Some empirical and methodological problems with the estimation of the impact of unemployment insurance liberalization on the measured rate of unemployment