25 April 2013

Bad policies -- or bad policy makers?

A few years ago, I debated the costs and benefits of ethanol with Ben Ho (his post with many comments plus my summary*). I used that debate recently in a paper (Economists owe ecology an apology) to illustrate how some economists favor policies that others (me!) wouldn't even consider -- e.g., a price subsidy that distorts decisions.

After I wrote a draft, I sent it to Ben to check for accuracy. Our subsequent emails reveal how differently we see cost-benefit analysis (CBA) and government interventions. This post (with his permission**) has part of our exchange. Tell me what you think.

Ben: I agree with your chapter, but I was so frustrated that you were so convinced that ethanol is bad and are so convinced that pursuing ethanol is like "striding off a cliff." My position was never that corn ethanol was good. My position was always that there is not enough data to reject the null hypothesis that corn ethanol is good.
Me: No, it's the subsidies. The cliff is NOT about corn. My null hypothesis is that subsidies should not be used if you can falsify the claim that they are good, i.e., find a single problem (beyond the obvious one of sourcing funding for it) in the subsidy. But you're relying on an overall CBA that (1) can never be resolved if you include long term data and (2) will always depend on assumptions. Note also that I'm not talking about corn ethanol, but subsidies. Ethanol -- or flower power -- or whatever is fine with me IF it competes...
Ben: I thought you might go with your default "do nothing" proposition. It sounds like a good principle, and in line with popular notions like the precautionary principle that is written into EU environmental law, and the classical liberalism of John Stuart Mills. But it's fundamentally irrational. Your position says we should do nothing relative to a status quo where there is no subsidy. But that's not the status quo. The status quo is that there currently is a subsidy. Getting rid of an existing policy has many costs as I teach in class right from Baron's textbook. It is costly in political capital and time. Congress and the president has a very finite amount of time and so changing policy has shadow costs. It is costly because existing companies have made sunk investments in existing policies, and therefore changes to policy are real costs. It is costly because recent theories of Kahnemann and Tversky argue that your reference point is set by expectations. And if your expectation is that ethanol subsides exist, then changes are coded as a loss. Therefore, the safer precautionary approach can arguably be the one that leaves existing policy be.

Furthermore, you discuss the costs and benefits of ethanol, but I could just as easily spin a story of the unmeasured benefits of ethanol, the importance of supporting US farmers, the pride Americans feel in home grown fuel, the advances in ethanol technology, the advances in ethanol infrastructure, the option value in case of future oil embargoes. And then we can weigh that against [your costs of] "groundwater depletion, surface water pollution, an enlargement of the dead zone at the mouth of the Mississippi, distortions in global food prices, deviations in land use, and (most recently) claims of engine damage."
Me: In short, your method pretty much allows government to subsidize whatever it wants. You could [spin that theory], but would you bet a month of salary on it? Even assuming you could quantify CBA?
Ben: I would bet that the benefits I listed outweigh the costs you list. Anyway, just very very very very back of the envelope. Spending on agricultural goods is 1-2% of GDP. As for oil, that's roughly 5% of GDP. So magnitude of effects lean in favor of oil.
My response [not in email] is that his benefits include "supporting farmers, pride, more technology and infrastructure, and energy independence," which are, respectively: a wealth transfer, worthless, the result of subsidies that may go to waste, and wishful thinking. His 2 vs 5 percent figures compares cash receipts but ignores substantial non-measured costs from groundwater depletion, surface water pollution, etc. (I'll allow that engine damage is not YET a big factor.)

Bottom Line: Ben thinks that a subsidy should be left intact because politicians are busy, companies have invested in subsidized activities, and people expect the status quo to continue -- unless it can be proven to be bad. That may sound good to an ethanol lobbyist or lazy politician, but it wouldn't last for a second at a profit-seeking company. Businessmen don't need proof of bad before they make decisions; they kill products that are not showing good results because they need to redeploy resources to products that DO produce benefits. Is government different, or do we just expect government to continue bad policies?***

* In that post, I bet that ethanol would be a de facto loser ("More evidence will come out that supports my POV OR congress will cancel the program as the economic/ecological disaster that it is. If the program expands AND a majority of evidence supports your POV, then you win.") It looks like he ignored that bet, which is handy given that I've won; see next post for more.

** Edited to save space, but I've tried to maintain the gist of the back-and-forth.

*** I just revised my paper on a groundwater tax that the Dutch government ended. SOME governments respond to costs and benefits.


  1. Interesting thoughts. Here is another take on the 'Null Hypothesis'. http://www.statisticsblog.com/2013/04/sudden-clarity-about-the-null-hypothesis/

  2. "non-measured costs from groundwater depletion, surface water pollution, etc"

    You mention that a lot, but in all our years of back and forth on this you have never cite a single study on what that number might be. My guess is that whatever it is, it will not be anywhere near 1% of GDP given that much of the externality is likely internalized by existing EPA regulations, and some estimates for the entire impact of the EPA is at most a couple percent of GDP.

    Also, is ethanol really that big a boogieman (a few $billion), compared to say the entire Farm Bill (~ $100 billion), which also puts out equally wasteful subsidies?

    To be honest, if I had a magic wand, yes I probably would eliminate ethanol subsidies and most farm subsidies.

    But two things: 1) there is a political system that realistically has to be accounted for when implementing policy.

    And 2) there really is a lot less evidence on most policy issues, certainly not enough to justify the confidence evinced by partisans on both sides of the issue.

  3. @Ben -- g/w pollution is not measured b/c it's either unknown or b/c remediation is too expensive. Why the 1% figure? Are you looking at values or costs?

    Yes, I agree that the Farm Bill should be killed with ethanol, so that's good, but I will not defend failures from the political system (farm special interest corruption).


Spam will be deleted. Comments on older posts must be approved.
If you're having problems posting, email your comment to me