The spending multiplier is a key concept in economic policy in general and Keynesian economics in particular. Ryan Murphy and Jesse Gastelle are involved in project to find out which models are more accurate at the moment of measuring the spending multiplier. As stated in their website:

The spending multiplier relates changes in GDP with changes in government spending. If the government spends an additional $1, will the size of the economy increase by less than $1, more than $1, or exactly $1? Estimates range from well below one to greater than four. In preparing the American Recovery and Restoration Act of 2009, Christina Romer used a multiplier of approximately 1.5 when predicting the future path of the American economy.

However, the literature on the spending multiplier coming out of academia and the public policy world lacks examples of out-of-sample tests. This is an essential method for sorting out which results in the social sciences have merit and which do not. We have compiled research – and continue compiling it – in search of a model with a spending multiplier greater than one which is able to pass the standard of an out of sample test.

The project is in its beta version and Murphy and Gastelle will appreciate any feedback and involvement that can help the discipline increase its knowledge on the spending multiplier.

You can find the project’s website here.