Strategic Environmental Policy and Strategic Use of R&D under Endogenous Regulation

Transcription

1 Strategic Environmental Policy and Strategic Use of R&D under Endogenous Regulation Preliminary draft Abstract We construct a model of strategic environmental policy with endogenous regulation in an international duopoly context. Firms are enhanced with the option to affect future environmental regulation through a prior investment in R&D. We highlight the role of investment in our results. Using commonly used functional forms we demonstrate that when governments use environmental policy strategically, welfare declines while pollution increases. We further illustrate that when emission standards are used as a policy instrument, instead of emission taxes, pollution, profits and welfare increase. Last but not least, we claim that flexibility in environmental policy is superior in terms of profits and welfare to a prior commitment to a certain standard. Keywords: Endogenous regulation, Strategic environmental policy, International trade, Imperfect competition, Commitment versus flexibility JEL classification: F1; F17; F18; Q53; Q56 Fabio Antoniou Department of International and European Economic Studies, Athens University of Economics and Business, Patission Str 76, Derigny Wing, 5th floor, Athens 10434, Greece. fantoniou@aueb.gr Panos Hatzipanayotou Department of International and European Economic Studies, Athens University of Economics and Business, Patission Str 76, Antoniadou Wing, 5th floor, Athens 10434, Greece. hatzip@aueb.g

2 1. Introduction In the last fifteen years, regional trade agreements (e.g. EU single market, WTO, NAFTA) led to a significant reduction in the use of tariffs and other direct trade policy instruments. These restrictions gave spunk for the investigation of alternative economic tools that could be used for trade purposes as well. Environmental policies fall in this category since in many industries affect costs significantly. The theoretical literature has contributed to these concerns by providing the fundamentals for the so called strategic environmental policy 1. Conrad (1993), Barrett (1994), Kennedy (1994), Rauscher (1994), Ulph (1996) and Neary (006), among others argue that when firms compete in quantities in international markets, ecodumping will appear during the setting of environmental policy. Put this concretely, governments set their policy instrument such that environmental policy is laxer in comparison with the first-best case. Recent empirical findings by Fredriksson and Millimet (001) and Ederington and Minier (003) argue that indeed there is strategic interaction in environmental policy setting and that environmental policy can be used as a secondary trade barrier 3. A common feature of these models is the timing they follow. The governments set their policy instrument at a certain level and then firms take their strategic decisions sequentially, about the level of R&D investment (where possible) and output. We characterize such a setting as exogenous regulation, since environmental policy cannot be affected by the decisions of economic agents. In our model we aim to approach the subject from a different perspective and we address the question what should happen and how the results are modified in the case where environmental policy can be subject to manipulation by economic agents decisions (i.e. R&D investment). Such a case, which is characterized throughout the paper as endogenous regulation, can occur when the governments are unable or unwilling to commit to an 1 Strategic environmental policy is based on similar mechanisms as strategic trade policy (see Brander and Spencer, 1985 and Eaton and Grossman, 1986), with the difference that an additional market imperfection (environmental externality) is added and that instead of production subsidies, environmental policy instruments are used as policy tools. Further extensions and limitations of these models after relaxing some assumptions have been done by numerous authors as Bradford and Simpson (1996), Nannerup (1998) Duval and Hamilton (00), Bayindir-Upmann (003), Greaker (003), Burguet and Sempere (003), Hamilton and Requate (004), Straume (006) and Barcena-Ruiz (006). 3 These results contradict the common belief held during the ninenties that there is no any significant correlation between environmental policy setting and competitiveness of the firms (see Jaffe et al, 1995).

3 F. ANTONIOU, P. HATZIPANAYOTOU initially announced policy level. When this is the case, firms R&D towards greener technologies should not only account for future output competition (e.g. Ulph, 1996) but should account for future environmental policies as well. In line with this is Dr Katsouyanni s (supervisor of an international group of scientists participating in the European program APHEA) claim: In fact the European Committee seems to have given way to the automobile industry pressures for laxer standards. Despite the fact that according to the scientists any reduction of Particulate Matters, which are mainly caused through the emission of gases by power plants, industries and automobiles, could prevent many deaths, European Committee decided to set too lax standards 4. Our model is a two period model involving two countries and one firm associated in each country where in each period governments and firms act sequentially. Firms invest in emissions reducing R&D in period 1 such that outcomes of period are affected. The main results suggest that if governments use environmental standards for strategic purposes as well, pollution and output will grow up, while innovation and welfare will be suppressed. The results do not change when taxes are used instead of standards. However, it is shown that when the latter are used as a policy instrument, welfare, profits and pollution are greater. Finally, we compare the cases where governments can choose whether to commit (exogenous regulation) or not (endogenous regulation) to a certain policy level. We obtain a striking result which suggests that we get a Pareto superior outcome in terms of welfare when both governments do not commit. The closest paper to ours is Puller s (006). In his paper are explored the welfare consequences of strategic behavior by firms in order to affect future regulation when the regulator have no means to commit to a specific standard. In contrast with our model (where an international duopoly is assumed), Puller assumes an oligopoly structure in a closed market and hence a single regulator 5. Under this 4 The results of APHEA will be presented in Epidemiology on July 007. According to these results, Particulate Matters (PM) concentration is responsible for numerous early deaths in Europe. Each year are estimated, approximately, early deaths due to PM..5. Actually within E.U. there is not any existing standard about PM..5. A standard exists only about PM 10 and it is commonly agreed that the new directives will lead to even laxer environmental policy (for details of what E.C. decided about PM standards see directive IP/06/163). Similar conclusions were held for the U.S. by NMMAPS (for details see 5 A similar staging in an open economy is used by Grossman and Maggi (1998) but in a strategic trade context without pollution externalities as in our model. They allow domestic firm s investment decisions to affect future trade policies.

4 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 3 assumption he illustrates through the use of simulations that welfare losses from the lack of commitment ability are mitigated and cleaner technologies are induced. In Section the model is presented and solutions are derived when emission standards are used as a policy instrument for the cases where governments act nonstrategically and strategically. In section 3 we calculate the solution for the case where taxes are used as a policy instrument and the results are compared with those in section. Next in section 4 we contrast the cases where the regulator commits to an initial policy level or prefers to be flexible. Section 5 provides some implications and concluding remarks. Proofs are given in the Appendix.. The model with emission standards under endogenous regulation.1 Stage game and notation We assume a symmetric two period game under complete information, where the governments and the firms act sequentially in each period. Our model consists of two firms, each one belonging to a different country, exporting their production to a third country where the consumers preferences can be mapped into a quasi-linear utility function which implies a linear inverse demand p = B-x-X, where B is the demand intercept and x, X are outputs for the domestic and the foreign firm respectively 6. Associated with the production of the domestic firm is a local pollutant. For simplicity we assume that one unit of production generates one unit of emissions and consequently emissions per unit of production equal unity. Following Ulph and Ulph (007), R&D in greener technology that can reduce emissions per unit of output is assumed to exist. Net emissions per unit of production are e = 1-β, where β is the amount of emissions reduction per unit of output by the domestic firm as a result of its own R&D and equals β = v/x, where v is investment in R&D and has convex cost, ic = gv where g is just a coefficient which will be assumed g 0.5. We observe that higher values of g imply a costlier investment and that marginal investment cost are increasing since ic = gv > 0. An abatement technology for this pollutant is assumed to exist. Hence total pollution is z = ex-α or equivalently z = x-v-α, where z is total pollution and α is abatement carried out with cost αc = 0.5α. Pollution is assumed to cause a damage to 6 Henceforth the corresponding variables and functions of the foreign firm and country will be denoted by upper case letters. We will introduce only the domestic functions since the foreign ones will be similar due to the fact that the model is symmetric.

5 F. ANTONIOU, P. HATZIPANAYOTOU 4 the local society given by d = 0.5kz, where k is a parameter determining the injuriousness of the pollutant. The convexity of the damage function describes a situation where every additional unit of pollution is more harmful than the previous one 7. Initially we assume that the government may use only an emission standard which is an upper limit on emissions and dictates the firm to pollute up to this level. Any extra emissions must be abated by the firm. Hence, emissions generated by the firm, z, must coincide with the standard set by the government 8. The usage of this standard is twofold. It is used to internalize the externalities caused by the production procedure, but at the same time can be used as an implicit subsidy for the firm, when it is set at a high level. We observe that there are two opposing forces affecting policymakers decisions. The externality internalization, which ratchets down (stricter) standards, and cost reduction, which ratchets up (laxer) standards. Initially, in period 1, governments announce a future performance-based standard applied to their corresponding firms. Governments announce emission standards that can be emitted but are unable or unwilling to commit. This could be due to the fact that at that time of regulation it could be too costly for the firms to implement it. For instance in 1990, it was adopted in California a policy that restricted vehicles to zero emissions but this standard was revised since automobile industries could not comply. Governments often announce infeasible policy levels that firms cannot follow and thus in the future they roll back their policies 9. Within the same period, following the governments announcements, firms invest in R&D, v, so they can comply with future standards when such standards are implemented. Innovation reduces net emission per unit and this implies that lower abatement should be carried out and thus marginal cost of abatement is lowered 10. In period of the game, it is the government s turn to update the announced level of regulation and implement a new emission standard, z. We assume that each 7 All the functions used will be the same as in Ulph s (1996) which are commonly used in the literature. We have chosen to work with these specific functional forms because it was impossible to derive any analytical solutions. 8 Note that when the regulator sets an emission standard, marginal cost of abatement should be positive and thus this standard must be binding for the firm. 9 Further examples can be found in Puller (006). 10 Johnson Matthey (00), one of the greatest producers of abatement catalysts, in its website announces a new abatement catalyst which reduces emissions of NO up to 90% and encourages firms to buy now this catalyst in order to comply in advance with future environmental policies and thus gain a competitive advantage in international competition.

6 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 5 government maximizes its domestic welfare, which is a linear combination of firm s profits and damage caused from pollution w d, (1) where ( B x X ) x cx ac ic, () denote domestic firm s profits. The model further assumes the firms have constant marginal cost in production which equals c for both firms. After governments regulation takes place, it is the firms turn to produce. Each firm determines its optimal level of production through maximization of profits given in equation (). At this point it is important to highlight the driving forces of the model. First of all we should note that when the firms maximize profits in period have as a choice variable only their output. Abatement cannot be chosen by the firms since the emission standard chosen by the regulator is binding. Abatement is calculated residually by α = x-v-z. (3) We observe that an investment in innovation in period 1 and a higher standard by the regulator have the same effect. Both reduce the level of abatement carried out in period. Therefore marginal cost of abatement is reduced. It is important to note that if the cost of abatement is not convex the analysis does not hold because investment and regulation do not affect marginal cost of abatement. It is this reduction of marginal cost of abatement that affects the behavior of the firms in period. The necessity of the assumption g 0.5 can be seen here. Supposing that this is not the case the firm has an obvious incentive to invest in period 1 in green technology so that abatement in period tends to zero because it is more expensive. Note also that g = 0.5 refers to a particular case where investment cost equals cost of abatement. As in Neary (006) abatement can be separated from production decisions and can be chosen either in period 1, acting as a pre-commitment effect or in period when production takes place.

7 F. ANTONIOU, P. HATZIPANAYOTOU 6 The effect that regulation and innovation have on the firm s behavior can be viewed in Figure 1 for the domestic firm. Residual demand is denoted by the bolded line. Total marginal costs are given by the summation of marginal cost of production and marginal cost of abatement, indicated by the kinked line cab. The kink at point A appears because from this point and forth the standard is binding. Firm s optimal production point is determined as usual at the point where marginal revenue crosses total marginal cost. When the regulator sets a higher standard in period marginal cost of abatement is shifted to the right and the new total marginal cost function is given by ca B. The optimal point is shifted to the right and the firm chooses a greater level of output. The losses from such a policy stems from the fact that price is lowered and are denoted by the blue shaded area. Benefits accrue from the fact that marginal cost of abatement is lower and are represented by the green shaded area. Clearly benefits exceed losses and thus firms increase their production. Similarly is affected domestic output by a greater investment by the own firm in period 1. In case foreign firm s output is reduced, residual demand shifts outwards and consequently optimal level of production increase. p Losses TMC B B TMC c A A Benefits MR x P(x,X) Figure 1 x Following Ulph (1996) two cases are examined. First, we examine the case where the governments are assumed to set the standards non-strategically and then we study the case where they do set them strategically.

8 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 7. Governments Act Non-Strategically Here we assume that governments act non-strategically when they choose environmental standards 11. That is they only use the environmental standard to internalize the environmental externality and they do not take into account the effect of their policy on the output of the rival firm. The problem is solved by backwards induction. The first order condition for the domestic firm in period is taken through maximizing profits in () with respect to output B c v 3x X z 0, (4) Note also that second order condition is satisfied since the profit function is concave (linear demand, constant marginal cost and convex abatement cost). Solving (4) and its analogue for the rival producer with respect to outputs we get the equilibrium output levels x=1/8(b-c+3v-v+3z-z), (5) where x / z 3/ 8, x / Z 1/ 8, x / v 3/ 8 and x / V 1/ 8 which imply that the own partial effects are stronger than the cross partial ones. Moreover, we observe that a higher domestic standard and greater investment yield higher output, while the opposite results are obtained when the corresponding foreign variables increase. This result emerges due to the fact that greater z and v act as a pre-commitment effect for more aggressive behavior of the domestic producer (see Fig.1) and reduces foreign output through the reaction function. In the beginning of period the government determines its optimal standard through welfare maximization given in (1). So the first order condition is dw x X d 0 dz z x z X z z (6a) x v z. 1 k (6b) 11 The fact that governments act strategically or not is known by the firms when they invest in R&D in period 1.

9 F. ANTONIOU, P. HATZIPANAYOTOU 8 The second term in (6a) disappears due to the first order condition for profit maximization. Similarly, the third term is set equal to zero due to the assumption of non-strategic behavior of the governments. The optimality condition in (6b) captures the equality between the marginal cost of abatement and the marginal damage in each country. Emission standards are set at the optimal Pigouvian level. The difference with other relevant papers (e.g. Ulph, 1996) is that since environmental policy is subject to manipulation by the firms investment decisions the optimal Pigouvian level is also endogenized. Solving a system of equations given in (5) and (6b) and their analogues for the foreign firm and government we can obtain the optimal standards as a function of domestic and foreign investment, from which we can derive the following partial effects z 3 5k v 3 10k 8k 0 z k and 0. (7) V 3 10k 8k Equivalent results can be obtained for the foreign firm. Interpreting these partial derivatives, we notice that emission standards and investment are strategic substitutes (see Bulow et al, 1985), implying that an increase in innovation in both countries leads the regulator to ratchet down the standards. The fact that standards and investment are strategic substitutes is anticipated by the firms when they change the optimal level of investment in period 1. Therefore, producers have an incentive to under-invest rather than to facilitate the regulator to ratchet down the standard. Naturally, there exist opposing forces that thrust for greater investment. These incentives rely on the profit shifting behavior by firms. As previously discussed, greater level of v increases domestic output and detects foreign output and standard. If the two latter effects dominate the first one then the firm will select a positive level of investment. Otherwise, no investment occurs and abatement coincides with production decisions in the final period of the game. These effects are represented through firms profit maximizing condition for period 1 with respect to v 1 1 In period 1 investment competition a conjectural variation that is similar to the Cournot competition model, i.e. v / V 0. For an explanation regarding conjectural variations and comparative statics in oligopoly models see Dixit (1986).

10 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 9 d z X z X X Z Z x x z 0 (8a) dv v z v X z v X v X Z v Z v x v x z v k((6 9 k) x (5 8 k) z) v. (8b) 6 g k(5 8k 4(5 4 k) g) The first term of equation (8a) denote R&D investment cost. The next two terms are negative and represent the ratchet effect. The fourth term is positive and stands for the reduction in foreign output due to the pre-commitment effect of investment by the domestic firm while the fifth and sixth terms are also positive and reflect the raise the rival cost effect. Finally the last two terms disappear due to the first order condition in period. Solving simultaneously the system of equations given by (3), (5), (6b) and (8b) we derive the solutions for our variables x X m ( B c)( k(1 k) 6g 4 k(5 4 k) g) z Z m ( B c) k(1 k(7 9 k)) v V m ( B c) k( k(1 k) 6g 4 k(5 4 k) g)] m ( B c)( k (5 8 k) (1 k)(1 k)(3 4 k) g), (9) where m = k (14+3k)+(1+k)(3+4k) g. Since B is greater than c and the parameters k and g follow the initial assumptions it follows that an interior solution exists. Differentiating the solutions with respect to g (i.e. as investment cost in R&D increases) we get that output and investment decrease, while standards and abatement increase. Similarly, if we differentiate with respect to k (more harmful pollutant), output and standards detract while abatement and investment rise 13. It is worth noting here that the level of investment chosen by the individual firms is not the socially desirable. To see that, assume that v was under the social planner s choice. Then the government would choose the optimal v to maximize welfare as follows 13 Partial derivatives, as all the derivations presented in this paper, are calculated through Mathematica.

11 F. ANTONIOU, P. HATZIPANAYOTOU 10 dw d D z 0 dv dv z v. (10) Yet, this equation when evaluated at the optimum (from the firm s perspective) is positive (since the first term is zero and second term is negative). Bearing in mind that the welfare function is concave, we conclude that the investment in R&D should be greater when it is subject to the social planner s choice than when it is not (v * g >v * ). Put it differently, when the firm chooses v there is an underinvestment problem because it only cares about the maximization of its profit. From (7) it follows that in case that v and V were subject to the regulators option, environmental policy would be stricter..3 Governments Act Strategically In contrast with the previous case, now when each regulator chooses the optimal environmental standard, takes into consideration how this will affect the choices of the rival firm. The procedure is similar to the preceding case. Output in the last period is unchanged and solutions coincide with those given in equations in (5). In period things are slightly different. First order condition for the government is given by equation (6a) with the only difference that now X / z 0. 9x 8v z. (11) 8(1 k) In contrast with the former case, we observe that output has a higher share in the best response functions 14. These relations do not reflect anymore the fact that the marginal cost of abatement equals marginal damage. If we substitute (5) into (11) and use the analogue functions for the foreign firm and government we get the new partial effects, z k 0 v (7 16 k)(3 3 k) and z 7k 0. (1) V (7 16 k)(3 3 k) 14 It is easy to check through these equations and find that z / Z 0. Hence emission standards are strategic substitutes.

12 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 11 The second effect in (1) is stronger than the one in (7) where governments did not act strategically, while the first is weaker. Put it differently the first partial derivative shows that emission standards and investment are still strategic substitutes but now the effect is a bit moderated in contrast with the second partial derivative. The explanation for this relies on the fact that when governments act strategically put more weight in output than in pollution. Moving towards period 1 we use again the profit maximizing condition in (8a) which yields 4 k(9(5 8 k) x (37 64 k) z) v. (13) 4 k(37 64 k) (7 16 k)(3 3 k) g Solving simultaneously the system of equations given by (3), (5), (11) and (13) and the analogue equations for the foreign firm and country we compute the optimal solution for the home and foreign variables xs X s n 9( B c)( 1 k g(7 16 k)(3 3 k)) zs Zs n 36( B c) k(3 8 k(5 8 k) vs Vs n s s n 8( B c)( g(1 k)(7 16 k)(3 3 k) 4 k (37 64 k)) ( B c)( 56 k (1 k) g( 1 8 k)(7 16 k)(3 3 k)), (14) where n = g(7+16k)(3+3k) +3k ( k). In order to obtain an interior solution it is sufficient that k>0.13. The signs of partial derivatives with respect to g and k remain unchanged from the former case. To get a better intuition we need to compare the optimal solutions in the two cases. The results are summarized by the following proposition Proposition 1: When governments act strategically, output and emission standards are increased more than in the case where governments do not act strategically, while investment in R&D, abatement and social welfare are greater in the latter case. (See proof in the Appendix)

13 F. ANTONIOU, P. HATZIPANAYOTOU 1 Proposition 1 does not provide any prediction about profits since the sign of the difference of profits in the two cases depends crucially on the value of k 15. In contrast with Ulph (1996) our model suggests that when governments set regulation strategically the incentive of the firms to invest and hence act strategically is dampened. Our results concerning welfare are aligned with Ulph s; in the case where both parties act strategically welfare declines. 3. Emission Taxes in the Non-Strategic Case We will now repeat the procedure followed in Section. but the regulators now instead of emission standards use emission taxes When an emission tax instead of a standard is used, domestic welfare and profit functions are w tz d, (15) 1 1 ( B x X ) x cx a v tz, (16) where t is emission tax per unit of emissions. In period, in contrast with the former case, firms have two choice variables, abatement (or pollution) and production since they are not restricted to a specific standard. Solving simultaneously the first order conditions of profit maximization for both firms we obtain optimal levels of output and abatement x=1/3(b-c-t+t) and a=t. (17) From the first order conditions in (17) we observe that firms abate pollution to the point where the marginal cost of abatement equals tax. Moreover, the equilibrium outputs do not depend directly on v and V. Hence, firms do not have a direct incentive to invest in order to shift profits on their own. In order to understand the differences between the two cases we introduce Figure which is similar to the previous one. Now, we introduce in the same figure 15 Numerous simulations that we have done illustrate that for values of k greater than 0.3 profits in the strategic case are greater than in the non-strategic case. For k<0.3 the opposite happens. This result suggests that only industries that cause great damage to the environment would lobby to achieve laxer environmental policy.

14 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 13 both total marginal costs when standards are used (cab) and total marginal costs with taxes (TT ). In the case of taxes, total marginal costs equal marginal cost of production plus the tax which equals marginal cost of abatement. We construct our diagram such that initially under both cases is produced x. However, in case the foreign firm reduces production, residual demand shifts to the right and the optimal production is given by x s when standards are used and x t when taxes are levied. We directly observe that output increases more if a tax is implemented which yields a reaction function with a greater slope 16. This property is of great importance and drives the results that will follow. p B TMC s c+t c T T TMC t A MR x MR x P(x,X) P (x,x) x x s x t x Figure The channel through which investment affects output competition is through taxes. Using a similar condition as in (6a) for the domestic government 17 kx kv t. (18) 1 k Solving for the partial effect of t with respect to v we get t (3 5 k) k v 3 10k 8k 0 and t k 0 V 3 10k 8k. (19) 16 For a formal proof in the general case see Neary (006). 17 In contrast with the case where standards are used we find that t / T 0. Hence taxes are strategic complements.

15 F. ANTONIOU, P. HATZIPANAYOTOU 14 The signs of these partial effects imply that firms can bring environmental policy closer to their interests, in contrast with the emissions standards case where greater R&D investment led to stricter environmental policy. In particular, an increase in v relaxes domestic and foreign government s tax. However, the first effect is stronger than the second, and hence a greater investment from the domestic firm leads it toward a more aggressive behavior in the output competition stage. Therefore firms have an indirect incentive to invest in green technologies prior to production. Using a similar condition as in (8a) for profit maximization with respect to v when taxes are used we obtain (3 k(7 3 k)) t k( 3 k) x v. (0) 6 g k(3 5k 4(5 4 k) g) Solving simultaneously (17), (18) and (0) we derive the equilibrium for our variables ( B c)(6 g k(6 6 g k(15 8k 4(9 4 k) g))) xt X t q ( B c) k( k(1 k) 6g 4 k(5 4 k) g) t T a A q ( B c) k(7 k(17 9 k)) vt Vt q ( B c)( k(1 k) 6g 4 k(5 4 k) g) zt Z t q, (1) where q = k(18+k(44+3k))+(1+k)(3+4k) g. The main results from the comparison between the cases of taxes and standards are summarized by Proposition Proposition : Assuming that governments do not use regulation policy strategically, when they choose emission taxes as an instrument instead of emission standards, outputs and R&D investment increase while pollution, abatement, welfare and profits are suppressed 18. (See proof in the Appendix) 18 This proposition can be also extended to the case where governments choose regulation strategically with similar conclusions.

16 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 15 Proposition is in line with Ulph (1996) and illustrates that, as in the exogenous regulation setting, firms are more aggressive when emission taxes are used as a policy instrument. Moreover, we provide a justification of why standards should be used instead of taxes. However, if the aim of the regulator is to prevent environmental detection, then emission taxes should be preferred. Using a similar equation as in (10) we get dw d D z t 0 dv dv z t v. () When we evaluate this at the optimal level of investment the whole term is negative. This implies that from a social perspective there is an overinvestment in R&D. When taxes are implemented, firms invest more in abatement capital and thus commit to higher levels of abatement than in the case where standards are used. Thus at the output stage firms become more aggressive and outputs rise Commitment versus flexibility So far the model assumed that governments lacked commitment ability. We will assume that governments can commit and apply their policy before firms take any actions. This transforms the model to an exogenous regulation model commonly used in the literature (see Neary, 006). Now, environmental regulation is implemented before R&D investment and as a consequence it cannot be affected by firms economic decisions. We will assume for simplicity that governments do not use environmental policy strategically. The results hold even when this is not the case. The timing is as follows. Governments set their standards and then firms choose sequentially investment in R&D and output. Equation (5) for output and (6b) for standards still hold. However, optimal condition for investment is altered d X x 0 dv v X v x v (3a) 9x 8z v 8 16g (3b) 19 Hence, possibly a mixed system of effluent standards and taxes could improve welfare.

17 F. ANTONIOU, P. HATZIPANAYOTOU 16 Equation (3a) in comparison with (8a) misses the ratchet and the raise the rival cost effect. Solving (5), (6b), (3b) and their analogues for the foreign firm and country simultaneously we obtain the solutions for our variables 8( B c)( k (1 k) g) xc X c i ( B c)( 116 g) zc Zc i, (4) ( B c)(1 9 k) vc Vc i ( B c) k( 116 g) c c i where i = 48g+k(3+64g). If we compare the results given in (4) with the one given in (9) we get the following Proposition Proposition 3: When governments commit to an emission standard, innovation in green technologies and output are enhanced, while pollution, abatement, welfare and profits are reduced. (See proof in the Appendix) The important implication of this proposition is that if governments commit to an international agreement they can deter pollution at the expense of lower welfare and profits. This can potentially provide a justification for the unwillingness that some countries show when they are invited to sign international environmental agreements. If governments prefer to be flexible in their policy level, a Pareto superior outcome in terms of welfare and profits emerges. Investment in R&D plays a key role in Proposition 3. Since the ratchet and the raise the rival cost effects are absent the firm uses investment only as a precommitment mechanism. Therefore R&D investment is higher in this case and this drives the results. Governments set stricter standards but the overall effect in output is positive. As a result competition is sharpened and profits are lowered. The benefits from the cleaner environment are not sufficient to compensate the profit losses and the

18 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 17 total outcome is negative 0. One way to see the key role of R&D investment is to differentiate the equations (A18), (A19), (A0) and (A1) with respect to g, which determines the cost of investment. These equations give the differences on the optimal values of the variables between the cases where the governments commit to a specific standard and the one where they do not. It can be checked out that as the cost of investment raises these differences decline. 5. Concluding remarks - Implications This paper bridges together the notions of strategic environmental policy and the possibility that the firms have to manipulate environmental policy through the use of investment in R&D. Our model suggests that when governments use environmental policy not only to restore efficiency caused from pollution but for rent-shifting purposes as well, a race to the bottom emerges which lowers welfare in both countries with a parallel degradation of environmental quality. Additionally, our model extends previous models results and provides a justification for the use of emission standards instead of emission taxes. When standards in both countries are used we obtain a Pareto superior outcome in terms of profits and welfare. Nevertheless, this superiority feature is achieved at the cost of lower environmental quality. The driving force for these results relies upon the fact that firms have a more aggressive behavior when taxes are used. In the last part of the paper we provided a rational for the countries to not sign any committing agreements restricting their environmental policy. They should instead prefer flexibility since they can be better off in terms of welfare even if environmental quality is reduced. This by no means should be perceived as a suggestion to the governments to not care about the environment. Our results throughout the paper indicated that the more harmful the pollutant is the stricter environmental policy must be, independently of the instrument used or the existence of commitment or not. The suggestion to the governments to be flexible in their policy decisions relies on the fact that firms are less aggressive in their investment decisions. When regulation is flexible governments credibly threat 0 The Pareto superior outcome obtained in the flexibility case is not Nash equilibrium. In other words, when the governments decide weather to commit or not to a certain standard, it is not necessary that governments have a unilateral incentive to not commit. In order to find the Nash equilibrium we should also take into account the asymmetric case where one government commits and the other does not. In such a case it can be shown that two Nash equilibria in mixed strategies exist involving asymmetric responses by the governments.

19 F. ANTONIOU, P. HATZIPANAYOTOU 18 firms that they will have to comply with stricter future standards if further investment takes place. The limitations of this model can provide stimulation for further extensions. The presence of more than one firm in each country, the existence of potential consumers of the produced good in each exporting country and the occurrence of a trans-boundary pollutant instead of a local one demand for a more complicated analysis. Governments should take into account of these factors when they select their policy levels. One of the basic suggestions of this paper is that exogenous regulation should not be taken for granted. Yet, if regulation is endogenous, many implications should be revisited, as for instance the way that asymmetric information is modeled. Nannerup (1998) used a screening model to resolve asymmetric information in domestic firm s cost. This is not possible under an endogenous regulation setting where the firms choose investment in R&D before policy is determined. In this case a signaling model, where the domestic firm signals its true cost through a distortion in investment decisions is the appropriate way to model asymmetric information. When this happens greener technologies are induced by the more efficient firm type and pollution declines since regulation and innovation are strategic substitutes! Appendix Proof of Proposition 1 Since the model is symmetric it is sufficient to illustrate the signs of the differences only for one country. The difference and the sign of optimal outputs, emission standards, the level of investment and abatement between the two cases (strategic and non-strategic) after some algebraic rearrangements are x s x 4 ( B c)(96 k (1 k) g (1 k)(1 k)(3 4 k)(7 16 k)(3 3 k) gk ( k(74 k( k)))) 0, (A1) mn 3 ( B c)(8 g (1 k)(1 k)(3 4 k)(7 16 k)(3 3 k) 4 k (446 k( k)) gk(1759 k(13909 k( k( k)))) zs z 0, (A) mn ( Bc) k(4 k (446 k( k)) g(1759 k(9697 k( k( k)))) vs v 0, (A3) mn

20 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION ( Bc)(96 k (1 k) g (1 k)(1 k)(3 4 k)(7 16 k)(3 3 k) gk ( k(74 k( k)))) s 0, (A4) mn Optimal welfare for the non-strategic case is given after substituting the optimal values obtained in (10) w * n 3 ( B c) (4 g (1 k)(1 k) ( 3 k)(3 4 k) k ( 1 k(37 k( k))) gk (59 k(37 k(857 k( k)))) 0. (A5) m Optimal welfare for the strategic case is given after replacing the optimal values obtained in (15) 3 3( B c) ( g (1 k)(7 16 k) (3 3 k) (37 64 k) 16 k ( k(347 8 k( k))) * 8 gk ( k( k( k( k)))))) ws 0. (A6) n The difference of optimal welfare in the strategic and non-strategic case is given by (A6)-A(5) and after some tedious algebraic rearrangements we get w s wn ( Bc) ( g g k g k g k g k g k g k g k g k g k gk g k g k g k gk g k g k g k k gk g k g k g k k gk g k g k g k k gk g k g k g k k gk g k g k g k k gk g k g k g k ) m n ) 0 (A7) After a careful examination of (A7) it can be easily seen that the overall sign will be negative Q.E.D. Proof of Proposition : First define as δ=(3+5k)(-k(1+k)+6g+4k(5+4k)g). The next four equations present the differences and the signs of output, investment, pollution and abatement between the two cases (taxes and standards)

21 F. ANTONIOU, P. HATZIPANAYOTOU 0 t x x ( B c) k / mq 0 (A8) vt v ( B c) k(3 4 k) / mq 0 (A9) t z z 6( B c) k / mq 0 (A10) 6( B c) k / mq 0 (A11) Welfare in the case of taxes after substituting equilibrium values of the variables is ( B c) ( k (7 k(347 k(597 k( k)))) * k(7 k(479 k(180 k(1717 k( k)))) g 4(1 k)(1 k) ( 3 k)(3 4 k) g wt q )). (A1) The difference of (A1)-(A5) is 3 ( Bc) k (4 g (1 k)(1 k)( 3 k)(3 4 k) gk(3 4 k)(9 k(14 k(397 k( k)))) k (7 k(37 k(664 k( k)))))) wt wn 0. (A17) m q We have shown that w or equivalently * d * * d *. This implies t wn t t n n that d d. Damage is a continuous and increasing function of standards. If t n t n we combine this with (A10) we observe that the right hand side is negative and thus 0 Q.E.D. t n Proof of Proposition 3: Define θ=6g+58gk+(-3+84g)k. The next four equations present the differences and the signs of output, investment, pollution and abatement between the two cases (commitment and flexibility) c x x ( B c) k / mi 0 (A18) vc v ( B c)(3 4 k) / mi 0 (A19) c z z 3( B c) / mi 0 (A0)

22 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 1 3( B c) k / mi 0 (A1) Welfare when governments commit to a standard after substituting equilibrium values of the variables is w * c ( B c) ( k (1 k) g)( 1 111k 18( 3 k) g) (A) i The difference of (A)-A(5) is 3 ( Bc) (3 k (37 k( k)) 4 g (3 4 k) (3 k(13 6 k(7 8 k))) gk(7 k(79 k(167 4 k(94 55 k)))))) wc wn 0 (A3) 4m i The proof for 0 is similar as in Proposition Q.E.D. c n References [1] Barcena-Ruiz, J. C. (006), Environmental Taxes and First-Mover Advantages, Environmental and Resource Economics 35, [] Barrett, S. (1994), Strategic Environmental Policy and International Trade, Journal of Public Economics 54, [3] Bayindir-Upmann, T. (003), Strategic Environmental Policy under Free Entry of Firms, Review of International Economics 11(), [4] Brander, J.A. and B. J. Spencer (1985), Export Subsidies and International Market Share Rivalry, Journal of International Economics 18, [5] Bulow, J., Geanokoplos, J. and P. Klemperer (1985), Multimarket Oligopoly: Strategic Substitutes and Complements, Journal of Political Economy 93, [6] Burguet, R. and J. Sempere (003), Trade Liberalization, Environmental Policy, and Welfare, Journal of Environmental Economics and Management 46, [7] Conrad, K. (1993), Taxes and Subsidies for Pollution Intensive Industries as Trade Policy, Journal of Environmental Economics and Management 5, [8] Dixit, A. (1986), Comparative Statics for Oligopoly, International Economic Review 7(1),

23 F. ANTONIOU, P. HATZIPANAYOTOU [9] Duval, Y. and S. Hamilton (00), Strategic Environmental Policy and International Trade in Asymmetric Oligopoly Markets, International Tax and Public Finance 9, [10] Ederington, J. and J. Minier (003), Is Environmental Policy a Secondary Trade Barrier? An Empirical Analysis, Canadian Journal of Economics 36(1), [11] Eaton, J. and G. M. Grosmman(1986), Optimal Trade and Industrial Policy under Oligopoly, Quarterly Journal of Economics 101, [1] Fredriksson, P. G. and D. L. Millimet (00), Strategic Interaction and the Determination of Environmental Policy across U. S. States, Journal of Urban Economics 51, [13] Greaker, M. (003), Strategic Environmental Policy; Eco-Dumping or a Green Strategy?, Journal of Environmental Economics and Management 45, [14] Grossman, G. M. and G. Maggi (1998), Free Trade vs Strategic Trade: A Peek into Pandora s Box, In: R. Sato, Ramachandran, R.V., Mino, K. (Eds.), Global Integration and Competition, Kluwer, Dordrecht, 9-3. [15] Hamilton, S. F. and Till Requate (004), Vertical Structure and Strategic Environmental Policy, Journal of Environmental Economics and Management 47, [16] Jaffe, A. B, S. R. Peterson, P. R. Portney and R. N. Stavins (1995), Environmental Regulation and the Competitiveness of U.S. Manufacturing: What Does the Evidence Tell Us?, Journal of Economic Literature 33, [17] Kennedy, P. (1994), Equilibrium Pollution Taxes in Open Economies with Imperfect Competition, Journal of Environmental Economics and Management 7(1), [18] Nannerup, N. (1998), Strategic Environmental Policy under Incomplete Information, Environmental and Resource Economics 11, [19] Neary, P. J. (006), International Trade and the Environment: Theoretical and Policy Linkages, Environmental and Resource Economics 33, [0] Puller, S. L. (006), The Strategic Use of Innovation to Influence Regulatory Standards, Journal of Environmental Economics and Management 5, [1] Rauscher, M. (1994), On Ecological Dumping, Oxford Economic Papers 46,

24 STRATEGIC ENVIRONMENTAL POLICY WITH ENDOGENOUS REGULATION 3 [] Simpson, D. R. and R. L. Bradford (1996), Taxing Variable Cost: Environmental Regulation as Industrial Policy, Journal of Environmental Economics and Management 30, [3] Straume, O. R. (006), Product Market Integration and Environmental Policy Coordination in an International Duopoly, Environmental and Resource Economics 34, [4] Ulph, A. (1996), Environmental Policy and International Trade when Governments and Producers Act Strategically, Journal of Environmental Economics and Management 30, [5] Ulph, A. and D. Ulph (007), Climate Change-Environmental Policies and Technology Policies in a Strategic Context, Working Paper (forthcoming).