Just came across this IMF Staff Position Note by Olivier Blanchard (and coauthors) entitled
Fiscal Policy for the Crisis.
The best part of this paper is contained in Appendix II: Fiscal Multipliers-A Review of the Literature, and Appendix III: Five Case Studies of Fiscal Policy During Financial Crisis.
The conclusion I arrive at in reading these appendices is that during a financial crisis: [1] fixing the financial sector is of primary importance; and [2] fiscal policies may be marginally beneficial at best, and seriously detrimental at worst.
In the body of the paper, the authors draw several lessons from history.
[1] Successful resolution of the financial crisis is a precondition for achieving sustained growth.Here, they give the example of Japan during the 1990s. Evidently, "fiscal actions following the bursting asset bubble failed to achieve sustained recovery because financial problems were allowed to fester."
In short, fiscal stimulus failed in Japan. But we don't really want to believe that fiscal stimulus doesn't work. It didn't work here because things were not set up to allow it to work. I wonder how one might modify an IS-LM model to formalize this notion?
But there are more examples. "Delaying interventions, as was also done in the US during the Hoover administration and during the S&L crisis, typically leads to a worsening of macroeconomic conditions, resulting in higher fiscal costs later on."
So I guess that if we wait too long, the fiscal multiplier now is severely negative. Not sure what to make of this (like, where does this show up in the IS-LM model?). And in any case, they have their history wrong: Hoover did in fact react quickly to the crisis (although many of his measures were frequently blocked by the Democrats-Roosevelt, in particular-and then later adopted wholesale by the Democrats). And even though "interventions" were evidently delayed during the S&L crisis, I seem to recall a fairly protacted spell of growth in the 1980s. Unimportant details, I suppose.
Ah yes, but there was the "prompt and sizeable support to the financial sector by the Korean authorities that limited the duration of the macroeconomic consequences thus limiting the need for other fiscal action."
Huh? The Korean government fixed the financial sector, thereby eliminating the need for future fiscal stimulus? So fiscal stimulus is not needed, if the financial sector is repaired. This is supposed to be taken as evidence that fiscal stimulus is needed?
[2] The solution to the financial crisis always precedes the solution to the macroeconomic crisis.Sure.
[3] A fiscal stimulus is highly useful (almost necessary) when the financial crisis spills over to the corporate and household sectors.Not sure where this came from; they cite no examples (and indeed, the examples above seem to suggest otherwise).
[4] The fiscal response can have a larger effect on aggregate demand if its composition takes into account the specific features of the crisis.In this regard, they note that the tax/transfer policies implemented early in the Nordic crisis did little to stimulate output. Sure, the Ricardian equivalence theorem at work. As for "evidence" as to how a different composition of expenditure might work, they make the following compelling statement: "In theory, public spending on goods and services has larger multiplier effects..."
I see. That would be the same theory (IS-LM) that has absolutely nothing to say about financial crisis?
And so, on the basis of this, the authors' conclusions are:
The optimal fiscal package should be timely, large, lasting, diversified, contingent, collective, and sustainable: timely, because the need for action is immediate; large, because the current and expected decrease in private demand is exceptionally large; lasting because the downturn will last for some time; diversified because of the unusual degree of uncertainty associated with any single measure; contingent, because the need to reduce the perceived probability of another “Great Depression” requires a commitment to do more, if needed;collective, since each country that has fiscal space should contribute; and sustainable, so as not to lead to a debt
explosion and adverse reactions of financial markets.
And so there you have it. The fiscal multiplier is greater than one. Unless, that is, the stimulus in question is any one of: not-in-time, small, short, undiversified, non-contingent, non-collective, and unsustainable.