It's also worth separating the descriptive and prescriptive aspects of Keynesian economic theory. On the descriptive aspects, it seems to be doing fairly well so far. In particular, Keynesian models predicted that we wouldn't suffer significant inflation from the stimulus program + QE/QE2, due to a liquidity trap, whereas some monetarist models predicted that high inflation would result (and the Austrians were even predicting hyperinflation). Since high inflation hasn't materialized, that seems like at least some evidence in favor of the Keynesian macroeconomic models.