Phillip Lane of the Irish Economy blog notes a new IMF working paper on The Effects of Fiscal Stimulus in Structural Models. He says:
This new IMF working paper provides interesting analytical insights into the determinants of fiscal multipliers. Also striking is the set of co-authors: it represents a joint collaborative effort across the IMF, ECB, European Commission, Federal Reserve, OECD and Bank of Canada.
Here are a few graphs from the report (I went overboard and there are 14 graphs below, most are on the continuation page to save space and reduce load time).
Interestingly, from the point of view of stimulating GDP, there is little difference between government investment and government consumption, and both work better than changes in taxes and transfers (one exception appears to be "targeted transfers")
[Note: if you cannot read the graphs, the models used are the EC's Quest, the IMF's GIMF, the ECB's NAWM, the Fed's FRB-US, the Fed's Sigma, and the BoC's GEM -- see the paper for more details. There are separate graphs for each of the seven fiscal policy instruments, the experiments are both with and without monetary accommodation, and both one year and two year stimulus packages are considered. Apologies for the space between the headers and the graphs -- it's in the originals]: