Cap-and-Trade: A Fly in the Ointment? Not Really

For more than two decades, environmental law and regulation was dominated by command-and-control approaches — typically either mandated pollution control technologies or inflexible discharge standards on a smokestack-by-smokestack basis.  But in the 1980s, policy makers increasingly explored market-based environmental policy instruments, mechanisms that provide economic incentives for firms and individuals to carry out cost-effective pollution control.  Cap-and-trade systems, in which emission permits or allowances can be traded among potential polluters, continue today to be at the center of this action.

Most recently, this has been in the context of deliberations regarding possible U.S. actions to reduced carbon dioxide and other greenhouse gas emissions linked with global climate change, as in HR 2454, the Waxman-Markey bill approved by the U.S. House of Representatives, as well as in proposals developing in the Senate.  (I have written a number of blog posts on this topic.  If you’re interested, please see:  “Opportunity for a Defining Moment” (February 6, 2009); “The Wonderful Politics of Cap-and-Trade:  A Closer Look at Waxman-Markey” (May 27, 2009); “Worried About International Competitiveness?  Another Look at the Waxman-Markey Cap-and-Trade Proposal” (June 18, 2009); “National Climate Change Policy:  A Quick Look Back at Waxman-Markey and the Road Ahead” (June 29, 2009).  For a more detailed account, see my Hamilton Project paper, A U.S. Cap-and-Trade System to Address Global Climate Change.)

But the transition from command-and-control regulation to market-based policy instruments has not always been easy.  Sometimes policy can outrun basic understanding, and the claims made for the cost-effectiveness of cap-and-trade systems can exceed what can be reasonably anticipated.  Among the factors that can adversely affect the performance of such systems are transaction costs.

In general, transaction costs — those costs that arise from the exchange, not the production, of goods and services — are ubiquitous in market economies.  They can arise from any exchange:  after all, parties to transactions must find one another, communicate, and exchange information.  It may be necessary to inspect and sometimes even measure goods to be transferred, draw up contracts, consult with lawyers or other experts, and transfer title.

In cap-and-trade markets, there are three potential sources of transaction costs. The first source, searching and information-collection, arises because it can take time for a potential buyer of a discharge permit to find a seller, though — for a fee — brokers can facilitate the process.  Although less obvious, a second source of transaction costs — bargaining and deciding — is potentially as important.  A firm entering into negotiations incurs real resource costs, including time and/or fees for brokerage, legal, and insurance services.  Likewise, the third source — monitoring and enforcing — can be significant, although these costs are typically borne by the responsible governmental authority and not by trading partners.

The cost savings that may be realized through cap-and-trade systems depend upon active trading.  But transaction costs are an impediment to trading, and such impediments thereby can limit savings.  So, transaction costs reduce the overall economic benefits of allowance trading, partly by absorbing resources directly and partly by suppressing exchanges that otherwise would have been mutually (indeed socially) beneficial.  But when transaction costs can be kept to a minimum, high levels of trading — and significant cost savings – are the result.

Since David Montgomery’s path-breaking work in 1972, economists have asserted that the post-trading allocation of control responsibility among sources and hence the aggregate costs of control are independent from the initial permit allocation.  This is an extremely important political property, but does this still hold in the presence of transaction costs?  This is a question I investigated in an article titled, “Transaction Costs and Tradable Permits,” which was published in the Journal of Environmental Economics and Management in 1995 (and which the publisher lists as one of the ten most cited articles in the journal’s history, going back to 1974).

The answer to this question is: “it depends.”  If incremental transaction costs are independent of the size of individual transactions, the initial allocation of permits has no effect on the post-trading allocation of control responsibility and aggregate control costs.  But if incremental transaction costs decrease with the size of individual trades, then the initial allocation will affect the post-trading outcome.

This is of great political importance, because it means that in the presence of transaction costs, the initial distribution of permits can matter not only in terms of distributional equity, but in terms of cost-effectiveness or efficiency.  This can reduce the discretion of the Congress (or other legislature or agency) to distribute allowances as they please (in order to generate a constituency of support for the program), and may thereby reduce the political attractiveness and feasibility of a cap-and-trade system.

Empirical evidence, however, indicates that transaction costs have been minimal, indeed trivial, in enacted and implemented cap-and-trade systems, including the U.S. EPA’s leaded-gasoline phasedown in the 1980s, and the well-known SO2 allowance trading system, enacted as part of the Clean Air Act amendments of 1990.

That’s good news, surely.  But nevertheless, going forward, choices between conventional, command-and-control environmental policies and market-based instruments should reflect the imperfect world in which these instruments are applied.  Such choices are not simple, because no policy panacea exists.

On the one hand, even if transaction costs prevent significant levels of trade from occurring, aggregate costs of control will most likely be less than those of a conventional command-and-control approach.  A trading system with no trading taking place will likely be less costly than a technology standard (because the trading system provides flexibility to firms regarding their chosen means of control) and no more costly than a uniform performance standard.

But the existence of transaction costs may make the choice between conventional approaches and cap-and-trade more difficult because of the ambiguities that are introduced.  With transaction costs — as with other departures from frictionless markets — greater attention is required to the details of designing specific systems.  This is the way to lessen the risk of over-selling such policy ideas and ultimately creating systems that stand the best chance of being implemented successfully.


Policies Can Work in Strange Ways

Whether the policy domain is global climate change or local hazardous waste, it’s exceptionally important to understand the interaction between public policies and technological change in order to assess the effects of laws and regulations on environmental performance.  Several years ago, my colleagues ­- Professor Lori Bennear of Duke University and Professor Nolan Miller of the University of Illinois – examined with me the effects of regulation on technological change in chlorine manufacturing by focusing on the diffusion of membrane-cell technology, widely viewed as environmentally superior to both mercury-cell and diaphragm-cell technologies.  Our results were both interesting and surprising, and merit thinking about in the context of current policy discussions and debates in Washington.

The chlorine manufacturing industry had experienced a substantial shift over time toward the membrane technology. Two different processes drove this shift:  adoption of cleaner technologies at existing plants (that is, adoption), and the closing of facilities using diaphragm and mercury cells (in other words, exit).  In our study, we considered the effects of both direct regulation of chlorine manufacturing and regulation of downstream uses of chlorine.    (By the way, you can read a more detailed version of this story in our article in the American Economic Review Papers and Proceedings, volume 93, 2003, pp. 431-435.)

In 1972, a widely publicized incident of mercury poisoning in Minamata Bay, Japan, led the Japanese government to prohibit the use of mercury cells for chlorine production. The United States did not follow suit, but it did impose more stringent constraints on mercury-cell units during the early 1970’s. Subsequently, chlorine manufacturing became subject to increased regulation under the Clean Air Act, the Clean Water Act, the Resource Conservation and Recovery Act, and the Comprehensive Environmental Response, Compensation, and Liability Act.  In addition, chlorine manufacturing became subject to public-disclosure requirements under the Toxics Release Inventory.

In addition to regulation of the chlorine manufacturing process, there was also increased environmental pressure on industries that used chlorine as an input. This indirect regulation was potentially important for choices of chlorine manufacturing technology because a large share of chlorine was and is manufactured for onsite use in the production of other products. Changes in regulations in downstream industries can have substantial impacts on the demand for chlorine and thereby affect the rate of entry and exit of chlorine production plants.

Two major indirect regulations altered the demand for chlorine. One was the Montreal Protocol, which regulated the production of ozone-depleting chemicals, such as chlorofluorocarbons (CFCs), for which chlorine is a key ingredient. The other important indirect regulation was the “Cluster Rule,” which tightened restrictions on the release of chlorinated compounds from pulp and paper mills to both water and air. This led to increased interest by the industry in non-chlorine bleaching agents, which in turn affected the economic viability of some chlorine plants.

In our econometric (statistical) analysis, we analyzed the effects of economic and regulatory factors on adoption and exit decisions by chlorine manufacturing plants from 1976 to 2001.  For our analysis of adoption, we employed data on 51 facilities, eight of which had adopted the membrane technology during the period we investigated.

We found that the effects of the regulations on the likelihood of adopting membrane technology were not statistically significant.  Mercury plants, which were subject to stringent regulation for water, air, and hazardous-waste removal, were no more likely to switch to the membrane technology than diaphragm plants. Similarly, TRI reporting appeared to have had no significant effect on adoption decisions.

We also examined what caused plants to exit the industry, with data on 55 facilities, 21 of which ceased operations between 1976 and 2001. Some interesting and quite striking patterns emerged. Regulations clearly explained some of the exit behavior.  In particular, indirect regulations of the end-uses of chlorine accelerated shutdowns in some industries. Facilities affected by the pulp and paper cluster rule and the Montreal Protocol were substantially more likely to shut down than were other facilities.

It is good to remember that the diffusion of new technology is the result of a combination of adoption at existing facilities and entry and exit of facilities with various technologies in place. In the case of chlorine manufacturing, our results indicated that regulatory factors did not have a significant effect on the decision to adopt the greener technology at existing plants. On the other hand, indirect regulation of the end-uses of chlorine accelerated facility closures significantly, and thereby increased the share of plants using the cleaner, membrane technology for chlorine production.

Environmental regulation did affect technological change, but not in the way many people assume it does. It did so not by encouraging the adoption of some technology by existing facilities, but by reducing the demand for a product and hence encouraging the shutdown of facilities using environmentally inferior options.  This is a legitimate way for policies to operate, although it’s one most politicians would probably prefer not to recognize.


What Role for U.S. Carbon Sequestration?

With the development of climate legislation proceeding in the U.S. Senate, a key question is whether the United States can cost-effectively reduce a significant share of its contributions to increased atmospheric CO2 concentrations through forest-based carbon sequestration.  Should biological carbon sequestration be part of the domestic portfolio of compliance activities?

The potential costs of carbon sequestration policies should be one major criterion, and so it can be helpful to assess the cost of supplying forest-based carbon sequestration.  This is a topic which I’ve investigated in a series of papers with various co-authors over the past ten years (“Land-Use Change and Carbon Sinks: Econometric Estimation of the Carbon Sequestration Supply Function.” Journal of Environmental Economics and Management 51(2006): 135-152, with Ruben Lubowski and Andrew Plantinga; “Climate Change and Forest Sinks: Factors Affecting the Costs of Carbon Sequestration.” Journal of Environmental Economics and Management 40(2000):211-235, with Richard Newell; and “The Costs of Carbon Sequestration: A Revealed-Preference Approach.” American Economic Review, volume 89, number 4, September 1999, pp. 994-1009.)   Most useful for policy purposes is probably the 2005 report Kenneth Richards and I wrote for the Pew Center on Global Climate Change (“The Cost of U.S. Forest-Based Carbon Sequestration”).  In that report, we surveyed and synthesized the best cost estimates from all available sources.

Human activities — particularly the extraction and burning of fossil fuels and the depletion of forests — are causing the level of CO2 in the atmosphere to rise.  It may be possible to increase the rate at which ecosystems remove CO2 from the atmosphere and store the carbon in plant material, decomposing detritus, and organic soil.  In essence, forests and other highly productive ecosystems can become biological scrubbers by removing (sequestering) CO2 from the atmosphere.  Much of the current interest in carbon sequestration has been prompted by suggestions that sufficient lands are available to use sequestration for mitigating significant shares of annual CO2 emissions, and related claims that this approach provides a relatively inexpensive means of addressing climate change.  In other words, the fact that policy makers are giving serious attention to carbon sequestration can partly be explained by (implicit) assertions about its marginal cost, or (in economists’ parlance) its supply function, relative to other mitigation options.

Among the key factors that affect estimates of the cost of forest carbon sequestration are: (1) the tree species involved, forestry practices utilized, and related rates of carbon uptake over time; (2) the opportunity cost of the land-that is, the value of the affected land for alternative uses; (3) the disposition of biomass through burning, harvesting, and forest product sinks; (4) anticipated changes in forest and agricultural product prices; (5) the analytical methods used to account for carbon flows over time; (6) the discount rate employed in the analysis; and (7) the policy instruments used to achieve a given carbon sequestration target.

Given the diverse set of factors that affect the cost and quantity of potential forest carbon sequestration in the United States, it should not be surprising that cost studies have produced a broad range of estimates.  Ken Richards and I identified eleven previous analyses that were good candidates for comparison and synthesis, and we made their results mutually consistent by adjusting them for constant-year dollars, use of equivalent annual costs as outcome measures, identical discount rates, and identical geographic scope.  We also employed econometric methods to estimate the central tendency (or “best-fit”) of the normalized marginal cost functions from the eleven studies as a rough guide for policy makers of the projected availability of carbon sequestration at various costs.

Three major conclusions emerged from our survey and synthesis.  First, there is a broad range of possible forest-based carbon sequestration opportunities available at various magnitudes and associated costs.  The range depends upon underlying biological and economic assumptions, as well as analytical cost-estimation methods employed.

Second, a systematic comparison of sequestration supply estimates from national studies produces a range of $25 to $75 per ton for a program size of 300 million tons of annual carbon sequestration. The range increases somewhat- to $30-$90 per ton of carbon-for programs sequestering 500 million tons annually.

Third, when a transparent and accessible econometric technique was employed to estimate the central tendency (or “best-fit”) of costs estimated in the studies, the resulting supply function for forest-based carbon sequestration in the United States is approximately linear up to 500 million tons of carbon per year, at which point marginal costs reach approximately $70 per ton.

A 500 million ton per year sequestration program would be very significant, offsetting approximately one-third of annual U.S. carbon emissions.  At this level, the estimated costs of carbon sequestration are comparable to typical estimates of the costs of emissions abatement through fuel switching and energy efficiency improvements.  This result indicates that sequestration opportunities ought to be included in the economic modeling of climate policies.  And it further suggest that if it is possible to design and implement a domestic carbon sequestration program, then such a program ought to be included in a cost-effective portfolio of compliance strategies when the United States enacts a mandatory domestic greenhouse gas reduction program.  Large-scale forest-based carbon sequestration can be a cost-effective tool that should be considered seriously by policy makers.

Of course, this raises the question of whether a policy that will bring about such biological carbon sequestration cost-effectively can be developed, whether as part of a cap-and-trade system, a related offset scheme, or through some other policy mechanism.  That is a question without easy answers (as I’ve noted in a previous post on the Waxman-Markey legislation), but the cost analyses I’ve reviewed in this post suggest that it is important to explore possible ways of incorporating biological carbon sequestration in future U.S. climate policy.