Structure of Earnings Survey
Introductory note | Contents |
Introduction: The New Face Of The Structural Earnings Statistics
(not edited for language)
The Czech Statistical Office (CZSO) in cooperation with the Ministry of Labour and Social Affairs periodically presents results of the structural statistics on earnings of employees that are based on sample surveys.
The structural earnings surveys differ from other earnings surveys in several aspects. Mainly, the earnings of individual employees are gathered, not only the overall sums on the level of enterprise or organization. The items of the gross earnings are collected together with personal information of employee such as age, sex, education level etc., and information on time worked/paid. The resulting statistics are very detailed and are used for in-depth analyses of the labour market.
As the volume of information gathered is very exhaustive and the survey is labour consuming, it is neither possible to carry it on all enterprises and organizations in the CR, nor to do it by means of paper questionnaires as usual. The survey has been carried only on the sample of units that are chosen randomly from the Business Register, when the probability of the choice depends on the size of the unit: all units with 1000 employees or more are taken, but units keeping less than 10 employees are not sampled. An electronic way of collecting is used - tailor-made acquisition software reads data straight from the computer systems of the reporting units.
First such a sample survey took place for the reference year 1996 (it was conducted by the Czech Statistical Office). Information on differentiation and distribution of earnings in the CR had come to experts and to public after many years. The survey was repeated for the year 1997. But it was a great burden for the respondents and, moreover, there was duplication in some indicators with already existing Average Earnings Survey (ISPV), which had been taking place as a continual quarterly survey on hourly earnings on behalf of the Ministry of Labour and Social Affairs. To decrease the burden of respondent, CZSO gave up conducting survey on its own and, since 1998, it uses the results of ISPV that was enhanced to affect all the important variables that were monitored by CZSO in preceding years.
For this year, another significant change has happened in the Average Earnings Survey (ISPV) methodology, and consequently in the data publication. It was made to ensure better accuracy of the results and higher comparability of the gross monthly earnings of individual employee.
The list of variables has been enlarged, as well for the reason to meet the requirements of the European Union on the labour market statistics.
The methodology of the gross earnings calculation has been improved to ensure better comparison on wage levels between individual jobs. The gross earnings now cover all wages and salaries (incl. premia and bonuses) and payments for days not worked (holidays, leave etc.) and payments for being on call to work during the whole year. The average monthly earning of single employee in the reference year is calculated as the yearly earnings divided by the volume of time paid expressed as a number of months. Time paid is a time when the employee was paid by the employer; it means that all absences (incl. illness) are removed. Such average gross monthly earnings (in CZK) are the best indicator for comparison of earnings levels among different jobs.
But, such earnings are not (and cannot be) the same as average earnings gathered by CZSO’s regular enterprise reporting where overall earnings sum is divided by overall number of employees in the enterprise, because the number of employees covers also employees temporarily not paid (i.e. sick or out of work). Another differences in earnings levels between structural survey results and any other sources may stem (apart of influence of absences and different sample scheme) from non-coverage of employees with normal weekly hours less than 30 as regards ISPV.
On the other hand, we should take into consideration that results from any sample survey are always imprecise because of sample error; moreover, that some sampled units did not send information required (i.e. non-response) or that some records could not be used for one reason or another. Consequently, some small distortions might happen from this source.
To improve the results, the grossing procedures were employed on the levels of economic subjects for the first time in this year. Groups of industries and size classes were used as weights for aggregates on all-over-the-republic level; for aggregates on regional level only size classes were used. For detailed results, the grossing procedures were not made because of increasing value of error. This situation is, therefore, mirrored in the new structure of this publication – it covers three parts:
· Part A incorporates aggregated results for the Czech Republic as whole. They are grossed up by weights based on industries and size classes. This part includes only simple breakdowns with small number of categories
· Part B incorporates aggregated results for regions. They are grossed up by weights based only on size classes. As well as the part A, this part includes simple breakdowns with small number of categories
· Part C incorporates results with small level of aggregation. They are not grossed up at all; it means that the results speak only about the sample. These results are detailed and the cross-breakdowns are included.
This is not all what has changed in the publication. The emphasis is now put on the earnings distribution (in several breakdowns). Not only arithmetic mean is used, but also median, that shows the earnings of an employee in the middle of the earnings group – and in this way – speaks better about the real earnings level in the category. Great emphasis is put on the breakdown by sex. Some new breakdowns very introduced in order to better describe factors that effect the level of employee’s earnings, e.g. length of service in the enterprise.
The most important breakdown in the structural survey is occupation. The Czech version of international standard classification (ISCO-88 COM) has been used. The results in the publication are broken down by unit group (4-digit level).
The objective of this publication is to look at earnings differentiation from various points of view. First of all it is focused on comparing average/median earnings between individual categories of employees, especially broken down like nowhere else, e.g. by occupation, by sex, by age and by educational attainment. Also available in this publication are the individual components of earnings and the comparison of differences in basic categories from this angle.
Prague, December 2003
Ivo M a k a l o u š
Head of Employment and Wages Statistics Department
Note: Where only individual data available, asterisk can be found in the tables. The results that are in brackets are supposed to be less relevant (a dominance of one enterprise or organization).