Testing and Applying Cointegration Analysis in Macroeconomics

Abstract: This thesis focuses on empirical applications and tests of macroeconomic theories. It consists of three fairly distinct essays, bound together by the common theme of the use of cointegration analysis in the empirical macroeconomic analysis. The first chapter is an introductory chapter, while chapter two, three and four contain the three essays. The first essay, Wage Formation and the Relation between Real Wages and Unemployment in Sweden, examines the wage formation process and the labour market in Sweden between 1982 and 2002. The analysis is made using a cointegrated VAR approach and focuses on the determinants of the wage level and the role of trade unions in the wage formation process. The long-run relation between real wages and unemployment is found to be negative, implying that when unemployment increases there is less pressure on the wages. Furthermore, the empirical estimates indicate that there is a large deviation between the development in real wages and the development in productivity, in that the growth in real wages exceeds the growth in labour productivity. It is also found that the development in real wages is to a large extent influenced by the development in consumer prices rather than producer prices. This feature is interpreted as a high degree of trade union influence in the wage formation process. Finally, the analysis also points at the presence of rigidities in the labour market. In the second essay, Testing Structural Hypotheses on Cointegrating Vectors: A Monte Carlo Study, two tests for structural hypotheses on cointegration vectors are evaluated in a Monte Carlo study. The tests are the likelihood ratio test proposed by Johansen (1991) and Johansen and Juselius (1990, 1992) and the stationarity test proposed by Kwiatkowski et al. (1992). The analysis of the likelihood ratio test is extended by means of a Bartlett correction factor. Under circumstances common in empirical applications, all tests suffer from large size distortions and have low power to detect a false cointegration vector, but the likelihood ratio test fares slightly better than the test by Kwiatkowski et al. (1992). Applying a Bartlett correction factor to small samples substantially improves the likelihood ratio test. The aim of the third essay, A Panel Cointegration Analysis of the Relation between Private and Government Consumption, is to analyse the relation between private and government consumption in 23 OECD countries between 1970 and 2001. In particular, it addresses the issue of whether government consumption is a substitute for or a complement to private consumption. The empirical analysis is made using panel cointegration analysis. Three main conclusions are drawn from the empirical study. First, the relation between private and government consumption differs considerable among the OECD countries. Second, in most countries, government consumption is a complement to private consumption and, third, in countries with a larger government sector, it is more probable that private and government consumption are substitutes.