Environmental and trade policies are closely linked. Even if pollution is considered only a domestic matter, environmental policies can affect industrial competitiveness in countries with the toughest policies. This can lead governments to set environmental policies for strategic trade reasons. This has two aspects: the targets for environmental policy (how tough they should be) and the policy instruments chosen to achieve those targets. These add a new dimension to economists' usual discussion of the efficiency of different policy instruments. The possibility of strategic behaviour by governments implies a need to coordinate environmental policies. But a second and more traditional reason for coordinating environmental policies is when there is transboundary or global pollution. Here there is a need for international agreements on environmental policy even if there is no strategic manipulation of policy for trade reasons.

These interactions of environmental policies and trade policies were the subject of a CEPR workshop held in London on 11/12 November. The workshop was organized by Alistair Ulph (University of Southampton and CEPR) as part of CEPR's research programme on `Environmental Policy, International Agreements and International Trade', supported by a grant from the Commission of the European Communities under its Environment Programme and organized with GRETA, Venice and SNF, Oslo.

In the first paper, `Voluntary Agreements in Environmental Policy', Carlo Carraro (Università di Udine and CEPR) and Domenico Siniscalco (Università di Torino and Fondazione Mattei) discussed voluntary agreements (VAs), bilateral contracts between firms and the government in which firms commit themselves to meeting environmental standards in exchange for a subsidy. The authors used principal-agent theory to analyse such contracts, aiming to design the optimal VA. They suggested that VAs should be used when reductions in pollution can be achieved through technological innovations, when there are technological spillovers in environmental innovations and when reducing pollution in strategic industries that produce tradable goods. Alan Ingham (University of Southampton) observed that VAs are not voluntary since the level of subsidy is typically such that firms are sure to take them up. More importantly, he suggested that the optimal way to obtain correct revelation of the type of firm (`clean' or `dirty') might have distortionary effects, with growth implications for the economy. Michael Rauscher (Universität Kiel and CEPR) questioned the `offer' of a subsidy by the government. He observed that often governments threaten very restrictive policies to induce firms to reduce emissions voluntarily without having to pay subsidies.

In `A Core-Theoretic Solution for the Design of Cooperative Agreements on Transfrontier Pollution', Parkash Chander (Caltech and ISI, Delhi) and Henry Tulkens (CORE-UCL, Louvain-la-Neuve) discussed the design of a scheme for sharing abatement costs through payment transfers between countries, using a cooperative game framework based on the concept of the `core' of a game. Particular features of this transfer mechanism are that total damage and abatement costs are minimized, and that no coalition can obtain lower costs with a different transfer or level of emissions. Alan Ingham commented on the assumption of strictly convex abatement and damage costs which implies that each country faces a unique minimum point. If instead the damage cost curve is S-shaped, as many ecologists suggest, then total costs could have two minimum points, a `dirty' and a `clean' one, and countries could end up with a `dirty' economy.

In `Emission Taxes in International Asymmetric Oligopolies', written with David Ulph and Anastasios Xepapadeas, Yannis Katsoulacos (Athens University of Economics and Business and CEPR) aimed to explain why some countries pursue tough environmental policies while others are lax, relating the different outcomes to different market structures. He observed that when countries set environmental taxes cooperatively and firms are asymmetric, the aim is to maximize total producers' surplus, implicitly assuming transfers between countries. Comparing this with a situation where countries set taxes in a non-cooperative way, the author concluded that policies in the cooperative equilibrium countries are too tough. Furthermore, countries with tougher policies have lower output. When governments act non-cooperatively then, compared to a first best, they are too lax, and again tougher countries have lower output.

In `Plant Location and Strategic Environmental Policy with Inter-Sectoral Linkages', Alistair Ulph and Laura Valentini (University of Southampton) analysed the impact of environmental policies on plant location in oligopolistic models of international trade that allow for inter-sectoral linkages between industries, and permit analysis of incentives for agglomeration. The analysis was of a three-stage game, in which first, two countries set taxes; second, all the firms of two industries, upstream and downstream, choose how many plants to locate and where; and third, outputs of both industries are determined. In this model, demand for intermediate goods is endogenous, depending on where firms locate. This means that environmental taxes imposed on one industry can have consequences for the other. Surprisingly, this effect can sometimes be positive: for example, a country imposing a small tax can attract other firms to locate there. Compared to a Pareto tax, countries may be too lax or too tough, depending on the importance of the damage cost compared to the gains in rent. Aart de Zeeuw (Tilburg University) suggested allowing for countries of different sizes, in which, for example, large countries have higher profits but more environmental damage.

In `Environmental Policy as a Game Between Governments when Plant Locations are Endogenous', Michael Hoel (University of Oslo) also analysed the noncooperative tax behaviour of two countries, and its effects on plant location. Hoel assumed that there are no transport costs (so that firms locate in one country only) and no inter-sectoral linkages. In his model, the only factor distinguishing costs of production is the environmental tax. Hoel confirmed that emission taxes set by non-cooperative countries can be lower or higher than those in cooperative countries, depending on the importance of damage costs. On the one hand, each country may want to attract firms, using lower taxes or standards and hence implying too lax environmental policies. On the other hand, if the damage cost is sufficiently high, each country might prefer that firms locate in the other country, hence setting too strict environmental policies. David Ulph (University College London and CEPR) commented on the move structure of the game, observing that, in reality, governments set taxes prior to firms' location choices. In the model, governments do not have to commit themselves, while firms are subject to some sunk costs, indicating their commitment to particular location choices. Ulph argued that in equilibrium it would then matter if governments set taxes above or below the optimum.

In `Standards versus Taxes in a Dynamic Duopoly Model of Trade', Talitha Feenstra, Peter Kort, Piet Verheyen (TiIburg University) and Aart de Zeeuw discussed the choice of policy instrument between taxes and standards, showing that there are cases where standards are preferable to taxes. In their model, each duopolist is located in a different country and decides its level of investment at the beginning of the game and its level of output (or energy use) in each period. Strategic interactions occur between the firms, a result of which is that they engage in an investment race. Governments decide whether to set taxes or standards at the beginning of the planning horizon, aiming to reach a fixed emission target. The firms react differently when a competitor is subject to a tax rather than a standard, since imposition of the latter implies less flexibility of reaction. The authors concluded that the reduction in flexibility is advantageous because it reduces the firms' incentives to overinvest. Carlo Carraro observed that this result is partly due to the fact that governments include only the producers' surplus in their welfare function. If they also included the consumers' surplus, then they would want firms to produce more, preferring taxes to standards.

In `Protectionist Interests and Environmental Policies in Open Economies', Michael Rauscher modelled the problem of regulatory capture. An incumbent government wants to maximize its political support, and, in his model, there are two kinds of lobbies: sector-specific factors of production and environmentalists. In addition, there are two traded goods (which are imperfect substitutes), firms are monopolistically competitive, and governments decide environmental taxes, product standards and tariffs. Rauscher argued that governments can use environmental policies to protect their domestic industries. But, counter-intuitively, environmentalists may find environmental quality being negatively affected by tight product standards, while industry lobbies may gain from tight environmental standards, even with foreign competitors not subject to these standards. The impact of industrial and environmental lobbies depends on the particular policy instruments, and some effects of regulatory capture are clear while others are quite ambiguous. Jim Rollo (UK Foreign and Commonwealth Office) stressed the importance of checking if externalities are in production or in consumption. He also observed that it is important to distinguish politicians from bureaucrats as policy-makers. The latter normally prefer command and control policies to taxes since they are less uncertain and depend on fewer unknown parameters.

In `Economic Policy and the Manufacturing Base: Hysteresis in Location', Anthony Venables (London School of Economics and CEPR) analysed the incentives for agglomeration caused by the combination of imperfect competition and inter-sectoral linkages. As in the Ulph-Valentini model, firms are monopolistically competitive in all industries, but in contrast, countries do not act strategically, that is taxes are not determined in a non-cooperative game. Using a simulation based on chemical industry data that considers two locations and one industry, Venables suggested that the industry in the taxed location may collapse with all firms locating in the other country. Hysteresis in location may arise, which has implications for removal of the tax, highlighting the importance of a country's industrial base and suggesting that the effects of economic policies are not always as intended. Domenico Siniscalco noted how this model is a typical example of tax competition among countries. When more countries try to coordinate their environmental policies, the incentive to free-ride becomes large because of the risk of losing the industrial base.

In `R&D Cooperation and the Stability of International Environmental Agreements', Carlo Carraro and Domenico Siniscalco analysed the problem of stability of international cooperation to protect the environment. Starting from the traditional definition of cartel stability, they considered two kinds of negotiations: one on pollution abatement, which is typically unstable, and one on R&D, which is typically stable. The main idea was that linking these two kinds of negotiations should make environmental cooperation stable. In a three-stage game; first, countries decide whether to cooperate; second, their preferred levels of abatement and/or &nbsptechnological spillovers are decided; in the third stage, firms decide their outputs. The result is that when countries negotiate pollution reductions, the stable coalition is rather small whatever the number of countries involved: because of the non-exclusivity of environmental benefits, the incentive to free-ride makes large coalitions unstable. In contrast, with R&D cooperation, it is possible to exclude non-members of a coalition from the benefits of cooperation: the incentive to free-ride is small and the stable coalition is the one with all countries. When these two negotiations are combined, the benefits of the second kind of coalition offset the free-rider incentives of the first, making it stable. David Vines (Balliol College, Oxford, and CEPR) asked if the combination of unstable and stable coalitions is the right way to solve the instability problem: a central issue in linkages is normally about punishments rather than the combination of two matrices of payoffs. Yannis Katsoulacos objected to the idea of negotiations between governments over R&D spillovers, observing that normally this is a matter of cooperation between firms rather than governments.

In `Environmental Regulations and Manufacturing Employment: A Microeconomic Study of Norwegian Data', Rolf Golombek and Arvid Raknerud (Centre for Research in Economics and Business Administration, Oslo) presented an empirical investigation of the impact of environmental regulations on manufacturing employment in Norway. Since the impact of regulations could differ across sectors and may also depend on firms' size, they analysed three separate sectors and firms of roughly equal size. Comparing patterns of unemployment and exit rates for regulated and unregulated firms in the same sector, they found that for two of the three sectors, regulation had a significant impact. Surprisingly, the regulated firm had a greater tendency to increase employment and a lower exit frequency than the non-regulated firms. In the third sector, the impact of regulation was not significant. &nbspAs a next step, Rosemary Clark (University of Birmingham) suggested checking whether the profitability of `green' goods holds for the whole sector and not just for these firms. L Alan Winters (World Bank and CEPR) suggested that a possible explanation of this result is that the government is providing a `hidden subsidy' to the regulated firms by, for example, favouring the purchase of their products over those of the unregulated firms. Michael Rauscher stressed the importance of the result since it challenges the traditional view of conflicts between the environment and employment.

The workshop concluded with a paper by L Alan Winters and Zhen Kun Wang (World Bank), `Carbon Taxes and Industrial Location: Evidence from the Multinational Literature'. Other papers had suggested that many countries are reluctant to impose a carbon tax because of the fear of losing competitiveness. OECD countries, for example, are worried that a partial agreement on pollution abatement has the effect of reducing their output but increasing the output of other countries, offsetting the benefits of their partial reduction in emissions. The underlying idea is that firms relocate to &nbspcountries where they do not pay a carbon tax, an idea that was questioned in this paper. In a survey of the empirical literature on multinationals, these authors showed that environmental factors are one element that multinationals take into account when deciding where to locate, but not the only one and not the most important. The effects of imposition of a carbon tax cannot be analysed without reference to other variables, such as differentials in costs of production. Considering the differences in environmental legislation, Winters and Wang argued, there is very little evidence of a `danger' of capital flights from developed to developing countries. In general, a carbon tax does not generate large increases in relative production costs and environmental policies do not necessarily affect negatively the production of domestic firms.