The PSI Blog

The personal costs of public-sector contraction: questions for the era of austerity
Genevieve Knight and Michael White


  • When public-sector employees avoid unemployment by moving to private-sector jobs,  they still face losses in terms of earnings and career prospects
  • Public-sector contraction and forced movement of public-service employees to the private sector has happened before – under the John Major government of the 1990s - so we can use the experience then to gauge the effects of public-to-private job moves now
  • The average earnings loss for employees moving to the private sector in the 90s was 4 per cent – equivalent to an annual pay cut of about £1000 at current prices.
  • Employees who moved to the private sector without prior training faced earnings cuts that were twice as big – eight per cent.  Public sector training was valued in the private sector and for those getting training this helped to keep down the losses from across-sector job moves.
  • Many employees had to take jobs in less-skilled occupations, at the same time that they moved from public to private sector. In this case, the earnings loss averaged 11 per cent.
  • Movers from the public to private sector were much more likely to have a fall in skill level than those who moved within sector.
  • The researchers conclude that maintaining public-sector training provision and its quality will be crucial in limiting the adverse effects of austerity and sectoral rebalancing.

According to the UK Office for Budget Responsibility, by 2018 there will be 1.1m fewer public-sector jobs than there were in 2010 - even with a guarantee that education and health services will be protected.  This contraction of public services has formed an explicit part of the coalition government’s austerity plan to restore UK public finances to health. It has generally been argued that this contraction will be at least balanced, and probably exceeded, by growth of employment in the market sector, since this was previously seen to take place in the 1990s. The underlying assumption is that this public-sector contraction will not lead to a large increase in unemployment, but rather to a restructuring, with many former public-sector employees finding new jobs in an expanding market economy. 

This idea of rebalancing seems to assume that the people who lost their public-sector job but still have a job are, in a way, going to be ‘all right’ – because they are not unemployed. But this ignores the personal costs that very probably arise when job changes and cross-sectoral moves take place in an unfavourable job market. These costs may include lost income, derailed careers, or personal dissatisfaction. Before we can come to a judgement about the equity and efficiency of restructuring policy, we need information about the scale of these costs. Moreover, if one uses the experience of the 1990s as the basis for believing that public-private restructuring can/will take place without mass unemployment, then it is also reasonable to turn to the 90s as a source of information about these other personal costs of restructuring.

By the time that data on the 2010-18 period shows how people displaced from the public-sector fared, the restructuring will have come and gone and it will be too late to make use of the answers. This is another reason for looking to the 1990s experience for some insights. Following the severe recession of the 1990s, when unemployment rose more sharply than in any post-war recessionary period, there was a contraction of public-sector jobs only moderately less severe than the present one. According to figures from the Office for National Statistics, they fell by 300,000 in 1993 and by 800,000 over the decade. Of course, every recession has unique features and it is easy to point out differences between the 1990s period and the present time (we mention some of these in our conclusions).  But we should be able to address the questions that we pose for current restructuring policy by examining the information from the 90s.

Our research (1) confirms that the 1990s reductions in the public-sector workforce did not lead to correspondingly large numbers of former public-sector employees moving into unemployment.  In fact, unemployment following a public-sector job was over the decade less than half as likely as unemployment following a market-sector job. This was a notable achievement for public-sector employers over the turbulent 90s, and also for the public-service unions that negotiated many job security deals during this period. It is true that the public sector did resort to a relatively high rate of early retirement over the same period, and this was probably a less desirable policy from the viewpoint of the UK economy. However, if one adds together the exits to unemployment and the exits to retirement to give an overall figure for moves to non-employment, the overall rate was still five percentage points lower in the public than in the market sector.

The other main source of reduction in public-sector employment, and the main focus of our investigation, was relocation into the market sector. In fact, movements from public jobs to market jobs were at more than twice the rate of movements in the opposite direction (21 per cent to nine per cent). Among those in the public sector initially, nearly half (46 per cent) of all changes of employer also involved a change of sector. For those in the market sector initially, the corresponding figure was 15 per cent. One of the reasons why cross-sector mobility was possible for the public sector may have been a high level of fluidity in the market sector, with 60 per cent of market-sector employees changing employment at some time in the 90s. Because many market-sector slots were being vacated, there were  more chances to come across from the public sector.

Why should one expect forced mobility to involve some cost to the individual who moves?  Because, over time, employees acquire skills and know-how that are special to their particular employer and/or job, and when an employee moves elsewhere some part of these skills and know-how becomes irrelevant or of lower value. The larger the shift, the bigger the risk of some loss. If individuals can choose and time their own moves, they can circumvent these risks by finding employers who value their existing skills and/or offer them other good opportunities for advancement. But when moves take place under external time pressure, there will be less chance of finding the favourable opening, and losses will be harder to avoid.

Mobility from public-sector to market-sector employment did indeed involve some immediate fall in earnings. On average, the movers from the public sector in the 1990s lost four per cent of usual earnings in the ensuing year, by comparison with public-sector employees who were able to stay put. In the context of massive falls in real earnings during the 2010s, this loss may not seem very severe. Even so, when translated into money terms, the loss averages £470 per annum at 1990 prices, or about twice that amount when translated into 2010 prices. Lost fringe benefits, the value of which we are unable to calculate, would doubtless add to this.

An important factor restricting the losses in earnings for sector-shifters was the ‘portability’ of training received in the public sector. Overall, in-job training raised earnings for recipients by around a modest one-and-a-half per cent in the following year, but those in the public sector who got training and soon afterwards changed sector raised their earnings by seven and a half per cent by comparison with those moving sector without prior training. So public-sector training emerges as a valuable insurance against job-displacement costs. But by the same token, employees moving out of the public sector without prior training faced considerably above-average earnings losses. In fact, these non-trained movers experienced an 8 per cent earnings reduction relative to non-trained stayers – twice the average loss. Fortunately, in-job training was well established in the public sector during the 1990s, with half of employees (52%) receiving some training in any one year. But among public-sector employees moving to the market sector, prior training was at a slightly lower level, 42 per cent, so the beneficial effect of training was not as high as it might have been if it had been more uniformly applied.  

The differences in outcomes for trained versus untrained movers show how important it is to consider sub-groups of employees rather than focusing solely on the overall average. Still larger losses arose in the 90s when public-to-market job moves were coupled with a substantial occupational downgrading (for example, senior managers or professionals who became lower managers or assistant professionals, or craft workers who moved into routine jobs). Such career setbacks are not that uncommon and are associated with sectoral mobility. One in five of all public-to-market job moves involved occupational decline, whereas it affected only eight per cent of those staying within sector. When sectoral moving and downgrading were combined, the average earnings loss was 11 per cent by comparison with staying within sector and occupation. This earnings decline also signals potentially significant economic costs through lost skills and lower productivity. The sectoral movement/downgrading double-whammy hit as many public-sector workers as unemployment in the 90s and should be treated as an equally serious issue.

So far we have talked about immediate or short-term consequences of mobility, but the medium term is also important. We examined earnings for movers relative to stayers at yearly intervals up to six years from the original move. The general picture is that initial earnings losses speedily fade out, so that after two or three years the average position of the movers is indistinguishable from the stayers; even the effects of occupational downgrading are ironed out. This demonstrates the considerable resilience of both individuals and the labour market during the 90s. More research needs to be done on the recovery processes, an inherently difficult task because of the great variety of career paths that can be followed.

Though this medium-term earnings recovery is somewhat reassuring, the medium-term analysis also reveals some further disadvantage for public-sector movers by comparison with market-sector movers. The employees who moved from the market sector to the public sector initially suffered a slight hit on their earnings, but by year two their earnings were already six per cent higher than if they had stayed put and earnings remained four to five per cent higher in subsequent years. Those who moved from the public to the market sector failed to achieve these career gains: their initial earnings loss was wiped out after a couple of years, but this did not then progress to a significant gain compared with those staying put. Employees in the market sector seem to behave rationally in career terms: they put up with unavoidable short-term losses of mobility only when they see medium-term prospects of gain. Public-sector employees could not be as choosy as this, because of the external pressures on the sector: their mobility had no medium-term career gains to offset the short-term impacts.

What practical guidance does this research offer the present decade? Our findings identify public-sector strengths and weaknesses in regard to mobility and these can be used as pointers for policy.

A conspicuous strength of the public sector in the 90s was in managing the headcount contraction without pushing employees into unemployment in large numbers. We conjecture that the capacity for this achievement came from the sector’s steady commitment to union representation and consultation procedures, plus a heavy investment in human resource management (HRM) including personnel planning and internal transfer systems (2). Another related strength is the quality of in-job training that protected the earnings of displaced employees because it was highly valued by other employers, including those in the market sector. On the debit side, training was not sufficiently uniform to protect all movers, and occupational downgrading inflicted serious losses of earnings (and presumably of skills) when coupled with cross-sectoral moves, a specific problem for which the 90s’ experience did not find a solution.

In order to draw policy lessons from past experience, one needs to understand recent and current changes. The 2011 Workplace Employment Relations Study provides much relevant information on developments since the 1990s - a valuable summary is provided by van Wanrooy and colleagues in their 2013 book, Employment Relations in the Shadow of Recession (3). Among the important changes that they highlight are decline in public-sector bargaining over pay and conditions and the introduction of public-sector pay-review bodies and pay freezes ordained by central government. This might make the market sector more attractive to public-sector employees, but the market sector has also changed, with the weakening of the finance industries and the prevalence of low wage increases. How these changes will work out hasn’t yet become clear. There have also been recent changes in public-sector retirement and pensions policy, with some attendant industrial conflict, and this appears to be emerging as a new critical area. 

Despite these uncertainties, we believe our study of the 1990s experience does provide some useful lessons. The findings suggest that if the public sector is to achieve the same positive adaptation in this decade as it did in the 90s, it will need to safeguard consultation processes and HRM systems, and maintain or even extend training provision while preserving its highly valued quality. The research also contains some more general messages. The medium-term analysis provides reassurance in revealing resilience following on public-sector contraction and cross-sectoral mobility during the 1990s. At the same time, though, the findings confirm that sectoral rebalancing was not without personal costs for individuals, and for some affected groups the negative effects were large. 

Genevieve Knight is a Senior Research Fellow with the National Institute of Labour Studies at Flinders University, Adelaide, South Australia and a Visiting Research Fellow at PSI.

Michael White is Emeritus Fellow at PSI.


  1. The findings summarized here come from a research study at the Policy Studies Institute, University of Westminster, supported by the Economic and Social Research Foundation under Grant number ES/K00476X/1. Data used in the research are from the British Household Panel Survey, 1990-2008, made available by the UK Data Service, with most of the analysis focusing on 1990-99. Pre-publication drafts of papers giving details of the analysis methods and results can be obtained from the authors on request.
  2. A feature of industrial relations in the 1990s was the number of ‘no compulsory redundancy’ agreements that were struck, especially in the public sector. As these agreements were fully honoured in practice, employers were obliged to develop alternative means of shrinking headcounts. See White, M.R. and Bryson, A. (2013) Job cuts, job guarantees, and unions, The Manchester School, vol. 81(issue 6), 855-75.
  3. van Wanrooy, B., Bewley, H., Bryson, A., Forth, J., Freeth, S., Stokes, L. and Wood, S. Employment Relations in the Shadow of Recession, Basingstoke: Palgrave Macmillan.

» All PSI blog posts