By education we distinguish four major categories: eight grades or less, vocational school, high school and university studies. Occupations are classified in 200 categories, but given that at such a level of disaggregation the forecast will necessarily be inaccurate, we aggregate occupations in larger categories as well.

In order to implement the project we use the Wage Survey Database which is a large administrative database  containing information on the workers’  demographic characteristics, occupation, number of hours worked and wages. As the demand for labour depends essentially on corporate characteristics, we link the data with a corporate database established by the Tax and Financial Control Administration (APEH) which has information on all companies registered in Hungary using double entry book keeping.  

Using regression techniques, we estimate the demand for labour.The estimated coefficients link the number of employees of particular demographic characteristics and occupations with corporate, sector specific and economic characteristics.  Such characteristics are, for example, the wages, the capital intensity of a company, export share in sales, total output of an industry and its growth rate and certain macroeconomic indicators.  Assuming a given realization of various indicators in the future, we determine the demand for labour of companies.  

The forecasts are expected to be inaccurate in the public sector which employs nearly one third of Hungarian employees.  It is hard to make demand for labour predictions in these sectors because the number of public sector employees depends on political decisions and not on competitive forces, as in the case of corporations.

The final products of the project are a study summarizing our forecasts, an appendix containing the projection results and the STATA do file describing the models used for the projection.