Faktoranalys av EU: s strukturella indikatorer

This paper describes a factor analysis applied at the data of the European Union structural indicators. These are the official indicators for wealth and economic development in the European Union with main use to evaluate the outcomes of the year 2000 Lisbon goals for making the European Union the world’s strongest economy by 2010. The purpose of the data is to serve for comparing and “learning by doing”-policies, but have, due to size and format, so far been hard to use for a good compare between member countries. The aim of the paper is therefore to, by using factor analysis applied on the indicators data for the last 10 years, find general patterns of macro economic combinations, which show the causality between policy prioritizations in the European Union. The outcome shows that the main part of the goal achievement is primary related to sustainable wealth, explained by Gross Domestic Product (GDP) and Price level index. Two primary patterns then describe how the rest of the indicators relate to wealth. The first one shows that Research and Development expenditure (R&D), greenhouse gas emissions and low poverty rate strongly relates to the wealth, and is more likely to describe countries with a long sustainable wealth. The second pattern instead shows high GDP correlated to a high poverty rate and independent of R&D, which interpretets as describing more developing economies, not relying on social security or in the transition to a more knowledge-based economy. The factor analysis also shows two obvious casual patterns: R&D strongly relates to education and all three employment-related indicators shows a strong connection between each other. Labour productivity surprisingly shows no mayor influence on the rest of the indicators, except for less significant unexplainable connections. Finally, the factor analysis shows a constant negative correlation of business investment and GDP and no correlations between labour productivity and GDP. This shows that the current model may be developed further by manipulating data, for example by time lagging some variables with a long term effect