On returns to training in Portugal

• Autor(a) principal: Pereira, Pedro Telhado
• Data de Publicação: 2004
• Outros Autores: Budría, Santiago
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.3/1120
• topic: Returns to Training; Selection Bias; Logistic Regression

This paper investigates the earnings effects of training in the Portuguese labour market. We use the Portuguese Labour Force Survey to classify training according to multiple criteria, including providing institution, purpose, duration, and content of the training activity. First, we establish some stylised facts about the extent and determinants of different types of training. We find that there are major differences in training participation across groups, with elder, low educated workers participating substantially less. Second, we measure the wage effects of training. We find that in Portugal returns to training are large and significant. The estimated coefficients are about 12% in the case of men and 37% in the case of women. We show that discriminating between gender, education level, experience, the public and the private sector, and industrial activity reveals important differences across categories of workers. Workers with low qualifications and long professional experience earn larger returns. On average, women receive larger returns than men, though they are subject to greater variation across education and experience groups. The average effect of training is similar in the private sector and in the public sector. Experience in the private sector and education in the public sector are key determinants of the returns to training. Further, training to improve current skills and training in a firm attract largest returns. Third, the paper investigates whether and to what extent training participation affects the probability of entering and leaving unemployment. We find that being trained does not affect significantly the transition probabilities.