Are women workers in the public sector paid too much?

Posted on March 19, 2017 by Tom Healy

Tom Healy, Director NERI
Tom Healy, Director NERI

Yes, you read that question correctly. I will seek an answer to this question at the end of this blog.  But for now, judging by the standard reaction to data on earnings when published by the Central Statistics Office, many commentators are very set on the idea that public sector workers are paid far more than their counterparts in the private sector. This is the only story we get to hear about. 

The practical conclusion drawn from this by the usual suspects is that public sector pay should be restricted or even lowered (as is happening in any case for new and younger recruits receive lower pay as well as ‘lifetime earnings’ in the form of deferred wage income in the form of pensions).  A major part of the story is the sharp difference in earnings as between women workers in the public and private sectors.

Comparing pay in the private and public sector is challenging. Simple comparisons without adjustment for the stark differences in the nature of work and composition of workforce make any comparisons troublesome. Not a little controversy even bordering on disagreement among labour market economists have arisen over the years.  A new analytical report by the Central Statistics Office provides us a comparison which updates previous information [ Econometric analysis of the public/private sector pay differential 2011 to 2014 ].  It is now possible to examine trends in pay and comparisons with the private sector for four crucial years following adjustments to pay and hours of work in the public sector.

It would be tempting – as some analysts try – to arrive at a single number estimate of the gap (or ‘premium’). Still others might try to selectively pick this number or that number to prove their case.  At the outset it must be pointed out that comparisons of prison staff with their ‘equivalents’ in the private sector (what are their real-world equivalents?) or of primary school teachers with their ‘equivalents’ in the private sector (there are hardly any private sector primary school teachers to compare with) is like searching for the statistical holy grail.  Still, it is possible to undertake a sophisticated statistical analysis using various models and approaches as well as drawing on a large amount of information on characteristics of workers such as: 

  • Occupation
  • Sex
  • Educational attainment
  • Public or Private sector
  • Nationality
  • Membership of a trade union
  • Age (and age-squared)
  • Size of local unit
  • Length of service with current employer
  • Overtime hours worked
  • Hours worked
  • Shift work
  • Supervisory status.

This is as close as one can get to a ‘like-with-like’ comparison where differences are ‘controlled for’ statistically.  The Central Statistics Office is in the unique position of being able to undertake this analysis – under the strictest legal and confidential conditions – drawing on, and merging, information from tax files, survey data and creating a very large data set for the entire working population (excluding, of course, by definition the shadow economy).

What the latest analysis shows very much confirms previous analysis but also reveals some new and very significant facts. Up to the last publication by the CSO in 2012 (with a slight update in 2015) the main conclusion is that there was a significant but diminishing ‘premium’ to public sector workers. The extent of this ‘premium’ very much depended on the choice of model used as well as the sub-sector or male/female. Now, the story has changed. 

On average, when other factors are taken into account altogether (age, sex, tenure, education etc.) public sector workers are not paid as much as private sector workers when the public sector pension deduction levy is taken into account.  All analysis in this blog refers to full-time employees aged 25-59.  It should be noted that, whereas, 70% of public sector workers have a third level qualification the corresponding figure in the private sector is 50%.  There are significant differences within the sample as between men and women and as between lower paid and higher paid workers.  And the difference between male and female workers is crucial to understanding the different earnings profile between the public and private sectors. Let’s take the controlled comparison without including the impact of the ‘pension deduction levy’ introduced in 2009. Chart 1 highlights three important points:

  1. There is a positive gap in favour of public sector workers when the pension deduction levy is not taken into account.
  2. The positive gap in favour of public sector workers narrowed considerably between 2011 and 2014.
  3. Women workers in the public sector were significantly better paid than their ‘like-with-like’ comparators in the private sector.  This is not the case with male workers where a small negative premium (or a ‘penalty’) opened up in 2014.


Let’s run the comparison, again, and this time with the pension deduction levy factored in (Chart 2). Now we get the following results:

  • There is a very modest positive gap in favour of public sector workers when the pension deduction levy is taken into account but not in 2014 the last year of the time period covered.
  • The positive gap in favour of public sector workers narrowed considerably between 2011 and 2014.
  • Women workers in the public sector were better paid than their ‘like-with-like’ comparators in the private sector.  The opposite is the case for public sector workers. 


What are to make of these outcomes and what explains the sharp difference in outcomes as between men and women? It is difficult to draw firm conclusions without more in-depth analysis of make-female differences sector by sector. However, it is clear that there is a large concentration of low pay among female workers in particular privately owned enterprises sectors. Even when controlling for age, education, tenure, etc. it is not surprising that there is a positive premium in favour of women in the public sector. High levels of female employment in two of the largest sub-sectors of the public sector (health and education) means that for given levels of education women workers tend to be better paid than women in other sectors of the economy (bearing in mind that direct job-by-job or occupation-by-occupation comparisons are not possible in many cases).

Given that many commentators attempt to jump from statistical analysis of differences to statements of what ought to be the level of pay in the public sector we are left with a curious question:

If, controlling for other factors, women are better paid than women in the private sector then the logic of arguing for an alignment of public sector pay downwards to levels in the private sector is that public sector female workers are overpaid.

Patently, this is nonsense. However, the logic is inescapable. The only way of getting around this unpleasant conclusion is to make the case for private sector female workers to be paid more than they are to bring them up to the levels of their colleagues in the public sector. That is, if we were to accept the logic of using statistically controlled wage differences to arrive at the optimum wage level. 

All of the nonsense arguments used against public sector workers find ultimate refuge in the argument that public sector pensions trump all wage comparisons and the ‘gold-plated’ state guaranteed pensions means that even a penalty against public sector workers in the 2014 data is outweighed by pension considerations.

Space does not permit a thorough and careful investigation of the complex and varied area of public sector pensions. Suffice it to say that pensions are better, on average, for public sector workers.  This advantage has been underlined by the collapse in ‘Defined Benefit’ occupational pensions in the private sector. What is the implication of this? That we should level down public sector pensions to the level of the lowest eligibility for those in the private sector? Or, should public policy aim at ensuring a decent universal basic income for all pensioners topped up where feasible by occupational add-ons on the back of contributions over a working life time?

One final reflection is in order. The CSO data show that the distribution of earnings in the public sector is much narrower than in the private. Again, this is linked to the pattern of very low pay among women workers in certain sectors such as hospitality and retail.  There is a very marked pattern of a bigger gap in favour of public sector workers for the bottom half of the earnings distribution.  Even in the case of female workers, those in the top 30% of the distribution are less well paid than in the private sector (controlling for various characteristics and including the pension levy).  It is not surprising that some parts of the public sector are experiencing difficulties in recruiting and retaining senior staff. It seems to mirror a similar difficulty in the case of recruitment of young third level graduates in some disciplines where pay can be much superior in the private sector.

One of the reasons women have greater income and job protection in the public sector is that maternity leave and flexibility in hours of work makes the public sector an attractive option for parents and especially mothers. The penalty in terms of forgone earnings or lost promotional opportunities while taking extended parental leave is not as great as is the case in the rest of the economy.

Perhaps the obsession with public-private sector pay differences and the associated attempts to set one group of workers against another needs to be seen for what it is – an attempt to drag all workers down.  The sharp gender difference shown up in the CSO data exposes the falsity of what passes for neutral evidence-based analysis to justify further cuts to public sector pay and conditions.

Digital Revolutionaries