In arguments over carbon trading, both sides often assume that past emission trading schemes have been notably successful. But in practice, trading schemes have lowered emissions more slowly than rule-based methods, and have discouraged rather than encouraged innovation. Even in the area where emissions trading shows some success — lowering gross compliance costs to industry — net costs are probably higher than rule-based alternatives.

Compare the success of the often-touted sulfur dioxide trading system the U.S., instituted in 1990, with the speed and quantity of reductions under rule-based systems during the same period. U.S. SO2 emissions dropped by 31% between 1990 and 2001 [1]. Over the same period of time, under old fashioned rule-based regulation, Germany reduced its emissions by 87% [2], Italy by 62% [2], and Western Europe as a whole by 57% [2].

In both absolute and per capita terms, Western Europe and the individual nations within it have less acid rain-producing pollution than the United States [3]. This was not true when they began their regulatory programs in 1982.

Another often-cited success, the U.S. lead trading system, represented only five years out of a 23-year process that began with the requirement that all cars made after 1974 run only on unleaded gasoline. By the time lead trading was introduced in 1982, only 20% of light-duty vehicles still accepted leaded gas. Even so, under the proposed regulations trading replaced, reductions that took until 1987 to complete (thanks to banked credits) would have been required by 1986 [4]. In addition, reporting-error rates were estimated between 14% and 49% — thanks to minimal verification and enforcement of reporting requirements. There were also cases of outright fraud, with purchased credits not matching credits sold [4]. Final elimination of remaining use of lead in gasoline was completed in 1996.

Grist thanks its sponsors. Become one.

In contrast, Japan completed a phase out comparable to the U.S. 1974-1987 process between 1970 and 1980 [5] — about a three year shorter time, and without the enforcement problems that come with a trading system. Slovakia and China were able to eliminate leaded gasoline in less than half a decade via entirely rule-based means — though, as their existing infrastructure investment in lead-using refineries was not on the same scale as rich nations, it is not an entirely fair comparison.

These two examples at least met their targets, albeit more slowly than comparable rule-based regulations. But they were comparatively simple and transparent systems, involving trading among only a few point sources. In contrast, the South Coast Air Quality Management District (SCAQMD) experimented with two programs that were closer to the Kyoto emissions trading in complexity. They involved multiple emission types (N2O and SO2), multiple types of sources (refiner, utilities and cars), and came nowhere close to meeting their defined goals.

The Center for Progressive Reform describes the programs and results well:

In the mid-1990s, SCAQMD launched the RECLAIM program, which allowed utilities and other major stationary sources to trade SO2 and nitrogen oxide (NOx) credits under a cap on total emissions, and the Rule 1610 “Car Scrapping” program, which allowed operators of large stationary sources to buy their way out of compliance with CAA controls by paying owners of old, dirty cars about $600 per vehicle to take them off the road. The RECLAIM program’s cap was set too high, in part because planners based initial allocations of credits on historically higher levels of pollution for covered sources, as opposed to the lower levels of actual emissions at the time the program began. Compounding this error, the program supplanted, as opposed to supplemented existing technology-based requirements, leaving no “safety net” to prevent excessive emissions from individual sources. As a result of these threshold mistakes, in the first three years of its operation, the program resulted in a decrease in actual emissions that was very modest–about 3 percent.

Grist thanks its sponsors. Become one.


Because the initial cap did not create a sufficient scarcity of allowances to motivate covered plants to install pollution controls, few installed controls that would enable them to generate additional allowances as the cap declined. Apparently, most owners and operators concluded that they could purchase credits later, as the cap declined. In fact, at one point, NOx allowances were so plentiful that sources gave 85% of them away for free.

The ultimate calamity for the system came in the spring of 2001, when a short supply of allowances pushed the price of NOx allowances as high as $45,000/ton. In the midst of the hysteria provoked by the California energy crisis, SCAQMD hastily pulled utilities from the system, giving them a three-year grace period to return to compliance with traditional regulatory requirements.

The SCAQMD car scrapping program contained similarly fundamental flaws in design, placing no limits on the amount of allowances stationary sources were able to purchase and failing to supervise the retirement of the cars that supposedly generated emissions reductions. The predictable result was the creation of extreme toxic hot spots containing intolerably high levels of pollution in the neighborhoods located in the vicinity of four marine terminals owned by Unocal, Chevron, Ultramar, and GATX. Exposure to these hot spots resulted in a cancer risk greater than 150 in 1,000,000 for people living in those neighborhoods, the vast majority of whom were people of color. Compounding these problems, SCAQMD auditors found rampant fraud in the program because owners of old vehicles were paid to retire their vehicles, the bodies of the cars were scrapped, but the engines were transferred into other vehicles that kept on running. Further, stationary sources appear to have underreported their emissions significantly, in order to save money by purchasing fewer allowances.

(See also note [6].)

So emission trading has a record of producing slower results than conventional regulation, with at least one example of complete failure to meet a goal. But doesn’t the increased flexibility at least encourage innovation? The empirical record says no:

The best empirical study of sulfur trading to date (“Regulation as the Mother of Invention: The Case of SO2 Control,” Margaret Taylor, Edward S. Rubin, David A. Hounshell, Law and Policy 27, No 2, April 2005, pp. 348-78, p. 372.) says …

… the majority of the performance and capital cost improvements in the dominant technology to achieve SO2 control occurred before the 1990 CAA …


Consequently, the weight of evidence of the history of innovation in SO2 control technology does not support the superiority of the 1990 CAA–the world’s biggest national experiment with emissions trading–as an inducement for environmental technological innovation, as compared with the effects of traditional environmental policy approaches. Repeated demand-pull instruments, in the form of national performance-based standards, along with technology-push efforts, via public RD&D funding and support for technology transfer, had already clearly facilitated the rapid maturation of wet FGD system technology that diffused from no market to about 110 GWe capacity in twenty-five years. In addition, traditional environmental policy instruments had supported innovation in alternative technologies, such as dry FGD and sorbent injection systems, which the 1990 CAA provided a disincentive for, as they were not as cost-effective in meeting its provisions as low sulfur coal use combined with limited wet FGD application

So SO2 emissions trading helped produce no major innovations, and actually provided a disincentive for technologies on the verge of maturation.

What about lead trading? It probably neither speeded nor slowed development of technology — since as with sulfur, mature technologies were already in place when trading was instituted, and unlike sulfur, there were no near-term alternatives that could be discouraged. However, credits from refineries with low lead emissions probably caused delays in implementing zero lead technologies because credits were available from low lead production [4]. In addition, the general delay caused by the trading system also probably delayed some deployment of other improvements.

The RECLAIM fiasco had obvious consequences; if it drastically slowed emissions reductions and failed to meet targets, it is unlikely it provided much incentive for innovation. Most major point sources did absolutely nothing, innovative or otherwise, relying on continued availability of credits instead. In short, a badly designed emission trading scheme resulted in the classic free rider scenario; a great many players did nothing, assuming others would reduce emissions enough to provide credits for them to buy. There were promising technologies being financed by auto registration fees, whose failure RECLAIM may have contributed to. These included low emitting burners and turbines. There was also an emerging NOx reduction technology, SCONOx, which was “more expensive than the dominant selective catalytic reduction method, but arguably could have penetrated the market if there had been stringent regulation generating less ‘spatial flexibility ‘ about where reductions were made.” [7]

In general, it is not surprising that emission trading discourages innovation. The whole point of spatial flexibility is to encourage use of all cheap means before turning to expensive ones. Simple procedures like using low sulfur coal will usually be cheaper than, say, replacement of coal plants with natural gas plants or wind turbines.

Emission trading does have one advantage: It apparently does lower gross compliance costs to polluters. (For example, it is generally agreed that lead trading saved polluters 20% compared to conventional regulation.) Of course some of that savings comes from trading rules being easier to cheat, rather than being easier to obey.

But more importantly, as Amory Lovins is fond of pointing out, we need to count all costs, and all benefits — to make decisions on the basis of the net, not the gross. Net costs with emission trading are probably not lower in most cases. And the cost/benefit ratio is probably worse in every case.

The dynamic cost of postponing technology innovation is going to be higher in most cases than developing the technology earlier. Normally the long-term trend in standards is to tighten them. If regulations or carbon taxes (or a combination of both) force development and deployment of emissions reduction technology in a few cases that can meet standards no other way, then the price will be lower at a later stage when standards tighten and it must be deployed more widely. Emission trading, on the other hand, by postponing such developments, often will ensure that when tighter standard go into effect the technology is not as mature, and thus much more expensive at the point when it has to be put into widespread use. (In the case of carbon credits, the intention is to continually tighten standards — so this argument applies even more strongly here than in other cases.)

Secondly, there is the question of total benefits. There are several ways conventional regulation tends to produce more benefits than emission trading:

  • Overshoot and efficiency gains: In the absence of emission trading, it is hard to comply exactly with a stringent regulation. If compliance requires significant capital investment, normally a company will look for something that provides more efficient production (either reducing inputs or increasing outputs) as a side effect of complying. The result often is an emission reduction greater than required, along with an improved process. I think you could argue that U.S. deregulation madness is part of the reason a lot of U.S. industries are losing competitiveness — not just to nations like China that have access to cheap labor, but high-wage countries which use more effective technologies.


    In addition, emission levels are often set more by political pressure than scientific evaluation — meaning that they are almost always too low. Overshoot, just on the emission level alone, probably provides additional benefits far outweighing additional costs — in terms of additional lives saved at low additional cost.

    Under emissions trading, any company that produces excess reductions will either sell them to another company, or bank them for future use. Thus trading systems almost never end up with a net long-term overshoot.

  • Reductions in other pollutants: most solutions that respond to stringent regulation reduce more than one pollutant. For example, substituting natural gas for coal, or wind for natural gas, reduces not just SOx, but NOx, CO2, and even methane.


    One way Germany achieved such a big sulfur reduction was a much greater implementation of wind than in the U.S. — in spite of having lower quality wind resources. Further, they began this before the price of wind technology dropped to levels competitive with natural gas. So they spent a lot more per unit of reduction than the U.S. did under sulfur trading. But Germany greatly reduced its greenhouse emissions, and air pollution of all types. And it also competes with Denmark for the large scale wind turbine market — something the U.S. does not do very well. The U.S. sulfur trading market was probably the cheapest choice we could have made. But can anyone argue it had a better cost/benefit ratio than the German investment?



[1] David Stern, “Reversal of the trend in global anthropogenic sulfur emissions”, Global Environmental Change 16 (2006) 207-220. David Stern’s paper (PDF) describes how he derived data for a database of human made SOx emissions from 1850 – 2001.

Stern’s spreadsheet for North American data is made freely available at (Excel Spreadsheet).

[2] David Stern’s spreadsheet for Western European data is made freely available at (Excel Spreadsheet).

Bless all analysts who put years into gathering their data and then freely share that data. Because anyone can have an opinion …

[3] I simply combined David Stern’s numbers with midyear population from the U.S. census International Database to derive per capita figures.

[4] Regulation of Fuels and Fuel Additives; Banking of Lead Rights, 50 Fed. Reg. 13116, 13119 (Apr. 2, 1985) for regulation that would have required phase out by 1986; for banking lasting through 1987 – Robert W. Hahn & Gordon L. Hester, Marketable Permits: Lessons for Theory and Practice; U.S. GAO, Vehicle Emissions: EPA Program to Assist Leaded-Gasoline Producers – David M Driesen, “Does Emissions Trading Encourage Innovation?”, Environmental Law Review, 33 ELR 10094 1-2003 Supra Note 143 – PDF

[5] Magda Lovei, Phasing out lead from gasoline: worldwide experience and policy implications, World Bank Technical Paper No. 397, Pollution Management Series, The International Bank for Reconstruction and Development/The World Bank Washington D.C., January 1998. Page 15. PDF

[6] Anne Egelston, Maurie J. Cohen, “California RECLAIM’s market failure: lessons for the Kyoto Protocol“, Climate Policy Volume 4 No.4, 27/May/2005.

[7] Larry Lohmann, Carbon trading, a critical conversation on climate change, privatization and power, development dialog no. 48, September 2006, Dag Hammarskjöld Foundation, Uppsala, Sweden. P109. Book-Size PDF

Reader support helps sustain our work. Donate today to keep our climate news free.