Paul Krugman blogs about fiscal multipliers, citing some new work on the topic by Ilzetzki, Mendoza, and Vegh. They have a new dataset with quarterly observations on GDP for 20 advanced and 25 developing countries, and I think they use a VAR with an identifying restriction suggested by Blanchard and Perotti.
Their bottom line seems to be that the size of the fiscal multiplier really depends on characteristics of the particular economy. Small, open economies with flexible exchange rates have a very low multiplier, as do heavily indebted countries. Contrary to what Krugman suggests, Ilzetzki and co. find that the post-1980 U.S. (in other words, after the Volcker disinflation and the commitment to price stability and flexible exchange rates) has a pretty low fiscal multiplier, around 0.3 or 0.4.
