The Great Wars, the Great Crash, and the Unit Root Hypothesis: Some New Evidence About An Old Stylized Fact

For decades, the prevailing sentiment among economists was that growth rates remain constant over the long run. Kaldor considered this to be one of the six important `stylized facts' that theory should address, and until the emergence of endogenous growth models, this was a fundamental feature of growth theory. This paper uses an endogenous trend break model to investigate the unit root hypothesis for 16 countries, using annual GDP data spanning up to 130 years. Rejection of the unit root, which is facilitated by the inclusion of a trend break, introduces the possibility of examining the long-run behaviour of growth rates. We find that most countries exhibited fairly steady growth for a period lasting several decades. The termination of this period was usually characterized by a significant and sudden drop in GDP levels. But rather than simply returning to their previous steady-state path, as predicted by the standard neoclassical growth model, most countries continued to grow at roughly double their pre-break rates for many decades, even after their original growth path had been surpassed.