The logic in favor of fiscal expansion when faced with a recession, in which aggregate demand has tanked precipitously, unemployment is high, business investments have dried up, and debt-to-GDP ratio is rising (and when monetary policy has run into the zero-bound) is that it utilizes idle resources (labor and capital), sustains (and even boosts, depending on the size of the stimulus) demand, and prevents government revenues from declining (and thereby increasing the debt/GDP ratio). It also provides the required traction for the economy to climb the steep slope of recession and return to normalcy.
Temporary deficit-financed, fiscal expansion, has been opposed by the "austerians" on the grounds that it would amplify the already large debt stock and deficits, drive down market confidence, crowd-out private spending, and finally unleash an inflationary spiral that would only serve to worsen the precarious economic prospects of an economy fighting a recession. These posts - here, here, and here - are among those replying to this argument.
However, as with all economic theories, the true test of its utility lies in its application to real world events/issues. In this case, what has been the historical experience with countries facing similar macroeconomic conditions? The two big comparable events are the post-Depression (including wartime) US and Japan during the nineties. I have already blogged about them here.
This comparison of the total US debt (public plus private) from 1929 to 1948 in billions of dollars and the total debt as a percentage of GDP shows how a robust economic recovery can quickly wind down the debt/GDP ratio and lower the real debt burden. While from 1929 to 1933 when everyone was trying to pay down debt, the debt/GDP ratio skyrocketed thanks to contraction and deflation, whereas during and immediately after WWII, despite the massive borrowing the GDP grew faster than debt, and the debt burden ended up falling.
As a recent WSJ article opined, the persistence of decade-long fall in consumer prices in Japan had raised questions about the dynamics of deflation. The inflation-adjusted Phillips curve predicts not just deflation, but accelerating deflation in the face of a really prolonged economic slump.
However, though standard Phillips curve models claim that falling inflation results in rising unemployment, the experience of Japan with rising unemployment in the nineties led to a surprisingly mild, long drawn-out deflation instead of the expected deep, destructive and concentrated (depression-style) deflationary spiral. Japan's bitter and frustrating experience with economic stagnation and slowly falling prices since the early nineties raises fears that the US too could be stuck in the groove of sustained gradual deflation.
There is also mounting evidence that "prolonged periods of economic weakness are, with almost no exceptions, associated with falling inflation rates". Further, as the inflation rate goes toward zero, it seems to get "sticky", and as in the case of the US now, slight positive inflation persists in the face of an obviously depressed economy. Krugman has written that this slow movement towards deflation is also in keeping with theories of downward nominal wage rigidity (probably due to bounded rationality, there’s some downward inflexibility in prices and wages even after expectations have had time to fully adjust).
See also this episodic comparison of recessions in the US which shows that inflation falls with rising unemployment. However, as the aforementioned nominal wage and price stickiness theory would suggest, the decline in inflation (or disinflation) has been muted when the core-inflation rate has been low.
Update 1 (14/10/2010)
Jon Hilsenrath has this excellent account comparing Japan and the US.