Public Sector: The Public Sector includes:
For the purposes of this analysis commercial semi-state organisations have been categorised to the private sector.
Nace Rev 2: The economic sector classification (NACE) is aligned to the CSO’s Earnings Hours and Employment Costs Survey (EHECS). The economic sector classification used for the EHECS is based on the ‘Statistical Classification of Economic Activities in the European Community (NACE Rev.2)’ which can be accessed on the Eurostat website. The NACE code of each enterprise included in the survey was determined from the predominant activity of the enterprise, based on information provided to the CSO.
Gross Annual Earnings: Total annual earnings represent the total gross annual amount (before deduction of tax, PRSI and superannuation) payable by the enterprise to its employees. This information is obtained from Revenues PAYE Modernisation dataset. It includes bonuses and benefit in kind (BIK). It excludes pension payments and severance payments. In the small number of cases where an employee has been made redundant in the course of the year the employee’s annual income excludes statutory redundancy payments but includes non-statutory redundancy payments.
Weekly Earnings: Weekly earnings are calculated by dividing the gross annual earnings by the number of weeks worked as declared on the PAYE Modernisation (PMOD) dataset.
Usual hours worked: Number of hours per week usually worked.
Size class of the enterprise: Number of persons working at the enterprise (1-99 & 100+).
Further information and descriptions on the LFS variables are available in the background notes to the LFS release.
The tables in Detailed OLS results and Detailed Quantile Regression results chapters present the detailed results of the various models described earlier. The dependent variable for all models was the natural log of weekly earnings, and the explanatory variables were:
The models analysed are presented both including and excluding size of the enterprise as an explanatory variable.
The columns labelled “Estimate” in the following regression results tables contain the estimated parameters (i.e. β coefficients) from the regression equations. For the continuous explanatory variables (e.g. length of service with current employer), these estimated parameters can be interpreted as the percentage change in weekly earnings per unit change of the explanatory variable. For example, in Table 7.1, the estimated regression coefficient for “length of service with current employer” is 0.008. This value may be interpreted as follows: holding all other variables constant, average weekly earnings increase by 0.8% for every additional year’s service with the current employer.
The estimated models contain two explanatory variables which were analysed on the log-scale (log of over-time hours and log of hours). These coefficients can be interpreted as the percentage change in weekly earnings as a result of the percentage change in the relevant explanatory variable holding all other variables constant. For example, in Table 7.1, the coefficient for “Ln Hours” is 0.608. This value may be interpreted as follows: holding all other variables constant, for a 1% increase in hours worked per week, average weekly earnings increases by 0.608%.
For the dummy explanatory variables (e.g. sector of employment), interpretation of the estimated parameters is more complicated. For example, in Table 7.1, the coefficient for “public sector” is -0.055. Generally, in the literature, this figure would be interpreted as a -5.5% discount for public sector employees. However, the strict interpretation is that the estimated coefficient measures the premium in terms of log weekly earnings rather than weekly earnings. To estimate the premium in terms of average weekly earnings we need to get the anti-log of the estimated coefficient and subtract 1. For this example, we find the antilog of -0.055 ≈ 0.94648. Subtracting 1 from this we obtain -0.0535 or -5.4%; the public sector discount is -5.4%.
The estimated coefficients for the categorical variables in the regression models compare average weekly earnings for each of the categories in comparison to the reference category. For example, the reference category for nationality is “Irish”, therefore this is used as the base comparison group for each of the other nationality classes. For example, in the first column of Table 7.1, the coefficient for “EU excluding IE and UK” is -0.093. This value may be interpreted as follows: holding all other variables constant, an employee from “EU excluding IE and UK” would be expected to receive approximately exp(-0.093)-1 = -0.0888 or -8.9% less in weekly earnings than an “Irish” employee.
The reference categories used in the regression analyses for the categorical variables are as follows:
Learn about our data and confidentiality safeguards, and the steps we take to produce statistics that can be trusted by all.