How can we measure the difference between pay in the public and private sector

As outlined in “Why statistics are tricky: the case of average public sector pay”, comparing public sector and private sector pay is more complex than just comparing the averages for hourly pay in the CSO report on Earnings and Labour costs (edition for Q3 2011 – Q4 2011 here).

The differences in the averages will be affected by the types of jobs within the public sector. For example in 2007 24.5% of private sector employees were in Sales and Other occupations (mainly manual labour) compared to 8.1% of public sector employees (NES 2007). The average is also affected by pay – but this could be a few high earners at the top, or higher pay over all or both. The raw difference cannot tell us.

We will look at three reports, the results of which are summarised here.

Kelly, McGuinness and O’Connell, “Benchmarking, Social Partnership and Higher Remuneration: Wage Settling Institutions and the Public-Private Sector Wage Gap in Ireland”, 2009

This report looked at 2003 and 2006 employment figures, giving first raw differences between public and private sector wages and then examining a cohort of full-time, permanent employees aged between 25 and 59, analysed the differences in renumeration based on individual characteristics (eg. gender, education, time employed). Job content was not included in the evaluation.

The results from the 2003 NES were found to give similar results to a previous study (by Ernst & Young and Murphy, 2007), finding public sector employees earned between 6% and 10% more in weekly earnings than those in the private sector, with a premium of 2% to 6% for men and 10% to 15% for women. The premium declined at the upper end of the distribution, becoming a discount at the top.

By 2006, public sector employees earned 22% more than the private sector, with the premium increasing to 23% for men and 21% for women. In addition, the discount that had been present at the most senior levels had become a premium of 8% to 13% for both men and women.

These figures do not incorporate pension information since no such information was available for the 2006 figures. The authors state that in 2003 96.4% of public sector employees and 61.5% of private sector workers on their sample were deemed to have an employer sponsored pension.

National Employment Survey 2007 Supplementary Analysis

This report did not attempt to compare salaries for similar jobs between the public and private sector. It explicitly mentions aiming to maintain comparability with the Kelly et al report. It found that “when differences in individual and employment characteristics are controlled for, public sector employees are paid on average 19.1% more than private sector employees.”

Among these, the premium females got was greater than that for males (22.9% vs 14.8%) and for low-paid workers was greater than high-paid workers (for all workers, the premium for the bottom 10% was 25.7% and the premium for the top 10% was 11.4%).

Analysis done on the same age group as Kelly et al (2009) found a premium of 12.6%, 10.4% for males and 15.1% for females.

Foley and O’Callaghan, “Investigating the public-private wage gap in Ireland using data from the National Employment Survey 2007”, 2009.

This report analysed the NES 2007 figures further. Following Kelly et al, they worked on a cohort aged from permanent full-time employees aged 25 – 59. They found a public-private differential ranging from 11.5% using unweighted data, to 15.9% using weighted data for all employees (ranges were 12.6% to 19.2% for males, 18.0% to 21.4% for females)

They cited evidence that, in general, workers in large organisations are paid more. When they allowed for size of organisation, the public-private differential reduced to a premium of 10.0% to 14.5% (7.2% to 13.7% for males, 12.7% to 13.9% for females).

In addition some jobs were not readily matched with private sector jobs (primarily Gardaí, Prison Officers, and members of the Defence Forces). When these were excluded the average public-private wage gap reduced further, gave a premium of 11.9% to 13.9% (7.7% to 10.2% for males, 16.8% to 17.8% for females). If enterprise size was allowed for as well the premium was then calculated as 7.1% to 8.4% (2.7% to 5% for males, 11.5% to 10.1% for females).

The authors also found that, when analysed, the differential between public and private pay peaks at the bottom (3rd percentile) at 20%, then reduces until it becomes negative at the 95th percentile on.

The authors point out that there is no consensus on how to estimate the difference between public and private sector pay, and ‘[t]aking all of this into account, we advise caution in attempting to estimate one definitive “answer” for the average premium.’

International Comparisons

The ECB have produced the ECB Working Paper 1406 (December 2011) titled “The Public Sector Pay Gap in a Selection of Euro Area Countries“. (Dr. Constantin Gurdgiev blogs on this paper here and here).

This does not analyse public/private pay gaps but compares the raw gaps in various EU countries. It examines the the net and gross pay of public servants, the differences between the public sector work force and private sector work force (age, qualifications, etc) and the differences in pay between the public and private sector in Austria, Belgium, France, Germany, Greece, Ireland,
Italy, Portugal, Slovenia and Spain. Greece, Ireland and Spain have similar premia, with Italy and Portugal higher and the rest lower (see Appendix II, table III).

The paper referenced previous work on public/private gap, including those outlined above. The paper found that the pay premium in favour of the public sector is generally higher for women, for lower paid workers and in Education and Public Administration rather than Health. The difference decreases when monthly wages are considered rather than hourly wages, and if the comparison is made to large private companies rather than all companies.

This paper does not use NES data, so caution must be used and the figures can not be directly compared to the other reports. For example the ECB paper shows the private sector as having 27% managers vs 33% in the public sector. This contrasts with the figures in Fagan and O’Callaghan (2009) based on the National Employment Survey says that 14.7% are managers in the private sector and 4.3% in public sector. The paper concludes:

Although we used the term “premium” to denote the differential obtained after controlling for the observable characteristics of the workers (to distinguish it from the pay gap observable in the raw data), it must be stressed that this definition might not be entirely appropriate, as there may be other relevant characteristics that affect the differential, which, because of data limitation, cannot be controlled for even in the analysis with micro data (these include, among others, fringe benefits, which are typically higher in the private sector, or pension rights generally higher in the public sector, but also non-monetary factors, such as job security that is generally larger in the public sector). Furthermore, this study shares some of the same shortcomings as in the existing country specific studies, so some caution must be taken when drawing conclusions.


  1. From the reports above, we can see that the raw average difference between Irish public sector and private sector pay is very different from the adjusted differences when personal attributes are allowed for. In 2007, the difference in hourly rate was 47.6% and the difference in adjusted rate was 19.1% (according to the NES Supplementary). Assuming that the process of adjustment was similar to Kelly et al (2009), the rate also allows for fewer hours worked in the public sector on average.
  2. It might be argued that these studies are flawed since they consider the characteristics of individuals rather than job characteristics. Conversely people may argue that additional qualifications don’t mean one should have greater pay. Kelly and colleagues argue, “there is little theoretical or empirical grounding to the apparent assumption that wages will be primarily determined by job characteristics.” They add, “Within economics it is widely accepted that accumulated human capital is the principal factor determining an individual’s productivity and, hence, their earnings.”
  3. However it should be noted the adjustments do not eliminate the pay gap. Even when protection services are omitted and the comparison is made against large companies only, a gap remains of about 7% to 8%. Omitting protection services only changes this gap to 12% to 14%. So there is still, based on these reports, a substantial difference between public and private sector pay.
  4. This difference is not across the board; the models all suggest the gap is highest at the bottom and lowest at the top of the pay scales, and larger for women than men. Kelly et al (2009) suggest there are also disparities across sectors in the public sector.
  5. The ECB report shows that not all public sectors in Europe have the raw gap seen in Ireland. France, Germany, Belgium and Austria all have lower gaps, and while the age differences in those countries are less than in Ireland for public sector versus private sector (decreasing the raw gap), the education attainment is higher (which would increase it).
  6. If, as Fagan and O’Callaghan recommend, we are cautious with all these reports, despite the levels of analysis done, we should definitely be cautious using the raw averages in the Earnings and Labour Costs report.


CSO National Employment Survey

CSO National Employment Survey 2007 Supplementary Analysis

Elish Kelly, Seamus McGuinness and Philip O’Connell “Benchmarking, Social Partnership and Higher Remuneration: Wage Settling Institutions and the Public-Private Sector Wage Gap in Ireland” ERSI 2009

P. Foley, and F. O’Callaghan, “Investigating the public-private wage gap in Ireland using data from the National Employment Survey 2007”, Statistical and Social Inquiry Society 2009.

ECB Working Paper 1406 (December 2011) “The Public Sector Pay Gap in a Selection of Euro Area Countries

1 comment for “How can we measure the difference between pay in the public and private sector

Comments are closed.