According to the Coase theorem if a property right is well defined and transaction costs are low

Excess Costs through Initial Allowance Allocation

Standard textbook treatments of emissions trading refer to the Coase theorem in declaring that the initial allocation of emission allowances has no efficiency implications but only distributional consequences. With the hybrid EU emission regulation design, however, where not all sectors of the economy are covered through the EU ETS, the Coase theorem no longer applies. Responding to emission market segmentation, member states may strategically partition their national emissions budgets between trading and nontrading sectors. For example, a country that expects to be a net seller of allowances can find it profitable to restrict the number of allowances issued in order to raise the equilibrium price of allowances in the market. Due to the partial coverage of the EU ETS, such terms-of-trade considerations simultaneously induce different marginal abatement costs across countries in the sectors outside the EU ETS (as well as between the non-ETS segments and the ETS). Although such strategic partitioning could be a source of efficiency losses, the applied economic analysis of the actual NAP indicates little scope for price manipulation.

Another more important source for efficiency losses induced by the differentiation between ETS and non-ETS sectors in the first and second periods appeared to be political economy pressures by specific interest groups. The segmentation of the national emission budget between ETS and non-ETS sectors in connection with the free allocation of emission allowances to the ETS sectors provided lobbying incentives for politically influential energy-intensive firms. During the first period, EU member states handed out at least 95% of emission allowances for free to their ETS installations. In the second period – from 2008 to 2012 – this threshold has been reduced to 90%. In phase 1, less than 1% of allowances was auctioned; in phase 2, this fraction increased just to about 4%. Even though the EU Commission had the authority to check the compliance of NAPs with respect to the Kyoto targets as well as concerns on efficiency losses and competitive distortions, the member states ended up with an overallocation of allowances into the trading sectors. In other words, the emission reduction requirements for ETS installations had been set too low from an efficiency perspective vis-à-vis the residual reduction targets for non-ETS sectors in order to comply with the economy-wide emission reduction commitment under the Kyoto Protocol. The overallocation to the ETS became apparent toward the end of the first trading period when the allowance price dropped to zero.

Free allocation of emission allowances may also distort production decisions of firms; as grandfathering was the preferred method of allocation in the first trading period, it was not obvious upon which historical emissions the allocation will be based in the second or future periods. Firms anticipating that their current emissions choice may have an impact on the allowances allocated to them in future periods have an incentive to increase output above the economy-wide efficient level. Similarly, the anticipation that closing firms will no longer receive allowances created a similar distortion to operating decisions.

The free allocation of allowances to firms had distributional consequences that cumulated in a fierce debate on windfall profits of electricity companies. It came to the surprise of the public and some politicians that the opportunity costs of emission allowances were passed through in electricity markets, thereby creating a substantial burden to consumers while windfall profits accrued to the companies. The societally controversial though economically fully rational cost pass-through of emission allowance prices enforced the decision of the EU Commission to increase the share of auctioned permits in the third period. Similarly, grandfathering is displaced by benchmarking as the preferred allocation method.

Coasean Bargaining in the Laboratory

The challenge is to determine how far the assumptions of the Coase Theorem deviate from reality. It is difficult to empirically determine total transaction and bargaining costs, or how much information each party has about the other's benefits and costs from an activity. Experimental evidence now exists that confirms that the rules matter in generating efficient outcomes in Coasean bargaining. The evidence suggests that experimental testbeds can generate experience and data on the efficiency of bargaining rules and protocol. These data are useful to the policymakers interested in how applications of environmental decision-making processes affect outcomes and behavior.

Beginning with the independent work of Elizabeth Hoffman and Matthew Spitzer, and Yvew Prudencio, researchers have used the experimental economics laboratory to investigate the boundaries of the axioms that characterize the Coase Theorem. The assumptions required for the Coase Theorem to hold include (1) two parties to an externality, (2) perfect information regarding each agent's production or utility functions, (3) competitive markets, (4) no transaction costs, (5) costless court system, (6) profit-maximizing producers and expected utility-maximizing consumers, (7) absence of wealth effects, and (8) parties will arrive at mutually advantageous bargains when no transaction costs are present. Regan stated that the presence of assumption (8) alone is needed to prove the Coase Theorem.

Donald Regan argued that the assumptions of the Coase Theorem imply that parties to an externality will agree on a Pareto-optimal level for an externality-generating activity, and that the attainment of this agreement is generated through a mutually advantageous bargain between the parties. This assertion has largely framed the analytical structure by which the results derived in Coasean bargaining experiments are interpreted and presented; this framework is expected to generate efficient agreements in which the parties to the bargain achieve a mutually advantageous distribution of wealth. First, a review of the existing Coasean bargaining experiments is warranted.

Numerous experiments in the psychology arena throughout the 1970s investigated two- and three-person bargaining games, producing a literature that is arguably richer than the Coasean bargaining literature of today. Many of these experiments incorporated one or more of the axioms that typify Coasean bargaining. Although Pareto-optimal choices were generally more prevalent as conditions lined up more closely with the Coase axioms, no consensus emerged on how the gains from trade were to be divided among the parties to a bargain in these pre-Coasean experiments. Because the predicted mutually advantageous distributions of wealth did not occur in many of these bargaining experiments, the notion of rationality was challenged in theoretical and experimental economics.

Economic experiments probing the boundaries of Coasean bargaining began to emerge when an outside option became available for bargainers. An initial series of papers investigating the Coase bargaining problem in the laboratory was concerned primarily with relaxing certain axioms of the Coase Theorem in an effort to better understand the impact on efficiency and distribution of wealth. Hoffman and Spitzer, and Prudencio established the framework for these earliest experiments. The first result, and one that is generally seen throughout subsequent Coasean bargaining experiments, is that parties to a bargain can often develop a negotiated agreement that is efficient. This singular feature of these first Coasean bargaining experiments lent initial support to the first of two key behavioral outcomes implied by Coasean bargaining – namely, that two parties will agree on a Pareto-optimal level of an externality. Efficiency was the typical outcome in almost all of these agreements.

The second result differed considerably between the two original Coase bargaining experiments. Bargainers in the Hoffman and Spitzer experiment were characterized by settling on a distribution of wealth that was equal between the parties, while the bargainers in the Prudencio experiment focused more on giving the available gains from trade to the party that did not hold the property right. These results are different from each other largely because subjects in the Prudencio design only had knowledge of their own payoffs whereas the Hoffman and Spitzer design allowed each player to have complete information about the payoffs of both players. The Hoffman and Spitzer design suggested that the distribution of wealth was equitable (i.e., 50/50 split), whereas the Prudencio design was characterized by constrained self-interest (i.e., 80/20 split). The player with unilateral property rights took for herself what was available from the outside option and gave the rest to the other player. These results are consistent with those reported in ultimatum games.

The results of those first experiments had significant implications regarding the experimental validity of the distribution and rationality implications of the Coase Theorem. As such, Hoffman and Spitzer conceded that while their results did not support the precise implications of Coase's hypothesis regarding the distribution of wealth in almost all of the bargaining agreements, these discrepancies were not significant enough to detract from support of the Coase Theorem. But the fact remained that ambiguous results existed regarding the ability of bargainers to reach mutually advantageous bargaining agreements in a Coase setting. Experiments that provided subjects with a more complete understanding of the meaning of unilateral property rights observed behavior offered strong support for both the behavioral implications of the Coase Theorem. Additionally, Hoffman and Spitzer determined that moral authority and allocation mechanisms had no adverse impact on efficiency, but provided more mutually advantageous bargaining agreements. Their work established a general experimental design to assess the robustness of the most important axioms of the Coase Theorem in terms of efficiency and distribution.

With an experimental design now in place, further investigations probed the boundaries of the Coase Theorem. These papers suggested that the Coase Theorem was robust in its assumptions – efficiency was not dampened by large bargaining groups, asymmetric payoffs, or requiring contractual consent from the other player. Additionally, a Coasean bargaining setting generated efficient solutions to externalities in experimental markets. While the distribution of wealth could be manipulated by relaxing some of Coase's axioms, the evidence supported the two key behavioral outcomes implied by the Coase Theorem.

Once these experiments affirmed the importance of the laboratory approach to support the axioms of the Coase Theorem, research efforts refined the general design of the Coasean bargaining experiment to include a binary lottery to control for subjects' risk preferences. This refinement allowed for greater precision in studying the robustness of Coase's axioms by relaxing key underlying assumptions. Uncertain payoff streams had no impact on predicted efficiency, while imperfect contract enforcement did. Further, institutional structure played a key role in predicting the distribution of wealth among bargainers. Mutually advantageous distributions became more common with membership on a team directing loyalty and in a tournament setting with nonlinearly increasing payoffs.

These first Coasean bargaining experiments directed the flow of research from the laboratory to the real-world. But research is now motivated and framed in more direct reference to the resulting policy implications. Conflicts in the environmental arena are now the focus of current Coasean bargaining experimental efforts. Reflecting the real-world concern that delay can erode bargaining gains in an environmental conflict, Jason Shogren's work determined that efficiency was hindered by nonincreasing marginal delay costs. Some types of bargaining rules can further enhance efficiency of the private resolution of environmental conflict by serving as a substitute for bargaining experience.

Testbedding research holds the most immediate promise for Coasean bargaining experiments in shaping dispute resolution in the environmental arena. By providing a source of experience and data about how various rules might work in a collaborative setting, testbed research offers an initial evaluation of the performance properties of the rules to be examined in this Coasean bargaining setting and becomes necessary to design and refine institutional collaborative frameworks. With a focus on the impact of specific rules on efficiency and distribution in resolving environmental conflicts among private citizens, Coasean bargaining experiments can suggest efficient policy mechanisms in this arena. Just as the experimental laboratory is being used to study feasibility, limitations, incentives, and performance of proposed markets designs for deregulation in natural gas and electric power systems, the experiments presented in this chapter are intended to further refine alternative decision-making processes currently in use in the environmental arena using efficiency as the evaluation tool.

Concluding Comments

Although Maskin’s analysis clarified the nature of the bargaining problems in the presence of externalities and private information, it hardly rescues the Coase Theorem. On the one hand, when the number of agents is small, some or all of them are likely to have market power. They may not agree to pseudo-competitive fees, in which case the bargains are unlikely to be Pareto efficient. On the other hand, externalities that affect a large number of agents are likely to give rise to the free-rider problem. (Transaction costs are also likely to prevent efficient bargains.) The models presented in this section indicate that the solution in either case involves some form of government coercion. The combination of voluntary bargaining, individual rationality, and private information is unlikely to produce Pareto-efficient outcomes, contrary to Coase’s expectations for bargained solutions to market failures.11,12

These difficulties with the Coase Theorem notwithstanding, one should not end a discussion of the Theorem without giving Coase his due. Coase’s insight that the assignment of property rights in situations involving externalities can potentially lead to efficient outcomes is important. Among other things, it is transforming the commercial fishing industry.

Fishing stocks were on the verge of collapsing in the last quarter of the 20th century for a number of commercial fish such as cod and halibut. Governments responded by establishing commissions to regulate the catch in the fishing beds under their jurisdiction to maintain the stock of each fish at a sustainable level. The fishing for Pacific halibut off the coast of British Columbia is a case in point, although a large number of examples throughout the world serve just as well.

In 1923, the U.S. and Canadian governments established the International Pacific Halibut Commission (IPHC) to oversee the fishing of halibut. When the stock of halibut off British Columbia became dangerously low in the 1970s, the IPHC established a series of regulations to keep the stock of halibut sustainable. They began by issuing a fixed number of licenses—435, one per vessel—to limit the number of vessels that could fish the halibut beds. However, technical improvements in long-line gear—the primary method used to catch halibut—made the vessels so productive that the stock of halibut continued to decline. The IPHC responded with a limit on the total allowable catch (TAC) by all of the vessels combined, a limit designed to keep the remaining stock sustainable. The TAC did not work well at all. Once the fishing season began, each fisher had an incentive to go out every day and bring in as much fish as possible until the TAC was reached. Given the productivity of the vessels, the TAC was reached in six days. This meant that the vessels were often going out in rough seas and fishing close together, conditions that made a hazardous occupation that much more dangerous. Moreover, with all of the halibut caught within a week, the fishers were forced to sell their fish to the fish processors, which gave the processors monopsony power over the fishers.

Exasperated by the results of the aggregate quotas, the IPHC turned in 1991 to a system of Individual Transferable Quotas (ITQ) that was first adopted by New Zealand in 1986. Each licensed vessel was given an ITQ that determined the amount of halibut the vessel could bring in during the fishing season. The quota under an ITQ was determined by a formula based on the length of the vessel and its catch in the preceding four years. The sum of the individual quotas under the ITQs equaled the sustainable TAC for the halibut bed. In addition, the ITQs were split into two shares, and each share could be sold to other vessels, including those already licensed, although no one vessel could hold more than four ITQ shares.

The ITQ approach is very much a Coasian solution because it essentially gives each fisher a property right over part of the halibut catch. It also had predictable beneficial results. Fishers no longer had to rush to catch their fish. As a result, the fishing season was extended from six days to many months. Fishing became much safer. The fishers could avoid bad weather, and the number of vessels with ITQs decreased 29% from 435 to 309 from 1991 to 1994 as some fishers sold their ITQs to more efficient vessels and left the business. Moreover, fishers could bring fresh fish to market over a much longer period, which meant that they could receive higher prices for fresh fish rather than being forced to sell almost all of their catch at much lower prices to the fish processors. The total revenues for the fishers increased by $23 million from 1991 to 1994. The profits would have been even greater had there been no restrictions on the number of ITQ shares one vessel could have, but the IPHC instituted the limit to allay fears that a handful of the largest, most efficient vessels would buy up all the shares and drive the smaller vessels out of business. The goal was to preserve a long-standing way of life for the small fisher. However, this goal is misguided from a purely economic perspective because a larger vessel that really is more efficient can offer the smaller fisher a price for his or her two ITQ shares that would exceed the present value of the smaller fisher’s annual net income from fishing (Grafton et al., 2000; the data on the change in the number of vessels and the increase in revenues from 1991 to 1994 is on p. 689).

In summary, the establishment of marketable property rights to fishing beds is an instance in which assigning property rights to control for an externality (i.e., overfishing) is implementable, and it works about as well as the Coase theorem suggests that it would.

4.6 Microbehavioral models in the economics of global public goods

To the readers who have read through Chapter 2, The Economics of Public Goods and Club Goods, and Chapter 3, The Economics of Global-Scale Public Goods: Key Challenges and Theories, that dealt with the theories and concepts of the public goods developed during the seven decades since Samuelson’s seminal papers, this chapter may have come as a total surprise to some and a refreshing perspective to others (Samuelson, 1954). The microbehavioral theory, models, and applications elucidated in this chapter certainly provide an alternative, thought-provoking theory on the economic problems of public goods.

At the core, the traditional theories of public goods, also presently predominant, are centered around the creation of a correct price, often referred to as a shadow price, for the public good that is under consideration (Samuelson, 1954; Buchanan, 1965). The price should be enforced across the macro economy through a legal vehicle, which will ultimately resolve the problem of the public good itself, national or global, that is, restore a Pareto optimality of the economy.

As such the variables and parameters of the empirical models of the classical theories such as the dynamic integrated climate and economy (DICE) model are all macro-level variables and parameters: to name some, global welfare, global rate of time preference, global production, global damage, global abatement cost, global average temperature, and global emissions of carbon dioxide (Nordhaus, 1994).

In an utterly counterintuitive and surprising manner, the microbehavioral model clarified in this chapter builds a theory and model of a global-scale public good whose central character is an individual who is forced to make numerous optimal decisions for the management of her resources (Seo, 2006, 2016a). This chapter has shown, especially through the empirical applications, that the microbehavioral decisions can play a pivotal role in addressing the problems of a global-scale public good. More specifically, the host of microbehavioral decisions explained in this chapter provides one of the mechanisms through which the humanity can cope with the problems of a global public good, that is, a global climatic change.

A critic and a traditional theorist may ask: will the microbehavioral decisions such as the ones described in this chapter be sufficient for the humanity’s dealing with the problems of global warming efficiently? Although this chapter does not provide a satisfactory answer, it does indicate the possibility and even present the roads which ultimately would lead to the answer. The ensuing chapters, especially Chapter 7, The Economics of Globally Shared Goods, will prove quantitatively that an ensemble of microbehavioral decisions can successfully handle the problems of global warming.

The microbehavioral theory and model of global public goods are counterintuitive, as such, largely overlooked in the economics literature for the past seven decades, because the problem of a public good by definition occurs at the public level and therefore the intuition tells that it should be solved at the public level. For this obvious reason, the classical theorists of public goods have all taken the macro economy approach (Nordhaus, 2008; Weitzman, 2009; Barrett, 2010).

Take for example the polycentric governance approach for the commons by Elinor Ostrom (Ostrom, 1990). In the study directed at the local commons, she reported a local governance, that is, a rule or a custom, set up by a local community for dealing with the problem of the commons in the locality. In her approach, the local commons problem is solved by the local community through a pool of collective actions. The local governance approach, therefore, lacks an individual decision-maker whose decisions are, when added up across the individuals, powerful enough to ultimately address the problem of the public good.

The microbehavioral theory can be compared with the Ronald Coase’s bargaining approach for the social cost problems (Coase, 1960). The highly acclaimed Coase theorem, which was awarded the Nobel Prize in economics in 1991, encapsulates the thoughts of the Chicago School of economics and the following previously quoted paragraph from Milton Friedman (Friedman, 1962):

An obvious example is the pollution of a stream. The man who pollutes a stream is in effect forcing others to exchange good water for bad. These others might be willing to make the exchange at a price. But it is not feasible for them, acting individually, to avoid the exchange or to enforce appropriate compensation.

The bargaining solution is only effective when there is a limited number of polluters (upstream) and a limited number of victims (downstream), continuing with the language in the above quote. When there is either a large number of polluters or a large number of victims, a voluntarily negotiated solution to the pollution problem becomes nearly impossible. This is perhaps the reason why the Coase theorem is pertinent to the negative externality problem but not to the problem of a public good which by definition involves a great large number of individuals.

A bargaining solution is certainly impossible for the problem of a global public good, more concretely, that of global warming. The number of polluters as well as the number of victims are truly large. The microbehavioral theory and model, by contrasts, is developed to address the problem of a global public good and therefore may be less effective, as far as the explanatory power is concerned, in the problems of negative externalities such as the upstream polluter problem in the above.

At this point, I presume that the readers are greatly intrigued by the presentation in this chapter and especially in this section regarding the microbehavioral model of a global public good. The full arguments of the present author and of this book will be made explicit by Chapter 7, The Economics of Globally Shared Goods, of this book. A swift reader, however, would not have to wait until Chapter 7, The Economics of Globally Shared Goods, to figure out why the microbehavioral theory and model may provide a satisfactory remedy to the problem of a global public good. A short description in the next section would suffice to give her ample insights and unmissable clues to the arguments.

3 Later CLS

CLS also extended the method beyond private law doctrine to the powerful emerging new forms of policy analysis, legal economics, attacking certain core premises of the new science including the Coase theorem, cost-benefit analysis of legal entitlements, public choice theory, and submission to the will of others as rational consent (see e.g., Kennedy and Michelman 1980). Others extended it to institutions: Simon identified opposed models of welfare administration, mechanical rule-application by low-skilled workers and the professional informal discretion of caseworkers. He showed how the welfare system had evolved from discretion to rules with damaging results for clients and argued for restoration of discretion (Simon 1983).

Critics also revived and extended the legal realists' critique of the distinction between ‘public’ (coercive regulatory) and ‘private’ (freely–willed contractualconsensual market) relations. Following the realist R. L. Hale (see Fried 1998 for an intellectual biography), they argued that all ‘free choice’ is made by actors with property and power endowments and under baseline conditions structured by the legal system, even choice in the most intimate spheres of life such as marriage relations. Their aim was to supplement Foucault's (1977) account of the pervasive regulation of social interactions through ‘disciplines’ with an account of the pervasive structuration of transactions and interactions in business and personal life through law.

Much critical effort went into history (see Gordon 1984). Some critical historians worked in the tradition of the social history of subordinated groups, using historical examples to show how ruling groups had used neutral- and equal-sounding legal doctrines to extend and justify their rule. Horwitz (1977) showed how early-nineteenth-century law had reworked private law rules of property, tort, and contract to favor entrepreneurs in the early nineteenth century, then later froze the new rules against redistributive revision by imposing constitutional restraints on legislation. Klare (1978) showed how conservative judges had drained the National Labor Relations (Wagner) Act of the New Deal of its potential as a charter of workers' rights to participate in management. Critical historians often read the past as a repository of lost social experiments, long since crushed or forgotten, which might serve as models for reconstructing aspects of social life—as did Frug with the nineteenth-century model of powerful autonomous city governments (Frug 1980), and Simon did with ‘social–republican’ property models of landholding and producers' cooperatives (Simon 1991). Other histories explored the legal structuration of social relations, especially those of master and slave, employer and employee, husband and wife (see e.g., Steinfeld 1991, Olsen 1983).

By the mid-1980s CLS had entered a new phase. One group of critics, perhaps responding to accusations that CLS was ‘nihilist’ in its attachment to destructive criticism (‘trashing’), turned to concrete programmatic work. Unger's (1998) was once again the most ambitious, proposing large-scale programs for the democratization of the economies of both developed and developing societies; most others were small-scale proposals for reforming specific legal policy fields, such as worker participation in ownership and management (Stone and van Wezel 1988), treatment of workers in plant closings (Singer 1988), regulation of injurious speech (Abel 1994), intellectual property (Boyle 1996), disability law (Kelman and Lester 1997), and property rights in ex-Communist societies (Alexander and Skapska 1994).

Another group focused on developing critical theories of law to a new level of sophistication, integrating them with postmodernist and poststructuralist theories, especially those of Foucault and Derrida (see e.g., Balkin 1987, Heller 1984, Peller 1985).

4.10 Taxes vs Quotas

In the discussion earlier, we have been focusing on a tax as the obvious candidate policy instrument. Indeed the damage externality is a pure externality for which the Pigou theorem applies straightforwardly. What are alternative policies? The Coase theorem applies too as well but it does not seem possible in practice to define property rights for the atmosphere (into which emissions can then be made, in exchange for a payment to the owner). What about regulating quantities? Indeed the “cap-and-trade” system, which is a quota-based mechanism, has been the main system proposed in the international negotiations to come to a global agreement on climate change. A cap-and-trade system is indeed in place in Europe since 2005.bn There is a debate on whether a tax or a quota system is better, and here we will only allude to the main arguments. Our main purpose here, instead, is to make a few basic theoretical points in the comparison between the two systems. These points are also relevant in practice.

Before proceeding to the analysis, let us briefly describe the “-and-trade” part, which we will not subject to theoretical analysis. If a region is subject to a quantity cap—emissions cannot exceed a certain amount—the determination of who gets to emit how much, among the users of fossil energy in the region, must still be decided on. The idea is then to allocate emission rights and to allow trade in these rights. The trading, in theory at least, will then ensure that emissions are made efficiently. The initial allocation of emission rights can be made in many ways, eg, through grandfathering (giving rights in proportion to historic use) or auctions. To analyze the trading system formally we would need to introduce heterogeneity among users, which would be straightforward but not yield insights beyond that just mentioned.

The first, and most basic, point in comparing quotas and taxes is that, if there is no uncertainty or if policies can be made contingent on the state of nature, both instruments can be used to attain any given allocation.bo If a tax is used, the tax applies to all users; if a quota is used, regardless of how the initial emission rights are used, the market price of an emission right will play the role of the tax: it will impose an extra cost per unit emission and this cost will be the same for all users, provided the market for emission rights works well.

Second, suppose there is uncertainty and the policy cannot be made state-contingent. This is a rather restrictive assumption—there is no clear theoretical reason why policies could not change as the state of nature changes—but still an interesting one since it appears that political/institutional restrictions of this sort are sometimes present. To analyze this case, let us again consider uncertainty and an ex-ante period of decisions. To capture the essence of the restriction we assume that the only decision made ex-ante is the policy decision. A policy could be either a unit tax or a quantity cap. We assume that the quantity cap is set so that it is always binding ex-post, in which case one can view the government as simply choosing the level of emissions ex-ante.

The choice between a tax and a quota when there is uncertainty (or private information on the part of “the industry”) has been studied extensively in the environmental literature since Weitzman (1974) and similar analyses are available in other parts of economics (eg, Poole, 1970). One can clearly provide conditions under which one policy or the other is better, along the lines of Weitzman's original paper. Weitzman considered a cost and a benefit of a pollutant, each of which depended on some random variable, and the two random variables were assumed to be independent. He then showed that what instrument would be best depended on the relative slopes of the marginal benefits and cost curves. Follow-up papers relaxed and changed assumptions in a variety of directions, but there appear to be no general theorems that apply in the climate-change application to conclude decisively in one way or the other. In fact, we know of no quantitatively parameterized dynamic model that looks at the issue so what we will do here is simply provide a straightforward example using our simple static model and then discuss a couple of separate, and we believe important, special cases.

For our example, we use one type of uncertainty only: that of the cost of producing fossil fuel, χ. With the calibrated model and a uniform distribution around the calibrated value for χ we obtained the ex-ante utility levels for a range of taxes and for a range of emissions, both committed to before the randomness is realized. Fig. 12 shows the results: a range of tax values around the optimal tax outperform the optimal quota. In this case, the precommitted tax rate is a fixed value. If it could be set as a proportion of output, which is ruled out now by assumption since the tax cannot be state-contingent but output will be, it would be fully optimal also ex-post, since the best tax ex-post is always a fraction γϕ of output. Apparently, the ex-post randomness of output is not significant enough to overturn this result. It is straightforward to look at other types of shocks. Shocks to γ deliver more similar welfare outcomes for (optimal and precommitted) taxes and quotas.

Now suppose that we consider a case of a tipping point and that the uncertainty is coming from energy demand (through, say, a separate, exogenous and random TFP factor) or from the cost of coal production (through χ). If the tipping point is known to be E¯, and Γ(E) is equal to zero for E<E¯ but positive and very high otherwise, what is then the best policy from an ex-ante perspective? Clearly, a policy with an emissions cap would simply be set at E¯, a cap that may or may not bind ex-post: if the demand for energy is low, or the cost of producing it is high, the ex-post market solution will (efficiently) be to stay below E¯, and otherwise the cap will (efficiently) bind. A tax will not work equally well. One can set the tax so that the economy stays below the tipping point, but in case the energy demand is low, or its production costs are high, ex-post, output will be inefficiently depressed. Thus, when we are dealing with asymmetric payoffs of this sort (relative to the amount of emissions), a quantity cap is better.

The previous example would have emissions rights trading at a positive price sometimes and at a zero price otherwise. Thus, the system with a quantity cap leads to a random cost for firms of emitting carbon dioxide (beyond the price the firms pay the energy producers). Variations in the supply of emissions rights, decided on by regulatory action, influence the price of the trading rights as well. The experience in Europe since the cap-and-trade system illustrates these points well: carbon prices have fluctuated between over 30 euro and virtually zero since the system started. Such fluctuations are observed also in other regions with cap-and-trade systems (eg, New Zealand). Clearly, since optimal carbon pricing should reflect the social cost of carbon, such fluctuations are only efficient if the social cost of carbon experiences fluctuations. Damages from carbon emissions are likely not experiencing large fluctuations, but our assessments of how large they are of course change over time as scientific knowledge accumulates. The recent large drops in the price of emission rights can therefore be viewed as problematic from a policy perspective.

A cap-and-trade system could be augmented with a “central emission bank” that would have as its role to stabilize the price of emission rights by trading actively in this market, hence avoiding the large and inefficient swings observed in the EU system. Notice, however, that we would then be very close in spirit to a tax system: a tax system would just be a completely stable (provided the chosen tax is stable) way of implementing a stable price of emissions for firms.bp

3.6.3 Resolution of externalities through bargaining by affected parties

Parties affected by unregulated externalities will sometimes have the opportunity to make mutually beneficial agreements with those who generate the externalities. In the classic example, if a factory’s pollution causes harm of $1000 that can be prevented by installing a smoke scrubber that costs $100, then, in the absence of any legal obligation on the factory, one might expect a potential victim of pollution to pay the factory to install the scrubber. An agreement for any amount between $100 and $1000 would be mutually beneficial. Let us first consider this possibility and then evaluate its significance61.

If it is posited that there are no obstacles to reaching a mutually beneficial agreement concerning externalities, then that will occur. This tautology is one version of the Coase Theorem; Coase (1960) stressed the point that externality problems could be remedied through private bargains. A closely related version of the Coase Theorem asserts that the outcome regarding the externality – whether a smoke scrubber is installed or instead pollution is generated – does not depend on the legal rule that applies. For example, if the scrubber costs $100 and there is no law that controls pollution, a bargain as we have described it will come about and the scrubber will be installed; and likewise if there is a law that leads to installation of the scrubber, the same will happen62. The outcome, however, might be affected by the legal rule because of the level of wealth of parties. Most obviously, the potential victims might not have assets sufficient to pay for the scrubber, in which case the scrubber would not be installed unless a legal rule leads to this; moreover, legal rules may affect the distribution of wealth and thus the demand for goods, including that of being free from pollution63.

There are, however, many obstacles to bargaining. Bargaining may fail to occur when victims are numerous and face collective action problems in coming together. This is often the situation with respect to victims of industrial pollution. Similarly, in important contexts, bargaining will be impractical because victims will not know in advance who will injure them; this is the case for automobile accidents and most other accidents between strangers. Another reason that bargaining may not occur is that victims might not know that they are exposed to a risk (such as from an invisible carcinogen). Also, of course, the cost of bargaining between just one potential victim and one potential injurer who know of each other can discourage them from engaging in the process. If these reasons do not apply and victims and injurers do engage in bargaining, asymmetry of information may lead to bargaining impasses; for example, where a victim thinks that a smoke scrubber would cost a factory only $50 and it really costs $100, he may offer too little to the factory to reach an agreement. In all, these problems that reduce the likelihood of bargaining occurring, and also its success if it does take place, make the importance of legal rules to remedy externalities substantial.

2.1 The Modern Law and Economics Movement

The modern law and economics movement had its origin in the work of Gary Becker, Ronald Coase and Guido Calabresi. Becker (1968) used simple principles of labor economics to analyze the criminal's decision of whether to engage in crime, and derived implications for optimal penalties and optimal enforcement of the law. Coase's (1960) work established the proposition known as the ‘Coase Theorem’, showing that, in a world of zero transaction costs, the distribution of resources is invariant to the rule of law. However, Coase later made clear that he considers his work concerning how to minimize the effects of transaction costs in assigning the rule of law under the more realistic assumption of positive transaction costs to be more important. Calabresi (1970) wrote a classical study analyzing the costs of accidents and suggesting ways in which tort law might be formulated to minimize those costs. After these initial contributions, a notable and extensive body of work in the economic analysis of law was, and continues to be, authored by Richard Posner, who has applied the principle of wealth maximization to a plethora of legal problems, explaining either how the law can be understood as promoting wealth maximization, or how it might be amended to better achieve that objective (see, for example, Posner 1998). As will be discussed in detail below, in more recent years, work in the economic analysis of law has taken a variety of forms, from the traditional neoclassical analyses to socioeconomic and behavioral economic analyses.

Resolving Environmental Conflicts with Property Rights

The Coasean solution to environmental problems is typically framed as competing with, or an alternative to, the Pigovian solution. In reality, the comparison is slightly more complex. A careful understanding of property rights allows one to distinguish these approaches.

In Coase's famous 1960 article, ‘The Problem of Social Cost,’ he describes a series of bilateral land-use conflicts, largely found in nineteenth century English common law, in which one landowner transmits cost to a second landowner via the environment. These negative externalities include noise, vibration, odors, airborne chemicals, and even the obstruction of natural airflow. The English courts would assign property rights to one party and thereby alleviate the environmental externality, although not the interdependency.

Coase carefully selected these legal cases to show several underappreciated aspects of environmental externalities. First, the cases involved behavior by both parties that was individually reasonable but, together, incompatible. Coase was, therefore, able to break free from the tendency to confound moral judgments about the environment (it is wrong to pollute) from economic analysis about what behavior leads to social efficiency.

Second, Coase's cases shared a startling characteristic; each showed that the party most likely to be labeled ‘polluter’ was also the party that seemed to pursue the use of property that led to the highest social product. In other words, if the ‘victim’ of ‘pollution’ were protected, then a socially inefficient allocation of resources would occur.

Following Coase, other economists summarized his analysis into the ‘Coase theorem.’ One version of this theorem is that when property rights are well defined and transaction costs are low, the final allocation of rights will be the same irrespective of the initial allocation of rights and this final allocation will be efficient. The first aspect is sometimes known as ‘invariance’ because the final allocation of rights is not affected by which party is initially assigned the right. The efficiency aspect means that, after trading, rights will be distributed so as to maximize the value of the rights at stake and no further gains from trade will exist.

In other words, the Coase theorem means that institutions may assign rights to either the ‘polluter’ or the ‘victim,’ but markets will be used to transfer the property to the party who values it the most in the event that the institution failed to do so originally. This distinguishes the highest-value user from the lowest-value user of the unallocated property right.

As a result, the ethical dichotomy of polluter/victim thus becomes less important for economic analysis than the understanding of whose use of property will lead to greater social welfare. In any environmental negative externality situation, there is a party seeking to pursue a low-intensity use of the environment, while another seeks a high-intensity use. In effect, both parties want the same ‘stick’ to be their property, seeking to put that stick to very different ends. It may be that the low-intensity use leads to the highest-valued product for society, or vice versa. The Coase theorem tells us that it does not matter for efficiency which party receives the property right as long as markets can reallocate that right with minimal transaction costs (of course, it matters for distribution).

However, it is the second part of Coase's argument that, while underappreciated, likely applies best to most environmental problems today. Coase argues that when transaction costs are higher, markets may be unable to deliver the right to the party that values it the most. In such cases, social efficiency can only be achieved if institutions assign rights directly to the highest-valued users. Many environmentalists likely feared policies derived from Coase's analysis for this reason; if one assumes that high-intensity users will have the highest-valued use of the environment, then this analysis suggests that no rights ought to be assigned to low-intensity users of pollution. Others, however, feel that if the values of low-intensity users are correctly valued – say, with nonmarket valuation – then the Coasean analysis can be used to warrant institutions that restrain ‘polluters.’

The Coase framing of environmental problems was a direct challenge to the Pigovian framing. In effect, the Pigovian analysis automatically sought to restrain ‘polluters’ by taxing their activities. The Coasean analysis would restrain the lowest-value user, which might be the ‘victims’ in some cases and the ‘polluters’ in others, by assigning rights to the highest-value user. In terms of rights, however, the analyses are somewhat distinct. In the Coasean analysis, rights are to be assigned to formalize a presumptive rights regime, and the issue involves selecting to whom the property rights should go. In the Pigovian analysis, property rights are not assigned in the same way; rather, ‘polluting’ activities are entitled to continue as long as a penalty is imposed for each unit produced. It was not until economists operationalized the concepts in the Coasean analysis as tradable discharge permits, or ‘cap and trade,’ that price and quantity instruments became directly comparable. Quantity instruments are addressed in the following sections.

7.4.3 Institutions and Institutional Change

The thesis that institutions matter for economic growth is widely accepted by economists although there are some disapproving arguments.32 To the extent that institutions are everywhere and provide incentive structures to agents, the thesis is almost tautological. The real question is why and when growth-promoting institutions are adopted in some countries but not in other countries. To a large extent, this can be boiled down to studying the efficiency hypothesis formally formulated by North and Thomas (1973): institutions evolve to explore economic gains. Following this line, Yao (2004) formally shows in the framework of implementation theory that the efficiency hypothesis does not hold under a well-behaved political process without side payments. In reality, though, the political process can be perturbed and cross-group transactions are commonly used to buy support. For example, agents may engage in a Coase bargaining to obtain the institution that maximizes the social output. This is what Acemoglu calls “the political Coase theorem” (Acemoglu, 2003). However, as convincingly argued by Acemoglu et al. (2006), the political Coase theorem rarely holds in reality because the political dynamics often does not allow for the Coase bargaining.

The above concise review of the literature highlights the significance of the Chinese transition for the theory of institutional change. The key to explaining China’s largely peaceful yet efficient transition from the command economy to the market economy lies in two areas. One is the sense of crisis, and the other is the contingent institutions created in the process of transition. This does not mean that other factors are not important; rather, their significance is of second order. The sense of crisis served as the catalyst for the Chinese transition, and contingent institutions have helped draw different groups to the common agenda of reform.

In the mid-1970s, the CCP leadership faced two kinds of crises: one from the outside, and the other from inside. By the mid-1970s, it was clear to the CCP leadership that China had lost the race with not only the advanced capitalist countries, but also its developing neighbors, including its runaway province, Taiwan. To the old generation of leaders, this reminded them of the old saying that had rung in their ears for decades: “Lagging behind is to get bullied by others.” This strong Darwinian belief became one of the impetuses pushing for change. Inside China, the CCP’s legitimacy was withering away. The catastrophes of the Great Famine and the Cultural Revolution had depleted the CCP’s revolutionary dividends and by the mid-1970s, its top leadership had a strong sense of crisis of legitimacy. The drop in agricultural output in 1976–1977 set the alarm that another famine would fall upon the country and led directly to the ensuring rural reform (Yang, 1998). With procedural legitimacy out of the question, the only choice left for the CCP was to gain legitimacy through performance, i.e. delivering tangible benefits to the population. Turning the party’s gravity toward economic growth thus became a national consensus under the leadership of Deng Xiaoping. The reform movement was underway.

To go with the reform, however, there were still many hurdles to overcome. To avoid engendering its own rule, the CCP had to take a gradual approach to reform. This then created two problems in the transition period. One was the resistance of the social and political groups whose interests were tied to the old institutions, and the other was the incongruence of the CCP’s own political institutions and the new economic institutions. To overcome those two hurdles, many contingent institutions were created. A contingent institution arises as a response to solve the most pressing issue facing the decision makers. For that it may have to compromise with the existing constraints governing institutional change, so it is often imperfect and will disappear or evolve when the constraints are lifted.

The dual-track price system (DPS) introduced in Section 7.2 is a prime example of a contingent institution. It was certainly not an optimal institution, but in addition to avoiding hyperinflation, it has also managed to evade the backlash of the groups with strong vested interests in the command economy. The way the big-bang approach adopted to attack this issue was fast privatization that was thought would eliminate the political bases of those interest groups (Boycko et al. 1997). In contrast, the dual-track approach provided limited protection to those interest groups, creating what Lau et al. (2000) has called a “reform without losers.”

According to Lau et al. (2000), the efficiency of the DPS lies in its two features: one is that the quotas were strictly enforced and market resale of quotas was allowed. This is quite different from the similar reform of the Soviet Union studied by Murphy et al. (1992) that was not able to enforce the quotas. Because the quota prices were lower than the market prices, firms had no incentive to produce for quotas so the dual-track system would collapse. In the Chinese case, quotas were strictly enforced. In this case, administrative discipline helped China’s DPS to succeed. On the other hand, allowing the resale of quotas eliminated the inefficiency stemming from the misallocation of quotas. However, quota resale was one of the early sources of corruption in China’s reform era. The DPS therefore provides an example of corruption through “greasing the wheels.”

The DPS disappeared in the early 1990s primarily because the market prices had converged with the quota prices. The market prices dropped because there were more and more firms supplying to the market. In particular, the township and village enterprises (TVEs) played a significant role. They were not covered by the government plan and had to rely on the market to obtain supplies and sell their products. Their growth greatly enhanced the market track. Yet they themselves were one of the contingent institutions. On legal terms, they were owned by the government, but in effect, they were jointly operated by individual entrepreneurs and the government. In fact, many of them were so-called “red hat” enterprises: they were established by entrepreneurs, but to avoid the uncertainty surrounding private firms, they were registered as township- or village-owned firms. Because of this legal ambiguity, property rights were not clearly delineated within the enterprises. By the standard theory of firm, therefore, TVEs could not have worked. Yet they flourished and contributed to 40% of the national industrial output growth at their highest point (Lin and Yao, 2001). In the 1990s, when private firms obtained a firm legal status, though, almost all the TVEs were privatized.

We can provide more examples of contingent institutions; the rural reform, SOE privatization, and the remuneration scheme for government officials all experienced stages of contingent institutions. Like the DPS, many of them created new forces demanding for less distorted institutions and as a result they disappeared in the end. One issue worth more exploration is the corruption created by contingent institutions. Yao (2004) shows that efficient institutional change is possible if side payments are allowed. Interpreting from this perspective, corruption is one kind of side payment that buys the support for reform. However, this does not tell us why economic growth has not been seriously undermined by rampant corruption. Figure 7.19 compares China with 88 other countries during the world during the period 2001–2009, in terms of corruption and economic growth. Two panels are shown in the figure. The left panel is a scatter diagram of the average growth rate of per-capita GDP against the mean corruption perception index (CPI) in the period. CPI is constructed by Transparency International by citizen surveys conducted each year. Larger values of CPI indicate cleaner governments. There is a weak negative correlation between CPIs and GDP growth rates. China is identified in the chart and is one of the two outliers with low CPIs but high growth rates.33 However, the negative correlation may be caused by the correlation of CPI with a country’s initial income. To deal with this issue, the mean per-capita GDP growth rate is first regressed on a constant and the per-capita GDP in 2001, and then the residuals are plotted against the average CPIs again. Now the negative correlation vanishes, but China is still one of the few outliers at the lower end of the CPI. China’s average CPI was 3.4, qualifying it as part of the most corrupt 25% of countries. But compared with both countries with cleaner governments, and countries with equally corrupt governments, China gained a much higher average growth rate. Why has corruption not become a serious impediment to China’s economic growth?

The RDA regime certainly has helped mitigate the negative impacts of corruption. As the preceding review in this section has shown, the promotion tournament has a strong dose of meritocracy that aligns local leaders’ interests with the pursuit of economic growth. In addition, competition among different localities has placed a check on the ability of local officials to grab from the business. This check is reinforced by China’s manufacturing-based growth model. The political-economy theory of the resource curse (e.g. Bulte and Damania, 2008) asserts that resource abundance fosters a predatory state that suffocates growth. This thesis is built on two premises: the state has a monopoly on the extraction rights of natural resources, and resources, i.e., the prey of corruption, are not mobile. In a manufacturing-based economy, however, the preys are capital owners who can easily move to other places. Officials in different jurisdictions compete with each other; they would be very happy to take the capital driven out by their more corrupt neighbors. Empirical evidence shows that local government officials pay more attention to manufacturing growth than the growth of agriculture and services (e.g. Yao and Zhang, 2011), primarily because manufacturing does not rely much on locality-specific inputs, whereas agriculture and services do. Depending on the composition of the local economy and its reliance on locally provided inputs, therefore, there could be an equilibrium in which government officials assume the role of both a helping hand and a grabbing hand.

A general lesson emerging from China’s reform process is that developing countries may have to give up the pursuit for institutional purity and instead focus on institutional efficacy when they conduct policy reforms. Good institutions align agents’ own interests with the societal interests. But agents take actions in a web of institutions, many of which can be detrimental to economic efficiency, yet cannot be easily overturned in a short period of time because they are deeply rooted in a country’s history. Therefore, the new institutions have to adapt to the existing institutions. As a result, they may not be pure; but with a wise design, they can be effective in raising economic efficiency and creating forces supporting further reform.