This paper applies panel cointegration tests and panel vector error correction models for 17 OECD countries and considers cross-sectional dependence and structural breaks to investigate the interrelationship between an insurance market's development and real output, controlling for banking activities. We first obtain evidence of a fairly strong long-run equilibrium relationship among them. Second, we find that insurance market development has positive effects on real output and that banking activities have an unfavorable, if not negative, effect on real output. In fact, insurance market activity is much more productive than banking sector activity. Finally, there exists bidirectional causality between insurance premiums and economic growth in the long run, suggesting the existence of the feedback hypothesis for the insurance–output nexus.