No country is an island: international cooperation and climate change

Working Paper Series

Massimo Ferrari, Maria Sole Pagliari No country is an island:

international cooperation and climate change

No 2568/ June 2021

Disclaimer: This paper should not be reported as representing the views of the European Central Bank (ECB). The views expressed are those of the authors and do not necessarily reflect those of the ECB.

Abstract

In this paper we explore the cross-country implications of climate-related mitigation policies. Specically, we set up a two-country,two-sector (brown vs green) DSGE model with negative production externalities stemming from carbon-dioxide emissions. We estimate the model using US and euro area data and we characterize welfare-enhancing equilibria under alternative containment policies. Three main policy implications emerge: i) scal policy should focus on reducing emissions by levying taxes on polluting production activities; ii) monetary policy should look through environmental objectives while standing ready to support the economy when the costs of the environmental transition materialize; iii) international cooperation is crucial to obtain a Pareto improvement under the proposed policies. We nally nd that the objective of reducing emissions by 50%, which is compatible with the Paris agreement's goal of limiting global warming to below 2 degrees Celsius with respect to pre-industrial levels, would not be attainable in absence of international cooperation even with the support of monetary policy.

Keywords: DSGE model, open-economy macroeconomics, optimal policies, climate modelling

JEL Codes: F42, E50, E60, F30

ECB Working Paper Series No 2568 / June 2021

1

Non-technical summary

The discussion on the impact of climate change and of potential mitigation policies has gained momentum both in academia and policy circles over the last decades. Heatwaves, oods and natural disasters at broad are raising the awareness that the long-neglected costs of adverse climatic events might materialize sooner than expected. Scienti c studies, endorsed by international agreements, have estimated that countries should reduce their emissions by about 50% in order to maintain the increase in temperature below 2 degree Celsius over the next century (IMF (2019)). Research networks involving central banks and policy institutions have been created to discuss how to best tackle climate change and how to calibrate mitigation policies. A recent report by the Network for the Greening of the Financial System (NGFS), organized by 87 central banks and supervisory authorities, for example, provides a detailed analysis of the potential impacts of climate change onto the economy and the central banking operations (NGFS (2020)).

Despite there exist robust empirical evidences on the macroeconomic costs of higher emission levels (Nordhaus (1994) and Hsiang et al. (2017)), there are relative few structural macro models featuring emission externalities that can be used to analyse the trade-o of di erent containment policies. Moreover, most of the existing macro-literature mainly makes use of closed-economy frameworks with no cross-country interaction (e.g., Heutel (2012), Ferrari and Nispi Landi (2020) and Dietrich et al. (2021)). Against this backdrop, our paper provides four new con- tributions. First, we derive an open economy general equilibrium model where emissions and their spillovers can be studied in a structural framework. In our setting there are two countries, each of which produces brown" and green" goods. The two goods are perceived as similar by consumers, but the production of brown goods generates a negative emission externality. In this context, the cross-country dimension is particularly relevant because emissions produced in once country a ect also the other. As a consequence, only cooperative actions are successful in reducing climate risk at the global level. In economic terms, this is a coordination problem, as actions produce the maximum bene ts only if taken jointly. However, pro t-maximizing agents might refrain from doing anything as they would bene t more by waiting for other agents to act. Second, we estimate the model with US and euro area data to study how emissions patterns across the Atlantic have changed over the last 20 years. Notably, in our framework the social cost" of emissions, which is given by the GDP loss due to emissions in the steady state, amounts to around 1.2% of GDP in the US, a gure which looks more realistic and aligned to the empirical estimates compared to other calibrated models' predictions. Third, we compare di erent policies

ECB Working Paper Series No 2568 / June 2021

2

that can be deployed to reduce emissions: i) a change in monetary policy objectives, whereby the central bank pursues a double mandate of maximizing welfare and reducing emissions; ii) a change in domestic scal policy, whereby scal authorities start to directly tax emissions; iii) a change in trade policy,whereby one country implements tari s targeting polluting imports from the foreign economy. We additionally evaluate the consequences of coordinated and competitive actions by agents. Speci cally, we show that the non-cooperative equilibrium is characterized by an insucient level of taxation on emissions, as both countries attempt to entirely pass the cost of emissions containment on to their respective counterpart. Therefore, only coordinated policies can achieve the climate objective when scal and monetary policy interact. Notably, the best policy mix is the one where governments focus on reducing emissions, while the central banks intervene to reduce the welfare costs of environmental taxation.

• Introduction

The discussion on the impact of climate change and of potential mitigation policies has gained momentum both in academia and policy circles over the last decades. Heatwaves, oods and natural disasters are raising the awareness that the long-neglected costs of adverse climatic events might materialize sooner than expected. Scienti c studies, endorsed by international agreements, have estimated that countries should reduce their emissions by about 50% in order to maintain the increase in temperature below 2 degree Celsius over the next century (IMF (2019)). Networks among central banks and policy institutions are now discussing how to best tackle climate change and how to calibrate mitigation policies. NGFS (2020), for example, provides a detailed analysis of the potential impacts of climate change onto the economy.

In spite of the wide and growing empirical literature that tries to quantify the costs of climate events and to assess the relationship across emissions, climate disasters and economic perfor- mance, few structural models have been developed in this eld1. Such frameworks could be anyways useful to study the implications of policies that have arguably never been implemented before and, therefore, cannot be ascertained on the basis of past data. Among the existing contributions on this topic, Heutel (2012) is the rst to provide a structural model where emissions are endogenous and a ect output.2 One advantage of using structural models is that they

• Notably, there are several empirical contributions that focus on the assessment of physical and transitional climate risks in the nancial sector. See, for instance, Pagliari (2021) for a study of the relationship between climatic adverse events and banking performance in the euro area, or Allen et al. (2020) for the setup of an analytical framework to quantify the impacts of climate policy and transition narratives on economic and nancial variables that are necessary for nancial risk assessment. Finally, refer to Tol (2009) for a thorough survey of the
  early theoretical macroeconomics literature on climate change.
  2There indeed exists a wide literature on the quantication of the cost of climatic events, for example Nordhaus (2008). However models in the spirit of DICE or FUND have been developed to account for several sources of

ECB Working Paper Series No 2568 / June 2021

3

can be used to formally study optimal policy problems and identify trade-os between dierent specications of the policies of interest. The literature on optimal policy in DSGE model is vast, including the pioneering work of Woodford (2003) on monetary policy, but only recently these tools have been applied to the design of climate change mitigation policies (Benmir et al. (2020), Kotliko et al. (2020)). Against this backdrop, our paper provides ve relevant contributions. First, we show that there are relevant non-linearities in the relationship between carbon emis- sions, a common measure of climate externalities, and macro variables. For example, when the level of emissions is low there is a positive correlation between CO2 and GDP, because more output leads to more emissions, but the emission stock is too low to generate sizeable climate events. When the stock of CO2 is high, on the contrary, that correlation drops because more emissions directly increase the severity of climate shocks leading to GDP losses. These results suggest that empirical models, if anything, underestimate the real costs of climate externalities on full sample estimations. Second, we derive an open economy general equilibrium model where emissions and their spillovers can be studied in a structural framework. Specically, we set up a rich two-countrytwo-sector model, where, on the one hand, brown" production generates a negative environmental externality that is detrimental to both domestic and foreign output, while, on the other hand, green" production does not. This cross-country dimension is crucial in the debate on climate change, because only a fall in the global stock of emission would reduce the likelihood of a climatic disaster", whereas isolated actions might result insucient. In economic terms, this is a coordination problem, as eorts produce the maximum benet only if taken jointly. However, prot-maximizing agents might refrain from doing anything as they would benet more by waiting for other agents to act, without bearing any direct cost.

Third, we estimate the model with US and euro area data to study how emissions patterns across the Atlantic have changed over the last 20 years. In our framework the social cost" of emissions, which is given by the GDP loss due to emissions in the steady state, amounts to around 1.2% of GDP in the US, a gure which looks more realistic and aligned to the empirical estimates (Hsiang et al. (2017)) compared to other models' predictions (Heutel (2012)). Fourth, we evaluate the implementation of three alternative policies to reduce emissions: i) a change in monetary policy objectives, whereby the central bank pursues a double mandate of maximizing welfare and reducing emissions; ii) a change in domestic scal policy, whereby scal authorities start to directly tax emissions; iii) a change in trade policy, whereby one country implements taris targeting polluting imports from the foreign economy. We show that both monetary policy and taris are not eective in reducing emissions, whereas domestic taxation can achieve

pollution and types of climatic events, but often rely on reduced-form equations for production and consumption side of the economy, which make them less appealing than structural models for the welfare analysis of policies.

ECB Working Paper Series No 2568 / June 2021

4