A lot of good thoughts here. He's not impressed by Edwards either. I do think the time-inconsistency problem is a little tougher than Jonathan suggests. He seems to present it (and I could be wrong) as "Keynesians always reach for fiscal stimulus", which he does a good job pointing out is not true in any particularly activist sense (we think it's nice that there are automatic stabilizers in place, but aside from that the reaction to 2008 was very different than the reaction to other situations, as Jonathan points out).
But time-inconsistency is more than that. The point of time inconsistency is a question of whether you can credibly commit now to decisions in the future. In other words, if you responde every time with fiscal stimulus that's fine as long as you reel it back in in boom years. Edwards argues that Keynesians don't demonstrate an ability to commit to that. That's why I highlighted all the discussion of the Bush tax cuts. That was a great (recent) example of Keynesians making a current commitment to long-term deficit reduction while others were the ones that made the bad decision (at least in terms of closing the long-term debt problem). Keynesians committed, many others didn't - and Keynesians committed and across the board specifically cited the long-term debt as the reason for their decision.
I do have to say, though, I found this point from Jonathan bizarre:
"Something I note that most of these quoted economists have in common is
that they do not support the Bush tax cuts."
Right! I specifically looked for quotes where they opposed the Bush tax cuts which also specified why they opposed it! Let's read on:
"While there may certainly
be a case that the Bush tax cuts were ineffective because of how they
targeted different income levels and how this may have distorted their
effect, this is not what these economists seem to be arguing. Rather,
they seem to be claiming that in the face high (even rising) long-term
costs of the entitlement programs and defense spending, we should
WHAT?!?!? No! The entitlement programs are the increased spending that is going to happen. The argument isn't that we should increase spending! The argument is that because there is a long-term debt problem casued by the entitlements we should increase revenue! Read on:
"This, to me, seems like a bad (possibly mildly
contradictory) argument. Long-term debt ought to be reduced by annual
budget surpluses or by increase in real tax revenue from an increase in
real income, not by an increase in the tax burden."
And now I feel like I'm taking crazy pills. Yes, we should run a surplus or at least a smaller deficit. You do that two ways: (1.) reduce spending, (2.) raise revenue. Lots of Keynesians like Obamacare because they think it will "bend the cost curve" and reduce Medicare spending. People who know Medicare like the back of their hand think this too, so I'm inclined to at least entertain the idea that it will work.
But the other key is raising revenue, and the Keynesians have shown that on precisely this issue they are not the ones that have the time inconsistency problem. They are willing to do things in the short term that are painful in order to deal with long-term problems. That is the very definition of not having a time-inconsistency problem.
Two points I meant to raise in the initial post, but didn't:
1. It's important to note that the real problems with the long-term debt are not the result of Keynesian policy at all: they're the result of Medicare and to a tremendously lesser extent, Social Security.
2. A few times... in the Edwards post, in Henderson's post, and now in Jonathan's post, there is this sense that we have to run surpluses in good times and deficits in bad times. That's actually not true. We could sustainably run deficits from now until the end of the world, so long as the economy keeps growing. They just need to be smaller in boom times and bigger in recessions. But there's no obligation to run a surplus. If things are going really well, of course it would be nice to run a surplus and pay off the debt even faster. But there's no reason why we have to.
How to Proceed at Generic Programming
4 hours ago