Thoughts on the Stimulus

by Andrew Gillen

There is a debate raging in the eduworld over what impact the education portion of the stimulus will have for reform efforts. As I've followed along, I’ve realized that I am quite torn. As much as I would like to see serious reform of education, that is not the point of the stimulus. I'm also torn over the stimulus itself.

This picture, more than anything else, convinced me that fiscal stimulus is a good idea (It is slide 31 from a Richard Koo presentation which is available here) But I am not confident that the American Recovery and Reinvestment Act (ARRA) will be stimulating. This is in spite of largely buying the administration’s argument put forth by Romer and Bernstein that direct government spending is better (has a bigger multiplier) than tax cuts as far as stimulus is concerned. Note, RB is not an established fact per se (see this critique, for instance) but I think it’s a good baseline, and I also think that it’s directionally correct – by which I mean that under current conditions, the multiplier for government spending is likely to be larger than the multiplier for tax cuts. But the reason for the difference between the two is one source of my concern.

While the debate terminology can be confusing (velocity of money, marginal propensity to (fill in the blank) etc.), the argument that having the government spend money directly is more effective than letting taxpayers keep it basically rests on the historical tendency in times of recession for government spending to bring more unemployed resources (both people and equipment) back into active production than tax cuts. I think that would largely hold today as well, since in situations like this, your typical taxpayer is cutting back on consumption and increasing savings. Normally, the reduction in consumption would be at least partially offset by increased lending by banks (thanks to the increase in savings), but banks are largely paralyzed at the moment (according to Richard Posner, excess reserves have jumped from $2 to $725 billion).

So basically, I buy into the argument that for right now, government spending is better, at least in theory, because historically it brought more unemployed resources back into production. But, given this reasoning, it seems to me that any stimulus should be focused like a laser on unemployed resources, or at least resources that have a high probability of becoming unemployed in the absence of the stimulus. Browsing through recovery.gov, stimuluswatch.org (for leading spending candidates), and newspaper accounts of stimulus spending, I have trouble convincing myself that this is the case. Rather, as Megan McArdle wrote,

Mostly, Democrats took their wish lists, called them "stimulus", and look set to inflict them on the American people.

I’m sure some of the spending will employ formerly unemployed resources, but it seems like that is more of a coincidental afterthought, rather than the primary motivation and goal of the spending.

My second concern with the stimulus is that even if it is properly targeted at unemployed resources, it will delay and possibly prevent needed structural realignment. For instance, bailing out the auto makers is a no brainer as far as stimulating the economy is concerned, since it is full of either currently unemployed, or soon to be unemployed resources. But there is a very good chance that the auto sector is in need of serious change that the bailout would postpone, perhaps indefinitely. As Luigi Zingales said,

Keynesianism has conquered the hearts and minds of politicians and ordinary people alike because it provides a theoretical justification for irresponsible behaviour. Medical science has established that one or two glasses of wine per day are good for your long-term health, but no doctor would recommend a recovering alcoholic to follow this prescription. Unfortunately, Keynesian economists do exactly this. They tell politicians, who are addicted to spending our money, that government expenditures are good…

Once you give politicians a plausible excuse to go around spending even more money and rescuing firms, I have real doubts that you will ever be able to get them to stop. If you’re not convinced of this, go back and look at how Republican politicians have misused the Laffer curve, twisting it from the idea that lower taxes can lead to higher revenues, to the claim that all tax cuts will lead to higher revenues. I think many aspects of the education world are in a similar situation – hopelessly in need of reform, but with reform in good times being highly improbable.

Moreover, too much involvement in business affairs by Washington can reduce efficiency in strange and seemingly inexplicable ways. For example, check out this story, where it is explained that paper companies are adding unneeded diesel fuel in their production process to qualify for a tax break. (I for one am astonished to realize that using the tax code to run giant social engineering experiments can have unintended consequences. Who would have thought?)

Even more distressing is the heavy lobbying that has engulfed DC lately. With politicians not only open to, but actively encouraging folks to come up with ways to spend public money, the stimulus free for all is significantly enlarging the inefficient rent seeking/lobbying industry. A story in today’s Washington Post highlighting an instance of a 22,000% rate of return to lobbying effort has terrifying implications. HT: MR. Don’t let the small size ($3 billion) of the lobbying industry fool you, the example highlighted resulted in $100 billion in giveaways.

Between keeping inefficient firms alive, making efficient firms less so, and expanding the relative size and influence of the inefficient rent seeking sector, significant long term harm could be done. To the extent that the stimulus and bailouts lead to more inefficiency, our standard of living will be lower.

The worst case would be a stimulus that doesn’t focus on bringing unemployed resources back into production, while simultaneously leading to more inefficiency long term. I won’t pretend to be able to figure out if this is a valid description of our stimulus. But I am highly troubled by the fact that I can’t rule it out.