Methodology 

The main objective of the labour cost statistics is to estimate the economic costs of the work factor in the production process of goods and services of an economy.

The two basic statistical sources to make estimates of the level and composition of the labour cost are the Quarterly Labour Cost Survey (ETCL) and the Annual Labour Cost Survey (EACL) conducted by the INE, and on the basis of which Idescat performs a statistical exploitation and extends the results for Catalonia.

On January 1, 2009 the National classification of economic activities 2009 (CNAE-2009), came into force, which replaced the National classification of economic activities 1993 Rev.1 (CNAE-93 Rev.1) and which until then has been the instrument used to codify the activity sector variable.

The change of Classification of economic activities led to a break in the statistical series. Therefore retrospective series were produced in order to link the data as accurately as possible.

2.1. Adaptation of the Directory of Contributory Accounts to the CNAE-2009

To produce the labour cost statistics, the base used is the ETCL, the central operation of the integrated data system for the labour cost. The population scope used by the ETCL is the File of Social Security Contributory Accounts. This is an administrative source that includes, among other variables, the economic activity of Contributory Accounts classified according to the Classification of economic activities applicable at any time. The level of disaggregation used in the sample design is that of the first two digits of the Classification.

2.2. Information collection with double classification

To produce the retrospective series of the ETCL and EACL, the information for the ETCL sample was collected throughout 2008 on the basis of the CNAE-93 Rev.1 and the CNAE-2009 to provide a link year between the two classifications.

In 2008, the File of Contributory Accounts was not yet available with the double codification (CNAE-93 Rev.1 and CNAE-2009). Nevertheless, the Central Companies Directory (DIRCE) produced by the INE already had a version using the double codification for 2007. The base unit of the DIRCE is the company classified by its main activity and identified by its tax identification number (NIF), the identifier that is also used in the File of Social Security Contributory Accounts.

The INE decided to use the information available in the DIRCE to reclassify the File of Social Security Contributory Accounts to two digits on the basis of the following available elements:

• DIRCE with double codification of CNAE and NIF.
• Matrix of correspondences between the CNAE-93 Rev.1 at 5 digits and the CNAE-2009 at 2 digits.
• Probabilistic matrix of correspondences obtained from the DIRCE re-codification work.
• Social Security File with the variables NIF, CNAE-93 Rev.1 and number of workers in the File of Contributory Accounts.

Using these elements the Social Security File and the DIRCE were fitted to obtain a directory codified to 2 digits of the CNAE-2009 which is compatible with the DIRCE.

2.3. Calculation method

Using the same procedure described for 2008, the directory units used in the ETCL corresponding to the years 2000 and 2007 were reclassified to the CNAE-2009. These directories were used to produce estimations in the CNAE-2009 for the periods from the first quarter of 2000 to the fourth quarter of 2007 based on estimators post-stratified by the increase in the sample from 2008. Thus the result obtained presents a gap between the data prior to 2008 constructed using the post-stratified estimator and the data for 2008 constructed using the usual estimator. To correct this maladjustment, link coefficients were calculated in consideration of the following premise: for each activity sector (industry, construction and services), the values of the new estimator in the CNAE-2009 must present similar inter-annual variations to those published, with the calculation of corrector coefficients (variation rate obtained between the years 2007 and 2008 in the CNAE-93 Rev.1) for each activity sector.

The link coefficients are only obtained for the two main variables: total net labour cost and salary cost, as these determine the evolution over time of the series. The final link coefficient used is the geometric average of the two previous cases and is applied to all of the economic variables of the units from 2000 to 2007. Thus, all of the labour cost series are linked, regardless of their disaggregation.

In analogue fashion, a link coefficient is calculated to ensure that the labour cost per hour worked per activity sector respects the variation rate published in 2008 and this coefficient links the series of effective hours worked.