Yves Smith at Naked Capitalism asks, "Why is Washington Dithering With Unemployment High?" Brad Delong speculates that those in power or those who chatter about those in power have become disconnected from "mainstream America." Edward Glaeser provides another possible answer for DC dithering and indifference here: the uncertainties of macro policy.
If nothing else, Glaeser provides the finest example of dithering I have ever personally witnessed:
The fundamental problem with acquiring certainty about Keynesian intervention is that anti-recessionary spending is just not very amenable to clean, compelling empirical evaluation. Recessions aren’t that common, and there are too many moving parts. Times change, so it isn’t obvious that the lessons of the 1930s – not that we can agree on those, either – are applicable today.
And so we are left wading in ignorance. It is a great tragedy that the most important area of economic decision-making is also the area where we will always know the least.
A tragedy indeed. Don't just do something! Stand there.
Glaeser provides a plot showing change in state unemployment rate and per capita Federal Recovery Act funds received by state. He rightly concludes that one can't conclude anything from it. The main reason for this is that there is no counterfactual. We don't know what each state's unemployment rate would have looked like sans funds. If it went up after receiving funds, does it mean the funds didn't work or does it mean that the funds prevented it rising as high as it would have without the funds?
He seems to be suggesting that the only way we can know whether or not stimulus funds work is to randomly assign states: 25 receive stimulus funds, 25 states don't. Then stand back and watch what happens to their respective economies.
Leaving aside the political impossibility of selling 25 states on the idea of being in the (randomized) control group that receives no funds, I'm not convinced randomization would provide a definitive answer. Fifty highly complex and disparate data points (disparate on politics, culture, history, economic output (even in good times))? Randomization, when it works, balances two or more groups on both observed and unobserved covariates. The smaller the groups, the less likely it will be that one can achieve such balance. Not a big deal for unbalanced observables. Any unbalanced observed variable can be controlled. It's the unobservables that will be a problem. I believe a responsible researcher would most likely still have to figure out why some unemployment rates went up and some went down in the "treated" states.
It may be more useful to resign ourselves to the fact that whatever we do, we'll be dealing with observational, not experimental, data; to imagine that if additional funds are made available we will have 50 self-selected volunteers to receive an infusion of federal funds. Since there is no control other than pre-/post-, we'll be left to figure out what worked and how.
BUT, we'll have to try something, like an additional stimulus, in order to figure out what works and how.
So I'm not sure what exactly Glaeser is trying to say. That we must be "certain" before adopting some "radical" Keynesian policy or strategy? That some critical value of statistical significance designed to distinguish the effect of a Keynesian fiscal alternative from a monetary null must be exceeded before we can proceed to use additional federal funds to stimulate employment and demand? That without the results from a randomized controlled trial, we should remain frozen like rabbits in tharn?
This seems like the antithesis of both prudence and sound science. The current economic downturn isn't a random draw of a black ball from an urn containing white balls and black balls. There's no sampling distribution. Very specific policies and actions landed us here. Now we must decide not only what policies need to be put in place to prevent it happening again, but also what policies would best drive us out of the ditch faster and sustainably. All we have to inform us about this is an historical series of unique situations, sometimes involving this country, sometimes involving other countries, that have some commonalities with the current economic situation, but are not exactly like it in every detail. (Although, Japan looks close if Paul Krugman is to be believed (and I almost always believe Krugman)). Waiting for "certainty" or even some degree of confidence derived from an experiment before one is willing to act seems like an irrational expectation and likely to prolong and deepen the "tragedy" of involuntary unemployment.
So I ask myself, is this really about science and certainty?
To add more complexity to the mix, even if we found that recovery funds did significantly reduce unemployment, that wouldn’t necessarily justify their cost. If you hire thousands of people on make-work jobs, then you are wasting their time. That cost needs to be weighed against the benefits of countering the recession.
"Make-work jobs?" How are teaching, public health, road-building, grid updating, new energy, and law enforcement "make-work" jobs? Some of any well-targeted fiscal stimulus would presumably go to preserving such jobs and services. In an ideal world, some "make-work" jobs would improve roads, bridges, mass transit, and other infrastructure. This is a reasonable role for government to play, especially if private entities are unwilling or unable to coordinate and produce these things. It's "investment" in public infrastructure and in the productivity of future generations. As a share of GDP, public infrastructure spending has declined since the 1950s and remained fairly constant since the 1980's. Meanwhile, we're still driving across unsafe bridges.
How is it a waste of their time to take someone who is otherwise unemployed and pay them to do such work, especially if it provides us with something of long-term value and them with earned income? AND it stimulates demand. Where is the down side, pray tell? Even if we guess wrong, we'll have much needed improved infrastructure, mass transportation, new energy sources, better educated kids and safer streets and highways.I believe the problem for Glaeser (and many others who dither while unemployment is high) is a morbid fear that putting people to work with taxpayer money (much as we have done for investment bankers) will crowd out efficient private sector productivity and jobs....some day. I'm left to wonder why crowding out of public infrastructure and education by unproductive, inefficient speculation in private financial and mortgage markets is OK, but I digress.
Yes, additional federal stimuli may "crowd out" some worthy private projects some day. And such spending may lead to high inflation some day. We can worry about both when they actually happen. Until then, worrying about them seems like another irrational expectation.
In the meantime, at least 10% of our population would benefit from having a job. We would all benefit from their having a job. In more ways than one. I'm willing to bet that most of them won't regard "publicly-stimulated" jobs as a "waste of their time." They'll be happy to have their involuntary leisure time crowded out.