Tuesday, September 20, 2011

Saving Green by Going Green ... Or Is It?

Today, I received an email from the Environmental Defense Fund urging me to protect clean air by calling my congressman and urging him to vote against the TRAIN Act, a law that will create an independent committee do do cost-benefit analyses of new EPA regulations before implementation. Opponents of the bill argue that it is unnecessary because the EPA already does cost-benefit analyses of its regulations and the new law would just duplicate that effort and delay implementation of the regulations. On the other hand, proponents say that the EPA analyses may be biased (after all, they're not exactly a disinterested party) and that an independent analysis is necessary to make it unbiased. (Actually, the latest version of the act does a lot more than just call for cost-benefit analyses; it also explicitly blocks certain regulations, see here.)

What I was interested in is just what, in particular, proponents believed the flaws of the EPA studies were. The American Association for Clean Coal Electricity, (ACCCE), a power-company lobbying group, has a web page that discusses the issue from their point of view. They identify two perceived flaws: first, that the EPA considers only one proposed rule at a time and does not lump multiple proposed rules together in its analysis; and second, that the EPA does not consider other negative economic effects such as lost jobs. (Note that on the association's front page, they claim that the regulations the TRAIN Act will block will cost 1.4 million jobs. However, on the actual page that discusses the TRAIN Act, they say it will cost 1.4 million job-years, totalled over an 8-year period. These are very different.)

Anyway, the first criticism does not, at first, seem to make any sense. If regulation A has costs which exceed benefits, and regulation B has costs which exceed benefits, then added together, regulations A and B will collectively have costs which exceed benefits. The only way this will not be true is if either:

(a) The benefits of implementing both regulations A and B are less than the benefits of implementing A alone plus the benefits of implementing B alone.

(b) The costs of implementing both regulations A and B are greater than the costs of implementing A alone plus the costs of implementing B alone.

This, of course, raises the question of in what circumstances these can be true. For case (a), I can think of a simple example: suppose that both regulations will reduce exposure to the same pollutant, and the pollutant has a hormetic dose-response relationship. But for some reason I don't think that's the case that the ACCCE is thinking about. For case (b), I can think of a different case, that seems to be the case that the ACCCE is discussing. Suppose that both regulations reduce the production of electricity, and electricity (like most goods) has diminishing marginal value. Then just looking at each regulation individually, and estimating the cost by multiplying the current price by the amount of reduction (let's say), will understate the total costs. In the diagram below, the true cost is C+D but the "looking at each regulation individually) approach will give you something closer to C.
We can now estimate about how big this difference is. For the sake of argument, I will use the assumptions that are most favorable to the ACCCE's position. They mention that there will be a total reduction in coal power production of 30 to 100 gigawatts (GW) due to "these and other rules". 100 GW is equivalent to 876,000,000 MWh over the course of a year, or about 25 percent of the total U.S. electricity consumption 3,741,485,000 MWh per year. Of course this is not a good estimate of total electricity consumption lost because some of the capacity lost in coal gets replaced by other energy sources. If I am interpreting the chart labeled "2016 CATR+MACT impacts" of their own report correctly (it's on page 6 of the PDF, or page 5 going by the page numbers on the page), it looks like about 60 percent of capacity lost in coal gets made up in increased natural gas. So you end up with a total of about 10 percent reduced consumption. According to the review here, the short-run price elasticity of demand for electricity is about 0.2. So 10 percent reduced consumption corresponds to about a 50 percent increase in price. That means that the triangular area D is about 25 percent of the area C.

However, my understanding (at least based on what it says here) is that for most of these regulations the benefits exceed the costs by at least several times. So just a 25 percent error won't make a significant difference.

----

The comment about jobs, however, is more interesting conceptually, and I think they have it backwards. Here's how I am thinking about it. Let's say that electricity and labor are perfect complements, so a business can produce a "widget" by using one worker and one unit of electricity. Suppose that currently the business is producing X widgets, and so it is using X units of electricity, and the new regulation will increase the price by Y. Suppose you ignore the issue of jobs. Presumably that means you assume that the business will just produce the same number of widgets as before. Then the total cost is X times Y. But suppose you take jobs into account, and you take into account the fact that now the business will produce fewer widgets because the cost of producing them went up. But if they made this change, then that means the change was beneficial (compared to just absorbing the extra cost). In other words, the "reduction in jobs" is partially a benefit because it means that you are now using less of the more expensive electricity.

-------------------------------

Of course, conservatives aren't the only ones who often use faulty economic reasoning when talking about environmental issues. During the 2008 presidential campaign, Barack Obama claimed that oil companies had 68 million acres of land they were "not using" and that we needed to make them "use it or lose it." Most importantly, this claim was false: most of the 67 million acres of "non-producing land" was actively being explored and prepared, it's just that no oil was coming out of it yet. But even if it was true that oil companies were deliberately ignoring large portions of their land, why is that necessarily a problem? There are only two reasons I can think of as to why they would do that. One reason is because they think that oil will become more expensive in the future and they would rather wait and sell the oil when it's more expensive rather than extract and sell the oil now. But if that's the case, then the oil companies' actions would raise the price now (when it's cheaper) and lower the price when they get around to extracting it (when it's more expensive), thus reducing the volatility of oil prices over time. Isn't that a good thing; to save it for when it's scarcer? Another possible reason is if they are colluding to reduce supply in order to raise the price now. But that theory doesn't seem to hold water because oil is traded on a world market, and the vast majority of world oil and gas reserves are controlled by companies outside the United States, so it doesn't seem like U.S. oil companies could reduce the world supply that much just by drilling a bit less. And in any case, if the problem is that we are using too much oil, isn't it good if the oil price goes up because that means that people will have an incentive to switch to renewable sources?

1 comment:

Dan Mont said...

Hi Alex,

IT seems to me that if you have two interventions that the costs could exceed the benefits for each individually, but taken together the benefits could exceed the costs. And you could get this outcome without even having to assume an interactive effect between the two benefits.

Say that exposure to pollutant A gives you a 60% chance of cancer, and so does exposure to pollutant B. And assume it costs X dollars to clean up A, and another X to clean up B (that is, there is no joint production in the clean-up)

If the population is exposed to A and B then 84% of the population will get cancer. Getting rid of A for X dollars reduces that to 60%. (same for B) But getting rid of A and B together reduces the cancer rate to 0% for a cost of 2X. It could be that getting rid of A or B is not worth it, but getting rid of both IS worth it.