Many elements of Japan's economic stagnation the 1990s are reasonably well explained a modern, 'synthesis type', new-Keynesian model. The actual occurrence of some of the elements, such as deflation and zero interest rates, was, however, rare contemporary experience. As a result there has been a disconnect between what is theoretically understood by economists and the public discourse about Japan's 'lost decades'. The implications for policy are not very controversial theory - fiscal policy should work and monetary policy could work but not via interest rate transmission. But the length of time that Japan remained below its potential GDP fuelled a debate about the effectiveness of traditional policy measures. This gave rise to real questions about the first 'lost decade' (1991-2001): why didn't policy work? Was this some new mystery ailment for which known remedies were ineffective? These questions are still debated and they remaimportant because the popular reading of Japan's experience is ill-informed and is being mis-applied to current policy settings. This paper argues that Japan's economic experience of the last two decades does not provide unambiguous evidence that demand stimulus policies are ineffective. The paper describes the ways which economic theory has been changed by the lessons from Japan.