Emissions trading with transaction costs

Produced as part of the Competitiveness in the low-carbon economy CCCEP research programme theme

While the prevalence of transaction costs in past and existing emissions trading schemes is a well-documented fact, the related theoretical literature has to a large extent ignored this practical issue and its implications for policy outcomes.

In this paper the authors address this gap by developing a theoretical model of emissions trading in the presence of transaction costs. They calibrate the model to annual transactions and compliance data in the European Union emissions trading scheme (EU ETS) over its second phase (2008–2012).

The analysis shows that accounting for transaction costs when modelling permit markets has equally important implications for evaluating past policy outcomes and assessing the impacts of possible future design of policy.

Key points for decision-makers

  • The authors develop a consolidation procedure to aggregate transactions and compliance data from the installation to the firm level. They use this consolidated dataset to scrutinise firms’ annual market behaviour in Phase II of the EU ETS (2008–2012).
  • They uncover two important empirical facts: first, autarkic behaviour is pervasive: 30 per cent of regulated firms, representing 10 per cent of covered emissions, achieve compliance without entering the market to trade. This is especially the case among firms that are small in size or hold more permits than needed to cover their emissions. Second, those firms that engage in trade do so quite sporadically (typically only a few times per year) and only for sufficiently large transacted volumes (above a certain cut-off).
  • These observations that trading is impaired and that some gains from trade go unrealised are not in line with the basic tenets of emissions trading. The authors thus interpret them as pointing to the existence of both fixed market entry costs and variable trading costs.
  • The authors enrich a standard permit trading model by formally incorporating fixed and proportional trading costs. In their equilibrium framework, firms individually decide to participate in trading or not, and if so by how much, and the permit price results from firms’ interactions. The authors then analyse the sensitivity of the market equilibrium to shifts in the trading costs and firms’ allocations.
  • They formally investigate the properties of the market price impacts of such changes, and characterise situations where the trading costs alternatively depress or raise permit prices relative to frictionless market conditions (i.e. where there are no transaction costs).
  • The authors then calibrate their model to EU ETS Phase II transactions data and show that trading costs in the order of €10,000 per annum plus €1 per permit traded noticeably reduce the number and dispersion of discrepancies between practical observations (i.e. that trading is impaired) and the model’s predictions for firms’ trading behaviour (i.e. participation in and extent of trading).
  • In this illustrative context, their simulations also suggest that ignoring trading costs may lead modellers to underestimate the price impacts of supply-curbing policies, with the size of this underestimation bias notably hinging on the specific incidence of such policies on firms.