Disaggregating Okun’s Law : a case-study for Malta

Abstract

The dissertation analysis the relationship between output and unemployment, what is commonly known as Okun’s Law. The study applies an aggregate and a disaggregate version of the relationship, to data for the Maltese economy to identify the differential reactions of unemployment to the various expenditure components of GDP. The results suggest that analysing Okun’s Law in this manner, alongside the more popular aggregate version, enhances the role of such relation as a short-run policy tool guide. The two versions are tested for the period between 2004Q1 and 2014Q4. In line with expectations, the aggregate version shows that there is a negative relationship between changes in unemployment and aggregate GDP growth. The Okun coefficient is found to be low by international standards highlighting the several rigidities that exist in the Maltese labour market. Using the disaggregate version, results show that exports and imports had the largest impact on unemployment, highlighting the importance of the external sector on the Maltese economy. The components of internal demand, which include private consumption, investment and government expenditure, have significant but lower impacts on unemployment. These results arise because exports and imports represent very high shares of the total output in Malta. In fact, if one were to account for the relative shares-to-GDP, the unemployment rate seems to be more sensitive to the components of internal demand. Specifically, a €1 increase in government expenditure would bring about a decrease in the unemployment rate which is larger than that brought about by a €1 increase in exports. This outlines the importance government has in safeguarding employment especially during economic downturns. On the other hand, exporting companies, whose production process generally involves the use of hi-tech machinery, have a relatively lower influence on unemployment than the components of internal demand. Finally, a comparison of the two versions suggests that, the disaggregated approach explains actual developments better than the aggregate version when analysing periods of high fluctuations. Furthermore, when tested as forecasting tools, the disaggregated approach gives better results.