engleski

Predicting gross wages of non-employed persons in Croatia

We present the findings of a study aimed at building a model for predicting wages of non- employed persons in Croatia. The predictions will be used in the calculation of marginal effective tax rate at the extensive margin and in labour supply modelling. The database used is 2012 “EU statistics on income and living conditions”. The paper comprehensively explains the data source, variables, subgroups of employed and non- employed, and the results of the linear regression model, the Heckman selection model and the quantile regression model. The quality of predictions obtained by different models is compared and discussed.

gross wages ; estimation ; prediction ; unemployed ; inactive ; Heckman selection model ; quantile regressions ; Croatia

nije evidentirano