Empirical macro models in a time-varying framework

This dissertation consists of four empirical papers which employ as a common element time-varying vector autoregressions with stochastic volatility estimated with Bayesian methods. The first two chapters apply this methodology to the crude oil market, while the latter two chapters pertain to monetary policy. The first chapter investigates whether and how the dynamic effects of oil supply shocks on the US economy have changed over time. The salient feature that emerges from our empirical analysis is a remarkable structural change in the oil market itself. In particular, we find a considerably steeper oil demand curve since the mid-eighties which has important implications for comparing the macroeconomic consequences of oil supply disturbances across time periods. The second chapter is motivated by the observation of a dramatic rise in oil price volatility over time which has been accompanied by a substantial fall in oil production volatility. We investigate the sources of this opposite evolution of both oil market variables. Our main finding is that the observed ‘volatility puzzle’ can be rationalized by the fact that the price elasticities of both oil supply and oil demand have decreased substantially over time. In the third chapter, we explore the nexus between liquidity, inflation and asset prices in the Euro area in light of the monetary policy strategy pursued by the European Central Bank. We show that the impact varies considerably over time, depends on the source of increased liquidity (M1, M3-M1 or credit) and the underlying state of the economy (asset price boom-bust, business cycle, inflation cycle, credit cycle and monetary policy stance). The fourth chapter extends the time-varying methodology to factor-augmented VARs and re-considers changes in the US monetary transmission mechanism at both aggregate and disaggregate level. Our main results indicate that time variation is a dominant feature of key macroeconomic variables and their components. In analyzing the temporal evolution of disaggregate dynamics, often hidden by aggregate measures, we also uncover a considerable amount of heterogeneity in sectoral price responses which suggests that monetary policy actions exert an important influence on relative prices in the US economy