We implement a quantitative empirical test of the fiscal theory, FTPL, defined as a non-Ricardian model of fiscal policy (the ‘FTPL model‘) in which accumulated debt is ultimately devalued by a rising price level. Thus inflation is caused by expected fiscal policy. We used indirect inference to compare its match to postwar US data with that of a standard New Keynesian model with Ricardian fiscal policy. This FTPL model is first treated as a permanent regime, expected throughout the postwar period, with equilibrating real interest rates set by monetary policy; we also repeated this within a classical RBC model with market-clearing real interest rates. In both cases we found that there was no stable solution, due to the ‘doom loop‘ linking debt, interest rates and inflation. We then modelled FTPL as a temporary regime including a monetary policy with a weak response to inflation, expected to revert to the Ricardian New Keynesian (Orthodox) regime eventually. While this hypothesis was rejected over the whole postwar sample, it was accepted for the post-GFC period alone, with the Orthodox regime prevailing otherwise. However the model fitting the whole period best was the New Keynesian with an orthodox monetary regime and a fiscal regime the credibility of whose Ricardian commitment was limited by the debt-related probability of a future switch to the FTPL. This gives a key role to the FTPL in explaining postwar US behaviour.