainment, but the issue and the evidence are contested. Results seem to vary by income and family circumstance. Children in poorer families may gain from the extra income if both parents work, or even if only the mother works. But overall, the amount of time parents spend with their children matters.

In sum, we should be extremely careful about offering easy solutions or suggesting that education can solve the problem of poverty. But carefully thought out interventions targeted at low performers could, over time, make a difference.

Professor Howard Glennerster is co-director of the Centre for the Analysis of Social Exclusion at the LSE. This article is summarised from a paper presented to a seminar of the Institute for Public Policy Research.

Link between poverty and failure

Free school meals take up amp; GCSE Results:1995

% of pupils... No of schools...% A-C %... 5A-G

0-10...........987............ 58........ 95

11-20......... 894............ 42........ 90

21-30......... 453............ 31........ 84

31-40......... 231............ 25........ 79

41-50......... 169............ 22........ 76

51-60......... 98............. 20........ 73

over 60....... 74............. 18........ 70