I've been thinking about a few ways to get output based standards to get larger economy wide coverage and give equal cost/benefits for polluters/reducers during the next 42 years.  

1. Assuming total delivered energy remains constant,  the standards have to be tightened by 2.38 percent each year to get a 70 percent reduction in CO2 by 2050. This 2.38 will be adjusted based on actual energy use. I am certain that it will be cheaper to reduce either CO2 per useful thermal energy unit or electricity (not sure which one, but I'm sure it won't be equal). While not explicitly stated in the article, I'd suggest suggest taking this into account when re-averaging the output based standard on a annual basis.

2. This plan does not account for the way heat is used in industrial production of steel, aluminum, glass, cement and others. Heat is an intermediate product and tonnes of these commodities are the final product. Your version of output based standards incentivizes industry to generate heat efficiently but then is agnostic to how well the heat is used to make the final product. For example, depending on the process used to make aluminum, different amounts of delivered heat are needed per tonne produced. In certain energy intensive industries (perhaps any commodity industry that produces more than .5% of US CO2 output), give them their own standard for CO2 per tonne of commodity. These markets could be linked to the other carbon markets. Of course, if they can recycle waste heat, they should be given credits for those units of delivered energy.

3. On a similar vein, output based standards do not account for "negawatts". When I heat/cool my home, my output is not delivered Btus, but a house in which I am comfortable. So it does not account for the efficiency of my house converting delivered units energy into my desired output. Under your plan, if the price of delivered energy goes up, there is a "stick" to build more efficient homes, but no "carrot". I imagine including residences into OBS would be a nightmare. Perhaps the building code for insulation could be adjusted based on the CO2/unit energy spot prices and average regional temperatures. I'm not sure how to pay for this. Perhaps an government auction element in the carbon markets to raise money for these "negawatts". Perhaps a tax on heating oil. The same type of ideas could be used for retrofits, but it is more expensive to renovate to a certain code than to build to that code in the first place. I'm not sure how to bring competition into the insulation market.

4. The same can be said for appliances and fuel efficiency. I'd suggest regulation such as CAFE standards for cars and appliances. One market possibility is that manufacturers who produce items that are less efficient than the standard would have to buy credits from manufacturers above the standard. With assumptions about average use of these devices, it would be possible to link the car and appliance market to the broader carbon markets. However, I'm not sure if this level of regulation to get larger economy wide coverage is worth it.

5. I agree that hydro and nuclear producers should  get an allocation of credits for the electricity produced in the long term. All energy options must be treated equally to get the desired result. However, in the first few years, output based standards represent a huge unexpected windfall to ~31% of US electricity producers. I don't see how this windfall profit reduces GHG's. Perhaps we can phase in the amount of credits already installed facilities get over 10 years.

I don't really understand who owns transmission capacity. I'm guessing its the regulated utilities. Is there a way to incentivize improvements in the grid to remove line losses? I wonder how this could be linked to the carbon markets. I'm guessing that at some point, transmission investments will be competitive with other investments. Would the entire electricity model have to be changed or could this be achieved with the utility owned grid model?

1. You're absolutely right, and this has been in the details of our most wonkified conversations with the feds.  Note though that the way to do this mathematically is not to do a strict 2.38% reduction each year, but rather to signal long-term goals and do year-on-year adjustments.  That way, you can accomodate any load growth (e.g., if total MWh grow, you need to build that into the formula as well) and allow yourself to be surprised if the emissions fall faster than anticipated.  As I'm virtually sure they will, once you make it worth people's financial efforts to do so.

2. It's a good point, but over-complicates.  Most GHG inventories get very confused on the thermal side, for exactly this reason.  (Indeed, I know of none other than those we've assembled that even break CO2 down into three sources: power, heat & transport.  Much more common is power, transport and... cement, steel, fertilizer, residential heating, commercial heating, chemical synthesis, etc. etc.)  The reason why it can be simplified is that within industrial processes, the overwhelming majority of fuel is used in boilers, furnaces and pre-heaters that all run within a pretty narrow band of efficiencies, from 70 - 85%.  Very few of them actually meter their output, but could be easily retrofit to do so.  (The one exception is for feedstocks - like coke in the steel industry, or natural gas in the ethylene industry - that are inputs to the product, such that their carbon is physically entrained in the finished material.  But these are easy enough to exempt.)

3. I've had a lengthy exchange with Gar on this site on exactly that point.  He agrees with you, and my feeling is that it should be exempted.  The crux of my logic is that there is no easy way to measure the carbon impact of your negawatt investments, for the simple reason that they aren't burning fuel.  If you're going to true-up on carbon, you need to pay more to the guy in WV than the guy in ID, since the latter has a much lower grid-carbon signature.  Clearly, you have an economic incentive from lower fuel costs (and it bears noting that mathematically, even in a fully-priced carbon market, the majority of the fuel cost is still fuel).  Equally clearly, there are plenty of non-carbon benefits of energy efficiency.  As such, I'd argue that you are better off pricing carbon at the point of it's release, and then maintaining any number of other incentive plans for end-use efficiency rather than try to force fit all those incentives into a carbon bill that is mathematically challenged from the get-go.

4. Same point as previous.

5. It's a fair point, but you kind of have to pick your poison.  If you want to put an immediate penalty on carbon emitters (e.g., no grandfathered coal plants), it's logically inconsistent not to credit low-GHG sources that already exist as well.  Yes, this gives them windfall profits.  But politically, you may well need their windfall profits to counter the inevitable lobbying against non-grandfathering.  My personal feeling is that I'm a lot more concerned about getting GHG regs in place and GHG down than the size of Exelon's dividend, and so I'd favor the everybody-in poison over the everybody-out poison.  Others may disagree of course, but to my mind, anything that starts with grandfathering provisions is starting in the wrong place - and I don't see any political way to grandfather one side of the ledger but not the other.

I forgot about the lifecycle footprint of commodities. The simplest example is fossil fuel production. A facility that efficiently uses heat to transform shale into gasoline should not get more credit than a producer that inefficiently refines conventional oil. A comprehensive carbon policy would have to find a way to look at the entire footprint of fossil fuels (and other commodities) from well/mine to finished product.

Again, the key is the point of carbon release.  Take your two facilities.  One is making heat inefficiently and the other is making it efficiently.  Step one in intelligent policy is to get them both to chase efficiency and - since the OBS is only based on CO2 released - the entrained carbon is then passed along to be paid for by someone else in the system.  (e.g., it's not suddenly gifted out.  That shale process pays once for all the CO2 required to turn it into oil, and then again when you burn the gasoline, and therefore has an innate cost disadvantage against another process that has less overall CO2 release.)  In essence, the lifecycle CO2 is factored into an OBS much more simply than in any other regulatory model, since every point in the economy where CO2 is emitted includes an equivalent cost per ton of carbon.  (As compared to other models that only price once and assume that the cost will be passed through perfectly in all downstream commodities.)

Note also that entrained carbon in the finished material is not necessarily bad.  It might even be good.  If I turn coal into coke and coke into carbon steel, I've locked that carbon up in a way that's not going to get into the atmosphere.  Similarly, if I turn shale into oil, oil into carbon black and carbon black into tires, I've also locked up the carbon (at least until it gets to a tire-burning power plant at the end of it's lifecycle!)  There are certainly other ecological concerns with shale, but from a GHG perspective, what matters is the point when that CO2 is released into the atmosphere, which is precisely the point at which an OBS regulates.

I don't really understand who owns transmission capacity. I'm guessing its the regulated utilities. Is there a way to incentivize improvements in the grid to remove line losses? I wonder how this could be linked to the carbon markets. I'm guessing that at some point, transmission investments will be competitive with other investments. Would the entire electricity model have to be changed or could this be achieved with the utility owned grid model?

Complicated question.  In vertically integrated states, it is the utility.  In other regions, it is quasi-competitive.  The trouble with line-losses is that are a function of only two variables: current and resistance.  (The total loss is a direct function of the current squared, times the resistance).  As such, there are only two ways to get line-losses down:  Lower the current or lower the resistance.  

The former implies reducing load on the wires, which can be readily done through conservation and/or siting generation closer to the load.  Note that both of those activities are done external to the transmission utility.  The latter implies building more wires to lower the overall resistance of the system.  And if you can make money building those wires, the transmission utilities would be happy to do so.  The key though is that not all of the things we need to do to drive line-losses down are within the transco's purview.

Re: how to factor these into carbon markets, there are actually some novel models.  ISO-New England currently auctions system capacity credit (e.g., you get paid $X/MW of capacity you provide to the system, either as generation or load reduction).  To the extent that you are either putting generation at the load or reducing load, the MW you bring on line are given a bonus factor (I forget the exact number, but it's on the order of 5%), such that you get to sell into the market 1.05x your actual MW reduction.

In an OBS model, the more straightforward way to do this on power is simply to calculate on a delivered basis, such that the local generator gets credit for 100% of their output, but the remote central generator only gets credit for 91% of their output, consistent with 9% average national line losses (or other value as may be regionally appropriate.)

Your point about giving big hydro and big nuclear huge windfall profits is interesting. I never considered the political angle of getting their support and lobby dollars on side to push through the legislation.

On 3 and 4, I think you're right that OBS probably isn't the best solution.

On 2, I'm still not convinced. I read your discussions with Gar and Max on carbon policy 5. (did you have another discussion somewhere else?). I think the crux of the problem is your proposed OBS has sticks and carrots for only the heat and electricity sectors. Improvements in other sectors are incentivized by OBS to the extent that it raises the price of energy (ideally the price of energy won't rise under OBS) meaning that these likely won't happen. This isn't necessarily a bad thing, because the heat and electricity sectors are likely the lowest hanging fruit because they are industries (especially electricity) that have been protected from competition for years. I agree with the capacity factor argument that negawatts on the consumer side are not low hanging fruit. But I think the gray area is industrial commodities.

In an OBS that wants to deal with CO2 from fossil fuels, the only "commodity" that by definition has to be regulated is heat, because that is the useful output of combustion. Such a policy would incentivize efficient boilers in power plants, but would be agnostic to the heat to electricity process. It wouldn't care if natural gas was used to heat water to run through a vapor turbine, or used in a gas turbine. Clearly, this would be a very poor policy because the process of turning heat into electricity is one of the low hanging fruit in creating a low carbon economy and electricity is one of the largest GHG producing sectors. Also, many other paths not involving fossil fuels to make electricity exist and should be incentivized just as much as efficient conversion of heat. The point I'm trying to make is that stopping at the commodity of "electricity" in expanding OBS is arbitrary and the policy might be improved by adding steel, gasoline, aluminum, cement... The emission of GHG's from these sectors is also large, (about an order of magnitude smaller http://ecm.ncms.org/ERI/new/GHG.htm), and there may be low hanging fruit in these sectors as well. Your version of OBS creates a carrot and stick for boilers and electricity generation. This is a policy that "picks winning sectors", electricity and boilers, while treating all other solutions as less important.

The choice which commodities to regulate is somewhat arbitrary. I'd agree that Gar's example of tomato paste is a bit excessive, because it would be a huge mess calculating the savings from one brand of tomato paste to another. In sectors where the output is identical, this would be a lot simpler. I'd argue that its easier to measure output of a commodity than an output of heat. (but we'd lose the added benefit of adding heat measuring that would allow industry to understand how much heat they're wasting). To get an efficient outcome, we need to get as close to economy wide coverage as possible. I'd argue that adding  bulk commodities that produce a certain amount of domestic emissions (perhaps greater than 1%), it might be a worthwhile complication. A side benefit is that commodity producers who have best industry practices would get windfall profits and join the nuclear and hydro industries in supporting and lobbying for this bill. A downside could be some industrial facilities with numerous outputs, like an oil refinery, would be very cumbersome.

So yes I agree that adding more "commodities" to OBS does complicate the bill, but certain large sectors could be added relatively easily and might be worth the effort. Experts from other commodity industries would be able to say how difficult it would be to expand OBS to their situation and the type of GHG reductions that would be incentivized. It might give you more allies in the political arena and lower the total cost of reaching GHG reduction targets. Perhaps this has already been studied, and the low hanging fruit is so overwhelmingly in the electricity sector that my point is moot, but I have yet to see this.

There is an admitted challenge in an OBS dealing with transportation, but that's common to any GHG bill, for reasons I've noted before: namely, the costs of transportation are dominated by capital recovery, while the costs of heat and power are dominated by fuel.  Ergo, a price on carbon (=fuel) isn't likely to shift behavior in transportation, so we need some other measure.  But this is as true for carbon taxes as it is for an OBS.

I'm not sure I understand your agreement on 3 & 4, as you seem to be agreeing that OBS doesn't work for this, but I'm making a different point.  Namely, that the carbon should be priced at the point of release, not downstream or upstream where we have to either assume a GHG impact (if downstream) or assume the price will transmit through the system (if upstream).  This is not to say that there aren't societal benefits from end-use efficiency that gov't policy should reward - simply that those benefits go well beyond carbon and ought to be incented and quantified as such, lest we get into a scenario where we decide that we should only install efficient dryers in high-carbon regions of the country.

Final point: an OBS is not agnostic to a heat-to-electricity process.  Quite the contrary, as if a plant generates electricity, it increases the denominator in a lb/MWh OBS formulation, giving an incentive to turn that heat into something useful, provided it doesn't come with too many lbs.  Note though that you don't get to count heat once and MWh again.  You only get one measure.  (And yes, this requires a bit of audit protection to make sure it is not gamed, but this is nothing that we don't already do in plenty other spheres, from financial accounting to renewable fuels credits.)  But the fact that fuel is turned into heat as an intermediate step en route to power would cause you only to get a credit for the power.

Note also re: downstream commodities that all those commodities can only be produced from heat or power.  

Hopefully that answers without too much confusion - but the points you raise are all ones that are innate to the algebra of the model (if not quite obvious on first read.)

I created this lengthy example to draw the parallel with how the production of commodities are treated. Your proposed OBS is agnostic to how heat is used to refine aluminum, steel, oil... just as my straw man OBS is agnostic to how heat is transformed to electricity. The result of your proposed OBS is that there will be a wealth transfer that will accelerate deployment of efficient boilers in commodity production. Deployment of efficient processes will only be incentivized by the resulting changes in price for heat and electricity.

The question I was getting at is why should the electricity and boiler sectors be allocated extra capital as opposed to investment in improving industrial processes. To get an efficient policy outcome, all reductions in GHG's must be treated as equally as possible.

I've since realized there are a few features about electricity that make it deserve its special place in GHG policy.

1. Electricity is completely fungible. As long as the grid is working properly, I can't tell where my electricity is coming from. A unit of heat is the same, which make standards like CO2/MWhr very simple. Steel has different alloys, different grain sizes that make CO2/ton steel much more complicated. I've come to understand why an OBS that accounted for different commodities would over-complicate things.

2. GHG from the elecricity sector is 4x bigger than those from petrochemicals which is number 2. A GHG policy that puts a special focus on electricity is targeting the right sector.

3. You've shown that the electricity sector is incredibly inefficient because of monopoly regulation. The lowest hanging fruit are there.

4. Electricity can be made in numerous ways not involving fossil fuels such as hydro, wind, solar, something new... Since the electricity sector is so important, we want to incentivize all paths making clean electricity equally. These different paths don't really exist in commodity refining. They all need heat typically comes from fossil fuels.

5. Your point that CO2 is best regulated at the point of release.

However, I've come to agree that attempts to incentivize other reductions through OBS would over-complicate it. Once the lowest hanging fruit are picked (points 2&3 are no longer true), it might be time to come up with new policies. Until then, OBS is the cheapest path to GHG reductions.

Is this a good assessment? I also apologize for quibbling over a minor point.    

Indeed, this already drives behavior.  The industrial who chooses not to invest in high-effectiveness heat exchangers, or insualate their steam lines or recover their condensate (all of which would serve to drive up the use of heat per unit of product) is a rare breed - primarily because those factories have either tightened up or shut down.  An OBS - or any carbon pricing model - may be slightly additive to that pressure, but certainly won't remove or diminish it.

But come back to carbon.  From a GHG perspective, an electric arc furnace connected to a hydro plant has a very different signature than one connected to a coal plant.  On a purely carbon basis, the latter ought to have a greater incentive to conserve than the former, but in reality both will seek to minimize their electric purchases.  We could make a similar comparison with a paper mill connected to coal, gas and oil boilers.  Different carbon impacts, but all have incentives to lower their fuel use per unit of product, which an OBS - nor any other GHG policy model - diminishes.  

But an OBS uniquely amongst GHG policy tools places an equal incentive on all those conversion devices that actually release the CO2.  To your point about all GHGs being treated equally, that's exactly what an OBS does.  The minute you start using a carbon policy to shift downstream behavior, you run into the issue of placing a differential incentive based on upstream use.  

Gas may be a more interesting comparison than coal on this front, since the former has a lower GHG impact but the latter is cheaper.  For permitting reasons, no one's built new coal boilers in a long time - but for economic reasons, no one who has an existing, permitted coal boiler is rushing to replace it with gas.  Putting a price on carbon isn't going to change the fundamental spread between $3/MMBtu coal and $14/MMBtu gas, regardless of how it's done - and as such, if the indexing to efficient end use stuff is tied back to carbon, you don't solve the issues.  (e.g., carbon policy alone is not sufficient to reap the full suite of benefits from efficiency.)

Re: the electric sector's inefficiency, don't use this to presume that there isn't also a lot of low-hanging fruit in the industrial sector.  Industrials do not deploy capital towards non-core investments, and energy is invariably non-core.  This creates a situation where a heat exchanger that transfers heat from one reactor to another (e.g., your downstream efficiency-of-thermal-use stuff) will be built since that increases yield and is core.  But the boiler upstream is energy, and so most industrials own and operate pretty crummy utility islands not because they're stupid, but because their managerial focus on core activities also causes them explicitly not to invest in high-return energy conservation projects in their on-site energy island.

Last point: you're still confusing steel with heat.  The carbon embedded in steel (e.g., carbon-steel) is sequestered, and we don't need to worry about it.  The carbon that is burned in the process of making steel is made for it's heat value.  So a heat-indexed OBS covers the steel industry (and petroleum, and chemicals, and all the others) with a single metric that doesn't get into all the complexities of different products & grades.