The single most important policy package adopted under the UK Climate Change Programme is comprised of the so-called Climate Change Levy (CCL) and Climate Change Levy Agreements (CCA). The package was introduced in April 2001 and is expected to achieve carbon savings of 6.6 mega tons by 2010. The CCL is a tax levied on energy consumption by industrial and commercial users. Taxable fuels are coal, gas, electricity, and non-transport liquefied petroleum gas (LPG). For political reasons, tax rates per kWh vary across fuel types, with electricity being taxed at almost three times the rate of the other fuels (Pearce 2006). The implicit tax per ton of carbon ranges from £16 for coal to around £31 for electricity and gas. In order to mitigate concerns about potential detrimental effects on international competitiveness, the government offers facilities in energy-intensive sectors an 80% discount on their tax liability if they enter a CCA. These agreements stipulate sector-specific targets for relative energy use to be reached by 2010 (a few sectors chose absolute targets). Compliance with interim targets is monitored every two years.

Ours is the first microeconometric study of the effects of the Climate Change Levy package. We match observations from two restricted-access panel data sets maintained by the Office for National Statistics in order to estimate the impact of the CCA on energy use, interfuel substitution, and economic activity. Our primary data source, the Annual Respondents Database, is the most comprehensive business microdata set in the UK and contains economic information for a large sample of establishments. We add data on energy use and expenditures by fuel type, which is available from the Quarterly Fuels Inquiry for a representative sample of manufacturing plants. The resulting data set is augmented with publicly available information on CCA membership and covers the years before and after the introduction of the policy until 2004.

We address the problem of self-selection of establishments into agreements in two ways. First, we include plant fixed effects to control for unobserved heterogeneity across plants. Second, we exploit the fact that eligibility for the CCA is based on pollution intensity defined by EU Directive 96/61 (‘Integrated Pollution Prevention and Control’) rather than on energy intensity. Pollution intensity can thus be used as an instrumental variable for membership in a CCA, provided that it does not directly affect energy intensity. To minimize such concerns, we choose as an instrument the number of different water pollutants emitted by a facility. This information is obtained from the European Pollution Emissions Register.

Preliminary results are available on request.