Environmental Regulation Regime Shift A Game-Theoretic Analysis

Transcription

1 Environmental Regulation Regime Shift A Game-Theoretic Analysis Yusen SUNG National Taiwan University Department of Economics Taipei, Taiwan ( ) ysung@ntu.edu.tw September 18, 2017 Abstract Throughout the US environmental regulation history, centralized regulation by the federal EPA has been the norm. More recently, however, state governments have taken a more active role by imposing more stringent environmental standards than the federal requirements. In this paper we set out to explain how this phenomenon may occur. Earlier, Williams III (2012) posits that this may be caused by a policy regime shift from the early command-and-control to more recent incentive-based regulations. In this paper, we adopt both simultaneous-move and sequential-move regulation models, and consider both quantity and price policy regimes to see if and when his theory is valid. It is found that the policy regime shift theory can only hold in a sequential-move setting in which local EPAs lead. In the early emission cap regime, only federal regulation exists in equilibrim, with the local authorities abstaining. Whereas under the new pollution tax regime, the case is reversed with the local authorities being the pre-dominant regulators. In addition, we also incorporate policy implementation and enforcement costs in our analysis. It is found that equilibrium policies of the federal and state governments hinge on these costs. When policy implementation exhibits economy of scale, the federal EPA will take on the major role in environmental regulation. Keywords: environmental federalism, policy regime shift, policy and enforcement. JEL classification: H23, H41, D64 1

2 1 Introduction Beginning nearly five decades ago, dating back to the 1960s, a spectacular federal environmental regulation has proliferated in US, with the Congress enacting a whole range of environmental legislation establishing and centralizing new responsibilities for the federal Environmental Protection Agency (EPA). 1 These congressional mandates kept most of the fundamental policy-making authority for federal agencies. With the spawning of federal environmental statutes, federal regulatory program expenditures have also grown substantially. As a result, most states relied on federal funding to run local environmental programs, and federal-state conflicts in environmental regulation were rare. In the past decade, however, things are quite different. State environmental agencies have become more active in setting new pollution standards, sometimes even aggressively over-riding federal regulations. How can we explain this drastic change in federal-state regulatory relationship? Weconstruct a simple model inthis paper toshow that this change may result froman environmental policy regime shift. We find that in the traditional command-and-control (e.g., emission cap) regulation regime, a social welfare maximizing federal agency will impose tighter emission caps than the local states to account for the spillovers effects of the pollutants, therefore leaving the local states nothing more to do. Whereas in the new incentive-based (e.g., emission tax) regulation era, local welfare maximizing states who prefer lower pollution tax than the federal agency will set a tax rate just high enough to make the federal agency adstain. In the relatively small environmental federalism literature, many works, such as Potoski (2001), Wellisch (2000), Oates (2001, 1999), and Dwyer (1995), focus mainly on the so-called race to the bottom theory, which claims that decentralized regulation makes local governments compete for firm investment by lowering their emission standards below that of a centralized regime. Therefore local environmental regulations will be set too lax compared to the efficient level. The few papers on environmental federalism that compare incentive-based and commandand-control regulations (e.g., Nordhaus (2006) and Wellisch (2000)) assume that regulation is entirely done by either the federal or the state governments. None of them consider the case in which both federal and state levels may act to control the same pollutant. As 1 Early statues include the Motor Vehicle Pollution Control Act of 1965 and the Air Quality Act of

3 far as we know, the only prior work that recognized the importance of the policy regime shift is Williams III (2012). They forward a compelling explanation as to why local agencies became active in the environmental policy arena similar to ours. Our model differs from their in two important aspects. Firstly, we construct a sequential-move Stackelberg game, instead of a simultaneous-move Nash game. Secondly, we incorporate explicitly policy implememtation/enforcement costs in our model to capture the reality of a policy process. This paper is organized as follows: The analytical framework is first laid out in the next section. In the third section we make a preliminary investigation without the policy implememtation/enforcement costs. The fourth section then take into account of the policy implememtation/enforcement costs and solve for the Stackelberg equilibrium of our sequential-move policy game. The final section concludes and summarizes our findings. 2 The Federal-State Government Model Consider a hierarchical government with two regulation levels in a country: one federal/central environmental protection agency (the EPA player) and N identical state/local regulators (the local players). Firms in all local states generate the same pollutant that causes external damage to every other state. All local state firms possess identical pollution control technology, and thus have identical emission abatement costs C(e i ), which is twice-differentiable, decreasing, and strictly convex in emission level e i : C (e i ) < 0, C (e i ) > 0. (1) State i s marginal cleanup cost (MCC) is thus: C (e i ) > Pollution Externality External damages caused by the pollution are assumed global (or national), affecting each state by its aggregate level: D(E), E e 1 + +e N 3

4 which is twice-differentiable, strictly increasing, and convex: D (E) > 0, D (E) > 0. (2) It is well established in the environmental economics literature and textbooks that measures have to be taken by environmental regulators to curb the inefficiency. In our model, both the federal EPA and the local regulators may act to reduce firm emissions. 2.2 Independent Government Goals We find out first what each regulator would do if it is the sole regulator of the pollutant in its own jurisdiction. In this case, it is assumed that each government will strive for abatement efficiency in its jurisdiction only. This will then determine how it will react to regulatory competition from other environmental authorities. Consider the federal government (the EPA player, F) first. As a national authority, its policy goal is to minimize total national costs: min {e i } TNC 0 (e 1,,e N ) i C(e i ) + n D(e 1 + +e N ) which yields the necessary first-order condition for optimal local/state emissions: i : C (e i ) = n D (e 1 + +e N ) (3) Plainly speaking, full efficiency requires firms in all states have equal marginal cleanup cost, which must again equal to marginal national damage, and hence: e 1 = = e N Let this equated emission level be denoted by s. To achieve its objective, the federal player F may set a state-level emission cap s F : s F = s or alternatively may opt for a corresponding uniform emission tax t F : t F = t = n D i (n s ) Next we turn to the local State players. As we assume identical states, we will henceforth focus on symmetric strategy equilibrium of the State players. Each state then would like to minimize total local costs: min e i TLC 0 i(e 1,,e N ) C(e i ) + D(E) 4

5 which requires the state MCC being equal to marginal state damage: C (e i ) = D (n e i ) (4) Let this symmetric emission be denoted by ŝ. To attain local efficiency, each State player may either impose an emission cap s S : s S = ŝ or adopt an emission tax t S equal to local marginal damage: t S = ˆt = D (n ŝ) It can be noted by comparing (3) and (4) and by assumptions (1)(2) that central policies tends to be more stringent than local ones due to pollution spillovers: s < ŝ (5) t > ˆt (6) 3 Preliminary Analysis 3.1 The Simultaneous-Move Nash Game First, we consider both quantity and price price games in a simultaneous-move Nash setting. 3.2 Quantity Competition in Emission Caps In this case, the Nash reaction functions are, respectively: { s F (s S s max, if s S s ) = (7) s, if s S > s { s S (s F s max, if s F ŝ ) = (8) ŝ, if s F > ŝ with s max ( ) being a very large cap that cannot be binding for firm emissions. The Nash equilibrium is then: s F = s 5

6 s S = s max which means that the central EPA is the only active regulator in this game, with local EPA not abstaining, as shown in the figure below. s F S F s (s ) s max s^ s* F S s (s ) s* s^ s max s S Price Competition in Pollution Tax/Charge The Nash reaction functions are now, respectively: { t F (t S 0, if t S t ) = (9) t t S, if t S < t t S (t F ) = { 0, if t F ˆt ˆt t F, if t F < ˆt (10) and the Nash equilibrium is then: t F = t t S = 0 which means that the central EPA is again the sole regulator in this game, setting the higher tax rate it desires. The local EPA, meanwhile, stays silent, as shown in the figure below. 6

7 t F S F t (t ) t* t^ t^ t* F S t (t ) t S 3.3 The Sequential-Move Game: Central EPA as Leader We explain in the previous section that policy regime shift(from quantity to price policies) will not cause local EPAs to be more aggressive in environmental regulation: central EPA will always prevail whatever policy regime is. Next we turn to sequential-move models, in which central and local EPAs act sequentially, hoping to see if and when local EPAs may be more dominant than their central counterpart. We construct a two-stage model to analyze the interaction between the central and the local regulators. In practice, we have witnessed occasions in which either Fed or State took the leadership role in real regulation cases. For transboundary pollution problems such as acid rain (SO x ) and green-house gases (CO 2 ), it is often the central EPA who took the initiatives in controlling the pollutants. Whereas for discharges that cause mainly regional externalities, it is usually the local EPAs that took the first step in emission regulation Quantity Competition in Emission Caps First we consider the case in which the central EPA announce its policy first (as a leader), followed by the local EPAs (the followers). With the central EPA being the policy leader, it expects the local EPAs to act afterwards according to Eq. (8). Therefore, it will simply choose s F = s to drive the local EPAs out of the arena. The Stackelberg equilibrium is hence: s F = s s S = s max 7

8 exactly the same as the Nash ones. The quantity cap chosen by the central EPA leaves absolutely no room for the local EPAs to intervene in the policy regulation Price Competition in Pollution Tax/Charge In the alternative price policy regime, the local EPAs will react to the central EPA s pollution tax t F according to Eq. (10). Knowing this, the central EPA will choose t = t, higher than local EPAs preferred rate ˆt, to deter the local EPAs potential intervention. As a result, the central EPA is still the sole active player in the equilibrium. 3.4 The Sequential-Move Game: Local EPA as Leader So far, we have found no case in which local EPAs actually participate in envirenmental policy regulation. Finally, we come to the case in which local EPAs act before the federal EPA in a sequential-move setting. That is, the local players announce their emission policy first (as a leader). The central EPA, after observing local policies, then determines its national requirement (as the follower). Both players may, however, choose to abstain (or opt out) if they find it more appealing for their regulatory goal. To explain how central and local regulators may change their regulation strategies when policy regime shifts, we consider and compare both quantity game (in which players set emission caps) and price game (in which players set emission tax) Quantity Competition in Emission Caps In early environmental regulation history, traditional command-and-control regulation is the standard way of pollution control. Regulators simply set limits/caps on local emissions. State firms then have to comply with both requirements, and hence must not exceed the tighter limit: e i = min{s S,s F } In this case, in equilibrium, a regulator (Fed or State) will either impose a binding cap or abstain (not bother to regulate) as policy enactment is politically troublesome and unappealing. As such, there will only be one active regulator in Stackelberg equilibrium. Since the central F intends to set a more stringent emission limit than the States (by (5), s < ŝ), all State players will abstain and leave the regulatory duty to Fed player: 2 s S =, s F = s 2 Note that this conclusion is robust and still holds even if Fed leads the policy game instead. 8

9 This result serves to explain why the central agency has always been the major(quite often the only) regulator in the old times when emission standards were ubiquitously used as the regulation instrument. Because non-binding or lax state policies have no effect under this regime, the local agencies have no reason to engage in the regulation competition to override central policies Price Competition in Pollution Tax/Charge In recent decades, environmental economists have triumphed in persuading politicians and policy makers to switch from direct control to incentive-based price instruments (tax/subsidy or permits). Along with this policy regime shift, state regulators become ever more active in setting local policies. In this section, we will show that this trend accords with the equilibrium outcome of our model. In the case of emission tax, state firms have to pay taxes to any (federal and/or state) authority that obligates them to do so. Faced with pollution tax rate t F and t S set by respective regulators, each local firm will choose an emission level equating its MCC with the total marginal tax rate: C(e i ) = t F + t S Since Fed prefers a higher tax than State (by (6), t > ˆt), it can always choose t F to make sure total tax rate sum up to t. In this regard, states will be indifferent about its tax rate t S concerning firm emissions. The local governments, however, will not simply let go and leave the regulation arena. Even thoughlocalfirmswill befinedeventually attheaggregatetaxratet, itstillmatters to the state governments where the tax revenues go. Though local tax payments are considered pure transfer within local jurisdictions, payment by local firms to the federal government is arguably a loss to local states. Therefore, local governments conceivably would prefer to retain it rather than to render it to another government. To preserve largest possible tax revenues in local treasury, each state will announce the maximal rate: t S = t and hence leave no room for federal tax. As a result, federal government will not impose further emission tax beyond the state requirement: t F = 0 9

10 These preliminary findings suffice to explain why devolved (decentralized) authority became the mainstream in the 1990s when US shifted from traditional quantity policy regime to incentive-based (tax or permit) regime in environmental regulation. 4 Costly Policy Implementation and Enforcement Next we add new features to the base model in previous section for a more complete and realistic analysis. In a real-world policy-setting process, there are costs associated with policy implementation and enforcement to be borne by regulators, whoever central or local. For the federal regulator, there are costs related to implementing the optimal policy. This is assumed to be a one-time fixed cost: K which include costs of designing policies, negotiating with congress in democratic processes, smoothing industry resistance, as well as policy legislature, etc. Furthermore, once a policy is enacted, it needs to be enforced for planned effects. The costs of enforcement will notably depend on firm emission level: F(e), F > 0, F > 0 which comprise expenses for firm emission monitoring and violator prosecution. With this new model feature, the central EPA has to make a binary participation decision first as to whether to invest K in policy implementation. Upon participation, it then, in the next stage, determines the optimal central policy to enforce: min {e i } TNC 1 (e 1,,e N ) i C(e i ) + i D(E) + i F(e i ) (11) which yields the interior first-order condition: i : C (s ) = n D (n s ) + F (s ) (12) Compared with the no-cost case, the optimal central policy now is tighter due to costly policy enforcement: s is lower, and the corresponding optimal tax rate t is higher: t = n D (n s ) + F (s ) (13) 10

11 Similarly for the local states, policy implementation and enforcement involves real expenditures. As the political ecosystems differ in central versus local governments (due maybe to different jurisdiction sizes), we assume that state regulators face a somewhat lower implementation costs than the central EPA: K i (< K) However, policy enforcement process would be identical regardless of regulator identity (central or local), and hence state enforcement costs are the same as the federal cost: F i (e i ) = F(e i ) Quite like the central EPA, local states first decide whether to participate in the policy gamebyexpending implementation cost K i. If so, statesthenmove ontothesecond-stage enforcement choice: min e i TLC 1 i(e 1,,e N ) C(e i ) + D(E) + F(e i ) (14) which gives us the interior condition: C (ŝ) = D (n ŝ) + F (ŝ) (15) Similar to the central authority case, local regulation becomes tighter with the extra policy enforcement costs: state-optimal emission cap ŝ is lower, and the corresponding state emission tax ˆt goes higher: ˆt = D (n ŝ) + F (ŝ) (16) As the existence of policy enforcement costs tightens emission regulation in both centralandlocallevels, theformerstilltenstobemorestringentthanthelatter, byconditions (12)(15): s < ŝ (17) t > ˆt (18) 4.1 The Sequential-Move Emission Cap Game We first investigate the quantity control regime of environmental regulation, in which both central and local authorities use emission cap as the policy instrument. The same 11

12 logic in the preliminary analysis applies here. Since state firms are bound by the lower emission limit: e i = min{s S,s F } only one (either federal or state) can be binding. 3 and hence only one authority will be active. In the absence of regulation costs, the federal authority is shown above to be the sole regulator in the emission cap game. Now with costly regulation, the federal authority will choose to engage in the policy war if and only if social gains of its doing so outweigh the associated social loss: TNC 1 (s ) + K < [ ] TLC 1 i (ŝ) + K i i which means: [C(s )+D(S )+F(s )] + K < i [C(ŝ)+D(Ŝ)+F(ŝ) ] + i K i i or equivalently: K i K i < i [C(ŝ)+D(Ŝ)+F(ŝ) ] i [C(s )+D(S )+F(s )] i.e., the efficiency improvement (RHS, > 0) is greater than the excessive policy implementation costs (LHS). More specifically, if policy implementation exhibits economy of scale as the usual case, that is, K < i K i then: K i K i < 0 < i [C(ŝ)+D(Ŝ)+F(ŝ) ] i [C(s )+D(S )+F(s )] and the central agency will certainly act and set: s F = s Knowing what the federal agency will do, the states, though as leaders, will abstain in the first stage. 3 The case of s S = s F is not possible with the existence of policy implementation costs. 12

13 4.2 The Sequential-Move Pollution Tax Game Finally, we turn to the price regulation regime, in which all regulators use tax as the legal policy instrument. In this scenario, both central and local regulators may be involved when they all actively impose positive emission taxes (t F > 0, t S > 0). If this actually happens, assignment of enforcement duty can be arbitrary and clumsy. We assume that, whenever federal policy is in effect, it is responsible for enforcement. Otherwise, there exists only state regulation, and it is the local governments duty to enforce it locally. With the two-stage central-local Stackelberg framework presented in previous sections, we now solve for the sub-game perfect equilibrium of the policy game. We start backwards with the social welfare maximization problem of the federal authority in stage 2. Given state choice of t S and corresponding s S satisfying (16) in stage 1, the central government will intervene and set t F = t t S only if: 4 K TNC 1 (s ) TLC 1 i (ss ) Otherwise it will consider intervention not worthy and hence abstain. Therefore, there will exist a state tax threshold t and corresponding emission ē: C (ē) = t i satisfying K = TNC 1 (e ) i TLC 1 i (ē) above which the federal government will abstain. As such, the optimal stage-2 federal strategy will be: t F = { t t S, if t S < t 0, if t S t as shown in Figure 1 (left panel), and the final tax rate t will be: { t = t F + t S t, if t S < t = t S, if t S t as shown in Figure 1 (right panel). 4 Note that local policy implementation costs K i incurred in stage 1 (if t s > 0) are left out here as forgone sunk social costs for the federal player in stage 2. In addition, enforcement burden F( ) is deemed pure transfer (shifting from local to central level) and cancels out in the national perspective. (19) (20) 13

14 t* t* 45 F S t (t ) F (t > 0) F (t = 0) t t*-t t t S 45 a t b t* t S Figure 1: Federal choice t F and final tax rate t as a function of t S Back up in stage 1, expecting the federal agency to behave as described in (19), the state agencies will choose local tax rate t S to make final equilibrium tax rate t as close as to ˆt (as defined in (16)) as possible. Depending on the value of the threshold t, there can be two possible cases. Case 1: 0 < ˆt < t < t, the state s preferred tax rate ˆt is lower than the threshold t, such as point a in the right panel of Figure??. In this case, local states will set t S = t to drive central agency out of the game. For t S < t, the central Fed will intervene, and the final tax rate t = t is way too high. If on the other hand, the states set t S > t, the central Fed will surely abstain, but rate t = t S > t > ˆt also gets jacked up unnecessarily. Case 2: 0 < t ˆt < t, the state s preferred tax rate ˆt is higher than the threshold t, such as point b in the right panel of Figure??. In this case, the local states can simply choose their ideal tax rate t S = ˆt and the central government will stay out anyway. It is noted that, in both cases, the central federal authority will abstain: t F = 0 14

15 and local states are the sole regulators who get to keep all the tax revenues. In summary, in Stackelberg equilibrium, the local states will always be active and impose emission tax: { t, t S if 0 < ˆt < t < t = (21) ˆt, if 0 < t ˆt < t Subsequently in stage 2, the central Fed will abstain: t F = 0 In the end, equilibrium tax rate will always be inefficiently lower (suboptimal) than the case without policy costs. 5 Conclusion We find in this paper that, when quantity (command-and-control) policy is the de facto instrument used by all government agencies, the central agency will dominate in the policy arena as the sole regulator. In this case, local agencies will be silent and refrain from setting any emission standards at all. In contrast, when both central and local agencies compete with price (incentive-based tax or permit) policy, only local states will set regulations, with federal agency abstaining from imposing pollution taxation. These findings serve well to explain the observed trend of decentralized regulation in the past few decades. 15

16 References 1. Dwyer, J.P.(1995) The Practice of Federalism Under the Clean Air Act, Maryland Law Review, 54, Nordhaus, W.D. (2006) After Kyoto: alternative mechanisms to control global warming, American Economic Review, 96(2), Oates, W. (1999) An essay on fiscal federalism, Journal of Economic Literature, 37(3), Oates, W. (2001) A Reconsideration of Environmental Federalism, Resources for the Future Discussion Paper No Potoski, M. (2001) Clean air federalism: do states race to the bottom? Public Administration Review, 61(3), Wellisch, D. (2000) Theory of Public Finance in a Federal State, Cambridge University Press, Cambridge, UK. 7. Williams III, R.C. (2012) Growing state-federal conflicts in environmental policy: The role of market-based regulation, Journal of Public Economics, 96,