This is the fifth and final post in a series on the details required to get carbon policy right. See also parts one, two, three, and four.

So far, I’ve done a lot of complaining — which, in and of itself, is just, well … whiny. Here, then, is a solution.

First, a very brief review:

  1. A test of good carbon policy is whether it encourages the private sector to invest capital in projects that will reduce GHG emissions.
  2. “Additionality” confuses carbon policy, by preferentially shifting investment toward less economic GHG-reduction technologies.
  3. Carbon taxes provide sticks without carrots, and thereby provide no direct incentive to those who might otherwise use carbon pricing to invest in projects that lower GHG emissions.
  4. Long-term carbon pricing is necessary to encourage private sector investment. Spots alone will not.
  5. Although not covered in this series, it bears repeating that auctions trump allocation.

Unfortunately, virtually all of the GHG-reduction strategies currently in existence (e.g., Kyoto, RGGI) or being contemplated (e.g., Lieberman-Warner, California AB 32) fail one or more of the prior tests. Moreover, all those actual/proposed bills are really complicated, with many moving parts that are rife for gaming — or, more charitably, significant legislative error. Here, then, is a better approach: output-based GHG regulation.

Output-based standards: The simple version

The concept of output-based standards comes from the world of criteria pollutant regulation. In conventional environmental rules, pollutants are regulated on a so-called “input basis” — which is to say, the more fuel you consume, the more pollution you’re allowed to produce.

As illustration, consider a typical parts-per-million (ppm) standard. Let’s say that you are a regulated source (boiler, power plant, etc.) with a permit that allows you to produce no more than 15 ppm NOx. You build your plant, put the appropriate pollution controls on, and check in at 14 ppm. Now watch what happens if you try to increase the efficiency of that plant. The denominator of your permit (the “millions”) is a function of only two variables: fuel combustion and air stoichiometry (the amount of air consumed per unit of fuel). For most technologies, air stoichiometry is just about fixed, so your only real variable is fuel combustion — which means that the more fuel you consume, the more total pollution you are allowed to produce, given a constant ppm.

Now, let’s assume that you suddenly find a way to dramatically increase the efficiency of your process, thereby cutting your fuel use in half. Since stoichiometry is fixed, you also cut your air flow in half. Let’s also assume that you cut your NOx emissions, but “only” by 40 percent. That’s good, right? Lower fuel costs, less fuel combustion, lower NOx — you would think so. But since the denominator (the millions) has fallen faster than the numerator (the “parts”) you suddenly find yourself in violation of your permit. 14 ppm NOx in the efficient device has suddenly become 14 x 0.5 / 0.4 = 17.5 ppm NOx, and you can no longer run your plant. The result? You simply keep running your inefficient device, over-emitting NOx, over-burning fuel, and over-paying for energy.

This is a really dumb policy outcome. It’s also ubiquitous. What would be vastly better is to go to an output-based standard, where your pollution is a function of the useful output of your project (e.g., lbs/MWh). Drive up your efficiency, and it’s easier to comply with your permit. Drive down your emissions, and it’s also easier to comply with your permit.

Remarkably, virtually all jurisdictions still use input-based standards, although a few (like Texas) have shifted to an output basis. Now let’s look at carbon.

Output-based GHG standards

As typically framed, carbon emissions are regulated on an absolute basis (tons per year, or some variant thereof). This would seem to be a step in the right direction, but it’s not. Our goal is to reduce global GHG emissions, not local emissions. And regulation that sets local caps has a really hard time setting an appropriate value for anything beyond battery-limits emissions. Witness the ongoing debate on biofuels on Grist — if I replace my fuel oil boiler with a biomass boiler, do I get credit for eliminating 100 percent of my carbon emissions, or should I first have to pay for the CO2 associated with fertilizer production, soil carbon depletion, and non-sustainable harvesting?

This is a legitimately difficult question. And it’s hard on the positive side as well. Suppose I replace my boiler with an on-site cogen plant that generates all my heat and power. I no longer buy power from the grid, and I’m making power more efficiently than the power I used to buy, but locally, my emissions increase, and so I have to try to figure out how much credit I should get (in terms of CO2 displacement) for the power I’m no longer buying. Again, hard issues.

Under current carbon rules, these dilemmas are addressed through some sort of offset rule (raising additionality issues, as Gar has pointed out) or on rather abstract theories that, since carbon prices will eventually affect those other fuels, you’ll eventually realize some incentive. Which is sort of like promising to buy someone a birthday present once they turn 80.

All of these problems go away if we shift to an output-based standard.

Nationally, our greenhouse gases come from three sources: electricity generation (about 40 percent of U.S. GHG emissions), transportation (27 percent), and thermal energy generation (33 percent). (The latter category includes not only residential and commercial space heating, but also a wealth of industrial operations — drying, calcining, melting, etc.) Let’s take electricity first.

In 2006, the U.S. produced 2,393 million metric tons of CO2 emissions in the process of generating electricity, and we generated just over 4,000,000 thousand MWh of electricity. By division, that works out to about 0.6 metric tons per MWh. That means that anyone who builds a power plant with CO2 emissions exceeding 0.6 metric tons per MWh is making our life harder, and anyone who’s beating it is making our life easier.

Under an output-based approach, we simply set a standard for all power plants at 0.6 MT/MWh and mandate that anyone who’s over that number has to pay for any excess pollution, while anyone who is under gets paid for their good deeds. Who do they pay? Each other. Measurement is easy, since both parties have fuel bills and electric meters — so, from a regulatory perspective, we simply require that they submit audited records of both at the end of the year, along with evidence that they bought or sold to get to 0.6 as appropriate.

To make sure that this structure caps CO2 emissions (e.g., to ensure that it doesn’t increase as total MWh production increases), we simply reset the cap every year based on actual CO2 emissions. If total CO2 emissions have increased by 2 percent, we reset the cap in the following year to 0.6 x (1 – 0.02) = 0.588. (A variant of this structure is that you can recalculate the MT/MWh factor every year based on actual data. As the payment streams incentivize low-carbon generation — and discourage high-carbon generation — this continual re-averaging would have the effect of steadily lowering the allowable level of pollution — a cute trick that requires no regulatory approval to tighten emissions standards.)

Now watch what happens. The structure has immediately created both carrots and sticks. If you install a solar panel (0 MT/MWh), you have 0.6 MT/MWh to sell. (And if you can drive up the efficiency of your solar panel, you get more tons to sell.) If you build a coal (1 MT/MWh) plant, you’ve got to buy 0.4 MT/MWh. Additionality? No need — just good guys and bad guys, selling or buying GHG emissions. Spots vs. strips? Solved — after all, the coal plant wants a long-term fixed price just as much as the solar plant does — they simply have to agree on a price. Goal- vs. path-based regulation? Done. The price is on carbon, not a technology. If a sexy new yet-to-be-dreamt-of technology comes along, it gets the same access to carbon pricing as everyone else.

The same mechanism can be applied to thermal energy generation, although the math is a bit trickier. (And indeed, to accurately capture the value of cogeneration, you have to factor in the value of thermal energy.)

The benefits are that the government need only set up the rules and then provide the appropriate oversight. It can then get out of the way and let market forces figure out how to optimally price and deploy technology. In doing so, it will naturally drive capital towards the most cost-effective carbon reduction approaches.

The tricky bits

That said, there are a couple of complexities:

  1. The model doesn’t deal with transportation very well. In theory, you could calculate tons/mile, or some equivalent metric. In practice, though, that would be really hard to monitor. Unless someone has a better idea, this probably means that you’d still have to impose some type of carbon tax on transportation fuel for economy-wide coverage. (Note, however, that the tax could be set to equal the carbon-content of a given fuel multiplied by last year’s average carbon price in thermal and electric markets on a $/ton basis, providing a linkage between the two markets.)
  2. For similar reasons, residential fuel use is hard. Again, it may be that a sector-specific tax is ideal.
  3. Biomass would require some additional paper trail to quantify which fraction of the carbon associated with the biomass is truly renewable and which part is not.

I’ve got no way to fix these complexities — but compared to the massive complexities that work into tax and/or cap-and-trade models, this is comparatively quite simple. It also has the benefit of being fiscally neutral. With the exception of the wrinkles for transportation and residential (which could be eliminated), the buyers and sellers balance exactly, so there is no net economic cost — simply a wealth transfer from polluters to cleaners. This means that we get big reductions in GHG emissions with no economic pain (and, better still, we see no net increase in energy costs, since every increase in the cost of a dirty MWh is offset exactly by a reduction in the cost of a clean MWh.)

Eager to hear comments from the Grist community. But if I were king, I’d roll this out tomorrow.