A very nice paper by two of my favorite macroeconomists columbia.edu/~js3204/papers… But while I agree wholeheartedly with their emphasis on natural experiments, their hope that this can depoliticize macroeconomics seems a bit naive 1/ 
After all, meteorology is closely related to climatology, and both fields benefit from huge amounts of solid empirical research. Yet the politics of climate -- not just the response, but the scientific analysis -- are deeply polarized 2/
Why? For one thing, there's a lot of money at stake: skepticism about well-established science is good for the fossil fuel industry, so there's good money to be made in being a skeptic. 3/
So is macroeconomics really all that different? Yes, lack of experimental evidence makes it hard to pin down the effects of monetary and fiscal policy; but the experience of the past 10 years has provided enough evidence to clearly refute some views 4/
But to a first approximation nobody has admitted being wrong about anything, let alone changed views in the face of overwhelming evidence. And a large part of this is politics: the more or less Keynesian view supported by the evidence is seen as too pro-government to accept 5/
So yes, let's have better empirical work and more evidence. But I suspect that there will always be too much money at stake to truly depoliticize economics. 6/
