A novel multivariate composite estimator for the Labour Force Survey

This paper introduces a novel multivariate composite estimator for the Labour Force Survey (LFS). Unlike the univariate composite estimators used in some countries, the multivariate estimator takes into account the different probabilities of transitioning between labour market categories, such as employment, unemployment, or non-participation in the labour force. By directly estimating all categories, it avoids residual determination issues, where one category is estimated as the difference between the total population and the sum of others. The multivariate approach improves the accuracy of the population estimates for each category. Additionally, the paper introduces a method to account for time-varying biases associated with how long the respondents have participated in the survey, explicitly incorporating wave-specific effects and their evolution over time.