Subjective well-being is a complex phenomenon coevolving with events in important domains of life. Panel vector autoregressions are a suitable tool to analyze the underlying structure of changes in happiness and its coevolution with changes in income, health, worries, marital status and employment status. With this technique we can simultaneously analyze the impact of the aforementioned factors on each other for the German Socio-Economic Panel (SOEP) data set. We find that positive changes in the named life domains are followed by decreases in subjective well-being (except for health, which is followed by well-being increases). Moreover, positive changes in well-being are followed by positive changes in most life domains. These findings are robust to different model specifications and comparable to similar findings for the British populace. We also examine how the structure of happiness differs with respect to different "Big Five" personality traits. Personality plays an important role, especially with regard to high traits of Neuroticism and Extraversion.