Are we on the verge of a public pay explosion? Is the fiscal rectitude of Chancellor Gordon Brown going to cost him the support of many public sector workers? What has happened to public pay in the past two decades? And who has fared best and worst among public sector workers?

The public pay bill is around Pounds 100 billion - almost 30 per cent of total public spending. To the Treasury, keeping control of public spending means tight control of public pay. But public pay is largely governed, directly or indirectly, by review bodies whose job is to check that trends are broadly in line with the private sector.

If the public sector is allowed to fall behind, then, it is argued, it will not be able to recruit the most able people and the declining morale of those who work in it will lead to an exit of the best workers and declining productivity of the remainder.

So in a period of declining public pay relative to pay in the private sector, there is a contradiction between the explicit objectives of review bodies and the requirements of tight central fiscal control This year’s budgetary and pay-setting process has seen this contradiction at its most intense for three reasons.

First, public-sector pay always fares relatively worse in a boom than do salaries in the private sector and relatively better in recessions. Before elections (and 1997 was no exception), the private sector is generally in an upswing and the gaps between the two are increasing. This disparity was heightened by the implausibly tight budgetary targets of Kenneth Clarke, which Gordon Brown is largely committed to maintaining. So public pay pressures are always at their most intense after an election.

Second, the general perception, especially in the public sector, is that workers fared badly throughout the Conservative administrations between 1979 and 1997, and that the election of a Labour government might be expected to reverse this process, however different New Labour claims to be from its predecessors.

And finally, it is often suggested that the public sector has been demoralised by a substantial loss of employment in the 1980s, by the apparent marginalisation of public-sector unions, by contracting-out and autonomous agencies, and by a general perception that public service was of low status throughout that period.

There is evidence for some of these trends - for example, public sector employment accounted for well over a third of employees only a decade ago while it now accounts for less than a quarter (around 5 million workers).

In contrast, when you look at the “raw” figures of average pay in the public and private sectors, there is little evidence of a clear trend against the public sector since 1979.

The ratio of public to private-sector pay has remained roughly constant and indeed, because public-sector jobs are on average more likely to be unionised and to require better qualifications than the private sector, average pay in the public sector is higher than the private sector, especially for women. The case for treating public-sector pay claims favourably seems weak.

But averages conceal a good deal. The jobs lost to the public sector in the 1980s and early 1990s were not representative: they were typically manual jobs, and especially those of relatively unskilled workers in local government and in ancillary services. They lost their jobs to outside contractors as manual tasks were opened up to private competition.

So in making the comparison, we should control for the occupational composition of jobs in the public and private sectors. And since the public sector is, on average, losing lower paid jobs, the compositional changes are going to overstate the relative pay of public-sector workers.

This point is illustrated in the graph, using data from the annual New Earnings Survey but with our controlling for the occupational composition of the public and private sectors.

Now there is clear evidence of a downward trend in public-sector pay relative to private-sector pay since 1979, although of course the actual path is also affected by booms and recessions along the way. The case that the public sector has fared worse in both pay and employment terms in the long run as well as in the mini-boom leading up to the 1997 election, seems indisputable.

But even this average is unrepresentative. The public sector is composed of different kinds of workers ranging from senior civil servants and administrators through to cleaners. How have these different groups fared?

In making this comparison, it is important to compare like with like: we can compare cleaners in the private and public sectors but we cannot compare civil servants in the public sector with those in the private sector.

Even when there are groups in both sectors (such as teachers, doctors and qualified nurses) the demand for their services from one sector usually swamps the demand from the other sector and pay in both sectors is determined by the dominant sector.

So to make this comparison, we have to look at individuals in different jobs and their earning characteristics - their age, gender, educational qualifications, training and the like. Research funded by the Institute for Fiscal Studies examined whether there is an average premium or penalty attached to the pay of workers in different parts of the public sector relative to their pay in the private sector. The comparisons make interesting reading.

First, women almost always fare better in the public sector than the private sector. They have better jobs than in the private sector, a higher chance of being unionised, and improved chances of promotion. See below.

Second, some groups, such as civil servants, seem to earn more on average than people with comparable qualifications in the private sector. But some groups of men (in health and education) generally do worse, although women again generally do better in these parts of the public sector.

The reasons for these disparities are varied. It may be that some public sector groups are better able to ensure that pay reviews are implemented than others, or that the case for higher pay for civil servants has been better made in the past than for those of us in, say, education. And the disparities between men and women suggest that the public sector has an important role in reducing pay inequalities in an era when, on average, they have risen sharply.

The graphic below, right, shows what has happened to relative wages in the education sector over the last 12 years. Here average weekly earnings of three groups of male public sector workers (universitypolytechnic academic staff, FE college lecturers and secondary school teachers) are shown as a proportion of average non manual weekly earnings. It illustrates quite clearly the extent to which different public sector workers have had different experiences over the late 1980s and early 1990s.

In 1984 there was a clear ranking among theese three groups with by far the highest wages going to the university sector and secondary school teachers actually earning less than the average non-manual wage. By the mid 1990s these differences have narrowed sharply. Both FE and university lecturers have had relative wage falls but secondary school teachers have had relative wage rises. So what does all this tell us? The argument that public-sector pay has been on a declining path relative to the private sector is strong. But there is little evidence that the Treasury will relax its tight pay targets in the short run, unless economic growth is significantly better than expected. So public pay tensions will remain. The public sector has an important role in reducing inequality, by providing women with better paid jobs than in the private sector. But there are differences in pay among groups in the public sector.

Perhaps the time has come for an overarching review, which asks what we expect of a whole range of public-sector workers, and whether the pay system is delivering the right incentives.

Richard Disney is a professor at Queen Mary and Westfield College and a research fellow of the Institute for Fiscal Studies. Amanda Gosling is a senior research officer at the Institute for Fiscal Studies. The work cited here forms part of an ESRC-funded project on public and private-sector pay in the 1980s and 1990s.