Dissecting the stimulus debate

There are really two quite separate debates going on over the stimulus, but they're being jumbled together into one gigantic ad hominem.  My take on both--pardon for being perhaps a tad obvious, but I think the debate has had a tendency to wander off into the weeds, so it's useful to be a little general from time to time:

Question #1  Will a fiscal stimulus work?

Define "work".  If the question is "Can borrowing money and spending it increase our measured GDP figure?" then yes, it is trivially true that stimulus "works".  So why don't we borrow a zillion dollars and spend all of it?  We could quadruple our standard of living overnight?

Because fiscal stimulus "working" is more than a question of increasing measured GDP.  In every other context, liberals are all too aware of the limitations of GDP as a proxy for human wellbeing.  In the context of the stimulus debate, however, all those reservations seem to fly right out of their heads.

When we ask "does it work" what we're really asking is "Will it increase our overall well-being?"  And to answer that, I think you need to know a few things:

a)  Are we increasing actual output, or only measured output?  If you pay people to mow their own lawns, you increase measured GDP, but you didn't gain much.  More broadly, if the economy is at full employment, increasing government spending definitionally crowds out private production.

Of course, we are not at full employment now.  I think that Gary Becker's worries about how much of the stimulus will truly generate new jobs is well founded--many of the projects don't seem well designed to pick up labor in the markets that have been hardest hit, and certainly not right away.  But it seems hard to escape the conclusion that we are probably going to reduce some unemployment by spending this money.

b)  Are the increases permanent, or temporary?  I'm shocked to see Paul Krugman complaining that people are assuming the jobs created by the stimulus will only last a single year in calculating the cost of creating them--back when the Bush tax cuts were proposed, he got very, very angry at . . . . people who assumed that the jobs created by the stimulus would last longer than a year.

There are two ways of looking at fiscal stimulus.  One is to assume that the government is simply closing the output gap between what we could produce at full employment, and what we happen to be producing right now.  The second is to assume that we are in a liquidity trap, and that we need a whacking great positive shock to jolt us out of a permanently lowered output level.  You might think of the former sort of stimulus as a pacemaker, and the second sort as the ER docs grabbing the crash cart and shouting "Clear!".

The evidence for the former sort of stimulus is decently strong, though it's costly.  The evidence that the latter sort of stimulus can actually produce a permanently higher level of output--what we need to believe if we are to accept the need for a huge spending package, and the possibility that our spending will create permanent jobs--is, as I wrote in the linked piece, practically nonexistant.

c)  What are the costs?  Right now, very little.  The government is borrowing near zero interest rates, and the marketplace doesn't want corporate debt at practically any price, so I find it hard to see much evidence for "crowding out".

But in the future, things start to get costly.

i.  The government will eventually have to roll over the debt, and the interest rates will not be 0% anymore.  Taxes will have to be raised, or other spending cut.

ii.  If taxes are raised, we'll see deadweight loss--people working/saving/investing less.  Republicans tend to overstate these costs, but they are far from zero.

iii.  Future government borrowing on this scale may well crowd out other private borrowing, meaning lower rates of investment.

iv.  To the extent that these programs are not temporary--and many of them aren't (I'm looking at you, high speed rail!)  they will incur ongoing operating costs.  These, too, will require future increases in taxes or cuts in spending.

v.  To the extent that the money is shoved out the door quickly, and in political ways, many of these projects will be badly designed, wasting resources and costing the government a lot of money to eithe rfix or end them.

The costs make it especially important to know whether we're getting a permanent boost out of this, or just a temporary tide-over.  If we get a permanent boost, we'll recoup a lot of the wasted resources.  If it's temporary, there are much more efficient and targeted ways to deal with the problem of unemployment, like temporarily topping up and extending benefits, or providing relocation assistance.

As I say, I'm skeptical that we'll get a permanent output boost.  But let's assume we think it will, and that the boost is more than big enough to cover the cost.

Question #2What sort of stimulus will best provide that boost?I'm agnostic on the question of tax cuts vs. spending, which makes me an oddity among most econopundits.  The complaint that spending is spent while tax cuts are often saved leaves me cold, because I think this focuses too much on that measured GDP figure, and not enough on welfare enhancements.  Right now, most households that save $500 by putting it in the bank or paying down debt will gain a big boost in welfare, because they'll worry that much less about credit card payments, or potential emergencies. 

On the other hand, given that the banks have really cut back on the credit they're willing to extend, it is worth worrying that that stimulus will stop dead with the consumer--it won't provide income to any other consumers who can then breathe a little easier.  But that raises two further concerns:  will we stay in a liquidity trap (I'm not sure we will), and if we do, will the people the government buys from spend their earnings, or save them?  If the latter, the multiplier isn't too high.

What I'm not agnostic about--and neither should any serious proponent of stimulus be--is the difference between stimulus now, and stimulus two years from now.  Spending may have a higher multiplier, but if you want an output shock, the immediacy of a tax cut far outweighs any possible benefit of a high speed rail project that's going to be built just as soon as we can design it, and get the EIS, and clear the public hearings . . .

So.  It's time to admit what we already know:  proponents of the stimulus are in favor of this package in large part because they favor a fairly large transfer of resources to the public sector, and the stimulus is a good way to achieve that.  There is, in fact, nothing wrong with this belief, for all that I disagree with it.  And most of the opponents of this package are opposed just as reflexively.

Myself, I'm agnostic.  I am skeptical that the stimulus will result in a permanent increase in output or decrease in unemployment.  However, I have also recognized that we are going to have a massive stimulus whether I like it or not, so I might as well view this as an interesting natural experiment, rather than get into a lather.  Especially since it's quite possible that, as an empirical matter, I'm wrong.

On the other hand, I'd like to see the people favoring it commit to some empirical benchmarks to test their case in advance.  If unemployment is still rising two years hence, for example, I think that will be a strong sign that stimulus does not work as well as billed.  It's no good complaining that "It wasn't big enough", either, because where are the stimulus enthusiasts demanding that we front load the spending into a single year so that it will be big enough, rather than wasting half of it on weakly stimulative spending that is likely to have its biggest effect after the recession is over? 

What will change my mind?  I'm still trying to figure that out, idly considering exotics like the second derivative of the unemployment rate.  For now all I have is a crude intuition:  if things are dramatically worse in two years, then the stimulus will have failed as badly as monetary policy.