Competitive carbon emission yields the possibility of global self-control

Abstract

Despite many international climate meetings such as Copenhagen 2009, it is still unclear how annual global emissions can be reduced without requiring governments to micro-manage the emitting companies within their individual jurisdictions. Here we examine a simple, yet highly non-trivial, computer model of carbon emission which is consistent with recent activity in the European carbon markets. Our simulation results show that the ongoing daily competition to emit CO₂ within a population of emitters, can lead to a form of collective self-control over the aggregated emissions. We identify regimes in which such a population spontaneously hits its emissions target with minimal fluctuations. We then focus on the emission dynamics induced by a governing body which chooses to actively manage the capping level. Finally we lay some formal stepping stones toward a complete analytic theory for carbon emissions fluctuations within this model framework – in so doing, we also connect this problem to more familiar theoretical terrain within computer science.

Introduction

Climate change, and the need to reduce carbon emissions, are seen as urgent global issues [1], [2]. Short of inventing some new technology, the most significant initiative for bringing about emissions reductions is the introduction of carbon markets [2]. These markets tend to be so-called ‘cap and trade’ markets in which an emissions cap is established and a quantity of emissions credits up to this cap then assigned to the population of emitters. These credits can then be traded among the market participants. Such schemes have already been adopted by the European Union (EU), which claims that emissions trading is a cheap way to reduce emissions since reduction is made at source, and also a clear way to ensure that the cap is lowered; but others argue that enforcing the reductions is an expensive process because each emitter will need to be individually micro-managed [2].

Here we use computer simulations to explore a complex systems model of a simple carbon market, in which a group of adaptive companies compete to emit in a market with minimal global control. We show that within the basic constraints of the model, companies are able to organise themselves to hit their collective monthly emissions target, with minimal fluctuations in the aggregated emissions each month, and in the absence of any external regulator controlling the market. We also study for the first time, the effects of allowing a government to have dynamic control over their emissions cap, such that they can change its value during the year. We find that companies react quickly to changes in the monthly emission limit, and that there is limited difference between this behavioural change for both an incremental and a sudden cap reduction. This provides insight into the most efficient method for reducing the emissions cap within the carbon market, in real time, during a given year.

Given the complexity of this dynamical, multi-agent problem, the ultimate future goal is to derive full analytic expressions which produce quantitative agreement with the simulations. We finish the paper by taking several steps toward solving this open challenge. In particular, we provide a formal analysis of the model, and a derivation of explicit expressions for the fluctuation size, in the simplified case where the emissions cap L is approximately half the number of emitting entities, and the memory of each emitting entity is relatively small. Though not a general solution, this analysis provides a strong platform for future analytic extensions. Just as importantly, by mapping the emissions problem to a binary system we are able to connect up to more familiar topics in the computational science field, e.g. strategies as look-up tables, differences in strategies in terms of Hamming distances, and the global information and individual emitting entity memory as finite length bit-strings. This formal association should enable computer science researchers to apply known computer science measures and techniques to deepen the understanding of the system. More generally, it provides a novel link between the topic of carbon emissions and more formal topics within the fields of computer science and information theory.