Referring to Europe, here is an email (slightly edited for clarity) I sent last night to a libertarian friend:
…let’s say the private sector is so screwed up that at the margin it won’t grow, no matter what, though some going concerns hold govt.-protected monopoly power.
If govt. cuts spending by $100, that is $100 less of output and employment (admittedly may be wasteful), but still the numbers come out that way. None of those workers are reemployed.
If govt. raises taxes by $100, some of the firms with monopoly power, while their profits are now lower, still will produce roughly the same output. Neither output nor employment falls so much.
That gives you a composition difference, pretty much fully blameable on government intervention, and without requiring any belief in Keynesian economics. Govt. spending cuts are *worse* for short-run gdp than are tax increases.
I am not saying that is always the true story, but I don’t see anything in the 1990s Alesina results to convince me to believe otherwise. Europe (or rather parts thereof) is less dynamic these days. And I don’t see why libertarians are in such a hurry to dismiss that particular story.
I am not claiming that Keynesian effects have to be zero, but rather using that as a hypothetical starting point. And this is a thought experiment to raise some points about observational equivalence, not a series of empirical claims about the real world.
You also may notice that in this “model” contractionary fiscal policy lowers gdp. And while expansionary policy raises measured gdp, it doesn’t necessarily do the economy a lot of good. It’s like a hidden transfer payment with possible hysteresis benefits through the illusion of make-work, but it is hardly on its own a source of much turnaround.