Impacts of Emission Reduction Policies in a Multi-Regional Multi-Sectoral Small Open Economy with Endogenous Growth

TL;DR: In this paper, the authors explored the impact of environmental policies on capital accumulation in a dynamic endogenous growth setting with a small open multi-regional economy with an international tradable permits market and showed that capital accumulation is more negatively affected by environmental policies in the energy intensive sector.

Abstract: The burden sharing of pollution abatement costs raises the issue of how to share the costs between entities (country, region or industry) and how the pollution permits should be distributed between the parties involved. This paper explores this issue in the framework of a dynamic endogenous growth 2 sectors - 2 regions - 2 inputs Heckscher-Ohlin model of a small open multi-regional economy with an international tradable permits market. Given an emission-based grand-fathering sharing rule, capital accumulation is more negatively affected by the environmental policy in the energy intensive sector. We show that such a property does not necessarily hold with a production-based grand-fathering sharing rule. We also show that the impact on capital is likely to translate into the sectoral added value level after some time, specially if the economy is submitted to an increasingly constraining environmental policy driving up the ratio price of permits to price of energy. Finally, we show that the impact of environmental policy at the regional level depends crucially on the specialization of the region along the baseline.

Summary

1 Introduction

• The burden sharing of pollution abatement costs, e.g. in the Kyoto Protocol context, raises the issue of how to share the costs between entities (country, region or industry) and how the pollution permits should be distributed between the parties involved in the Protocol.
• One region is more specialized in the energy intensive sector.
• For a given sector, there are technological spillovers from one region to the other.
• Conditions are established under which the spread of specialization, that depends on the initial capital endowments and on the technological spillovers, might be reversed.

2 The model

• Given that the authors consider a small open economy, agents are price-takers and prices are determined by the Rest of the World and thus exogenous.
• Sector a produces capital goods, sector b produces consumption goods.
• National and foreign products of a certain type are supposed to be perfect substitutes.
• Emissions are linked to energy consumption, and for the sake of simplicity, e denotes simultaneously energy and emissions.

2.2 Endogenous growth and technical spillovers

• There is thus a interregional technological spillover at the sector level.
• The authors shall use this observation in some interpretations later on.
• Notice that ωjt is asymptotically geometric with coefficient θj 2−θj < 1, which implies that limt→+∞ ωjt = 0. ¤ This strong result follows from the fact that besides different initial sectoral dotation of capital, the two regions are identical : they face the same international prices and share the same technologies.
• In the presence of intra-sectoral spillovers, the divergence force coming from endogenous growth (namely, constant returns) is neutralized.

2.3 Specialization of the regions

• The authors also define the spread of specialization index at time t by the ratio of the specialization indexes of the two regions : σt = χwt χvt = kwat kwbt kvbt kvat.
• The purpose of the present subsection is to study what determines the evolution of the regions’ specialization through time.
• One observes the remarkable fact that the spread of specialization index depends only on the initial stocks of capital and on the spillovers parameters that are exogenous.
• As a consequence this index will not be affected by the environmental policies that are considered below.
• The authors have indeed the following proposition : Proposition 2.

3.1 Preliminaries

• The authors assume that the baseline tends in the long run to a balanced growth path characterized by the asymptotic constant growth rate g that applies to both sectors.
• As the authors aim to compare the impacts of an environmental policy (EP) between sectors and between regions, they state the two following important assumptions : 2This implies that in the long term (i.e. when the sequences are close to their limit values θj (j = a, b)) the ratio θb/θa is high enough.
• Then a remarkable fact is that the previous inequality cannot be satisfied whatever the spillover parameters (inversion of specialization is impossible).
• In a next subsection, the authors will also briefly consider a ”production-based grand-fathering” sharing rule in order to highlight how results can depend crucially on the chosen endowment rule.

3.2 Impact on the sectoral growth rates

• 5This could be justified if the rule applies at an aggregate level and if the firm is small and receives a fixed share of the total permits endowment.
• This is because the sectors j of both regions face the same exogenous prices and share the same technologies.
• The growth rate of the energy intensive capital sector a is more affected by the EP than the accumulation capital of the other sector if the endowment effect relative to sector a is not higher than the one relative to sector b.

3.4 The dynamic production-based grand-fathering rule

• To highlight how results depends crucially on the chosen endowment rule, let us consider briefly the dynamic production-based grand-fathering rule, where a firm’s current permits endowment depends on its previous period production6.
• The authors assume that this dependance is identical across sectors and regions and that it is internalized by firms when they maximise their profits.
• Because of assumption (33), the second term between brackets of the RHS is lower for sector a.
• But if pbt is increasing (as is observed empirically), then the first term between brackets of the RHS is lower for sector b, and the overall effect is undecidable without more information on the parameters.
• Configurations of parameters are possible where one obtains the opposite effect as the one obtained with the emission-based grandfathering rules analyzed above, i.e. with equality (39) is reversed (for example if the consumption good price increases sufficiently quickly because of a low exogenous technical progress ρt).

3.5 Impact on the sectoral added value

• In the remaining of the paper the authors return exclusively to the emission-based grand-fathering rules because they lead to more clear-cut conclusions.
• The second one measures the impact on the regional sectoral capital stock analyzed in the previous subsection.
• Indeed, because of the increase of total cost of energy, its share arises in the value of production.
• And this effect is stronger the higher βj.
• The above proposition is likely to be verified 14 in the case of a more and more constraining EP.

3.6 Impact on the sectoral revenues with transfers

• Let us first consider the non anticipated emission-based grand-fathering sharing rule.
• In this last equality appears sector j’s growth rate characterising the EP scenario.
• The inequality (49) necessarily holds after some time.
• Then the above proposition follows again by a similar reasoning.

3.7 Impact on the regional product

• In subsection 3.5, the authors have checked that vaijt vaBijt does not depend on i and that the inequality vaiat vaBiat < vaibt vaBibt is likely to be verified.
• There should not be any inversion of specialization at date t w.r.t. date 0. Now Proposition 2 states the conditions for a specialization inversion.
• This implies that in the long term, i.e. when the sequences are close to their limit values θj (j = a, b), the above inequality implies that the ratio θb/θa is low enough.
• This implies that in the long term, i.e. when the sequences are close to their limit values θj (j = a, b), the above inequality implies that the ratio θb/θa is high enough.
• If kwa0 − kva0 > 0, the regions have opposite specializations (i.e. kwa0 > kva0 and kwb0 < kvb0), then a remarkable fact is that the above inequality is necessarily satisfied whatever the spillover parameters.

3.8 Impact on the regional revenues with transfers

• 7Remember that the spread of specialization index is not modified by the environmental policy.
• In the previous subsection, the authors have checked the conditions ensuring that the inequality riat rBiat < ribt rBibt is verified.

4 Conclusion

• The model presented in this paper has been designed to study the burden sharing of pollution sharing in the context of a Kyoto-like protocol in a multi-regional multi-sectoral economy.
• In order to get the dynamic picture, the authors have considered time-dependent intrasectoral spillovers across regions and learning-by-investing in each sector.
• In such a context, the authors have shown progressively how the typical wisdom gathered for the static case can be altered.
• In particular, the authors have disentangled the main price-based and quantitybased mechanisms which determine the impact of environmental policy at different levels (sectoral value-added and regional notably).
• Nonetheless, a much more serious quantitative assessment is needed, and this would require a rigorous calibration of the model, including the exogenous price processes involved.

Full text

Département des Sciences Économiques
de l'Université catholique de Louvain
Impacts of emission reduction policies
in a multi-regional multi-sectorial small economy
with endogenous growth
R. Boucekkine and M. Germain
Discussion Paper 2007-10

Impacts of emission reduction policies in a
multi-regional multi-sectoral small open economy with
endogenous growth
∗
Raouf Boucekkine
†
and Marc Germain
‡
February 2007
CORE DISCUSSION PAPERS
2007/11
∗
Raouf Boucekkine acknowledges the support of the French National Agency for Research (ANR)
through an ACI contract on sustainable development and technological progress.
†
D´epartement d’Economie and CORE, Universit´e catholique de Louvain. Email:
boucekkine@core.ucl.ac.be
‡
EQUIPPE, Universit´e Charles-de-Gaulle Lille 3 and CORE, Universit´e catholique de Louvain. Email:
germain@core.ucl.ac.be .

Abstract
The burden sharing of pollution abatement costs raises the issue of how to share the costs
between entities (country, region or industry) and how the pollution permits should be
distributed between the parties involved. This paper explores this issue in the framework
of a dynamic endogenous growth 2 sectors - 2 regions - 2 inputs Heckscher-Ohlin model
of a small open multi-regional economy with an international tradable permits market.
Given an “emission-based grand-fathering” sharing rule, capital accumulation is more
negatively affected by the environmental policy in the energy intensive sector. We show
that such a property does not necessarily hold with a “production-based grand-fathering”
sharing rule. We also show that the impact on capital is likely to translate into the sectoral
added value level after some time, specially if the economy is submitted to an increasingly
constraining environmental policy driving up the ratio price of permits to price of energy.
Finally, we show that the impact of environmental policy at the regional level depends
crucially on the specialization of the region along the baseline.
Keywords: Pollution permits, Grand-fathering, Sectoral spillovers, Multi-regional econ-
omy, Endogenous growth
Journal of Economic Literature: D58, H21,E22,O40.

1 Introduction
The burden sharing of pollution abatement costs, e.g. in the Kyoto Protocol context,
raises the issue of how to share the costs between entities (country, region or industry)
and how the pollution permits should be distributed between the parties involved in the
Protocol. In Belgium, the debate opposes in particular the Flemish and Walloon regions.
W.r.t. the other regions, Wallonia is characterized by an industry which is more energy
consuming. From the point of view of Flanders, the bulk of the effort should thus be
made in Wallonia, where the abatement measures are assumed to be less expensive. On
the contrary, this solution is considered to be too unfavourable by Wallonia.
Now it is important to emphasize that the fact that a country’s activities are more
energy consuming does not necessarily result from their inefficiency. It can also result
from the specialization of this country in the production of relatively energy intensive
goods, a specialization conditioned itself by its comparative advantages and which should
benefit through international trade to all countries involved in it.
In the context of the Belgian burden sharing debate, Germain et al. (2006) develop
a model of a 2 sectors - 2 regions small open economy, where each sector produces one
good by using fossil energy and another sector specific factor
1
. Sector 1 is more energy
intensive and thus more polluting than sector 2. One region is more specialized in the
energy intensive sector. Hence, its energy consumption per unit of added value is higher.
Given an environmental policy that increases the price of energy (through an energy tax
or through the tradable permits price), the authors show as a first result that the energy
intensive sector is more burdened than the other sector, so that the region sp ecialized in
the energy intensive sector is more burdened than the other region.
Now this intuitive result does not necessarily extend when one takes into account the
tax revenues or the tradable permits endowments associated to the control of pollution.
The impacts of an emission reduction policy can indeed be modulated by bringing into
play such transfers assigned to the regions. In this respect the above authors compute
the permits endowments to the regions such that the environmental policy has both the
property to be efficient (i.e. abatement marginal costs are equalized between sectors and
regions) and fair (i.e. the relative losses of welfare are identical between regions). The
endowment is of course relatively more generous for the region specialized in the energy
intensive sector. And the difference between regional permits endowments is higher, (i)
the higher the national objective in terms of emission reduction, and (ii) the more the
regions are unevenly affected by the climate policy.
The analysis of Germain et al. (2006) presents the drawback to be static. However
a region or a country’s factor endowments and thus its sp ecialization are likely to evolve
through time. Consequently the impacts on this region or country of a long term environ-
mental policy (like climate policies) are also likely to evolve through time. On the other
hand, if the impacts of an environmental policy on a country’s welfare depend on its spe-
cialization, the dependence is likely to go the other way around. Indeed it seems probable
that an emission reduction policy that translates in an increase of the total cost of energy
will induce a change in a country’s specialization towards less energy intensive products,
and this seems the more likely to happen the more this country is initially specialized in
energy intensive goods.
In this paper, we shall consider a dynamic model allowing for specialization rever-
1
These specific factors can be considered as aggregates of all non-fossil energy factors (such as capital,
labour, infrastructure, non fossil energies,...) with different composition .
1

sal. One way to get such a property is to incorporate time-dependent spillovers across
economic sectors and regions. This might even be the easiest way to generate reversals
in specialization. There is an extensive literature about spillovers both at a regional or
international level. In particular, the empirical assessment of such spillovers have been
at the heart of a quite abundant empirical literature. An important and early contribu-
tion to the topic is due to Coe and Helpman (1995) who assessed the economic growth
impact of R&D expenditures in OECD countries. They found that such expenditures are
beneficial not only for the performing countries but also for the trade partners. Smolny
(1999) provided an empirical evaluation of international sectoral spillovers for German
and US industries using a broad panel of industry sector data for both countries over
the period 1960-1990. In particular, Smolny analyzed productivity convergence, trying
to disentangling the precise mechanisms behind. He found that most of the convergence
comes from total factor productivity convergence, and more importantly, that endogenous
growth models relying on knowledge spillovers are confirmed by the estimates. Lejour and
Nahuis (2005) studied in deep the sectoral nature of R&D spillovers and its impact on
economic growth. They stressed that the effects of sectoral spillovers do depend on the
specialization patterns.
There are also plenty of empirical contributions addressing the issue of spillovers’ extent at
a regional level. Among them, Van Stel and Nieuwenhuijsen (2004) is a good illustration
of this stream of literature. While they did not find any compelling evidence on the
growth effect of specialization, in contrast to Lejour and Nahuis (2005), they identified
some clear spillovers in certain sectors (specially in the sector services). As one can see,
the empirical debate on spillovers is currently much more centered on the kind of relevant
spillovers than on the existence of such spillovers, which seems largely admitted. One
of the crucial issues turn out to be whether intra-sectoral or inter-sectoral spillovers are
more important for economic growth. As to this precise point, the evidence is mixed. A
recent study by Malerba, Mancusi and Montobbio (2004) tends however to put forward
intra-sectoral spillovers. Using a unique panel data on R&D expenditures and patent
citations in 135 narrowly defined technological classes (or sectors) in France, Italy, Japan,
United-Kingdom and the US, over the perio d 1981-1995, they show that the effect of
intra-sectoral spillovers is 70% higher than the effect of national inter-sectoral spillovers.
We shall incorporate intra-sectoral spillovers in our model: the industries in a given sec-
tor of a given region are assumed to benefit from knowledge spillovers from the industries
of the same sector in the other region. As we shall see in Section 2, considering at
the same time inter and intra-sectoral spillovers in our model would induce the same
long-run capital accumulation in ALL sectors and in ALL regions, which sounds an un-
desirable outcome as it implies that the sectoral composition of the economy is irrelevant
in the long-run. Recent studies on two-sector growth mo dels tend rather to emphasize
that investment-specific technological progress (as opposed to the typical Harrod-neutral
technological progress at work in the consumption good sector) is likely to generate a per-
sistent pro ductivity gap between the capital good and consumption good sectors, which
should translate into different patterns of capital accumulation. This divergence is clearly
reflected in the downward trend of the relative price of capital, first pointed out by Gor-
don (1990), and later exploited in a two-sector accounting framework by Greenwo od,
Hercowitz and Krusell (1997).
The aim of this paper is to study the impacts of long term environmental policies in
the framework of a dynamic 2 sectors - 2 regions - 2 inputs (capital and energy) Heckscher-
2

Frequently asked questions

Q1. What are the contributions mentioned in the paper "Impacts of emission reduction policies in a multi-regional multi-sectoral small open economy with endogenous growth∗" ?

This paper explores this issue in the framework of a dynamic endogenous growth 2 sectors 2 regions 2 inputs Heckscher-Ohlin model of a small open multi-regional economy with an international tradable permits market. The authors show that such a property does not necessarily hold with a “ production-based grand-fathering ” sharing rule. The authors also show that the impact on capital is likely to translate into the sectoral added value level after some time, specially if the economy is submitted to an increasingly constraining environmental policy driving up the ratio price of permits to price of energy. Finally, the authors show that the impact of environmental policy at the regional level depends crucially on the specialization of the region along the baseline. 

Q2. What are the future works mentioned in the paper "Impacts of emission reduction policies in a multi-regional multi-sectoral small open economy with endogenous growth∗" ?

The model presented in this paper has been designed to study the burden sharing of pollution sharing in the context of a Kyoto-like protocol in a multi-regional multi-sectoral economy. In such a context, the authors have shown progressively how the typical wisdom gathered for the static case can be altered.