In all countries in which CO2 taxes have been introduced, some sectors
are exempt, or the tax is differentiated across sectors (see, e.g., ECON, 1997).
Typically, households pay the full tax rate, whereas export-oriented industries
pay either nothing or a symbolic rate3.
Very few countries have actually implemented a CO2 tax, and (unsurprisingly)
tax exemptions are more systematically analyzed in these countries, such as
the Scandinavian countries. Concerns about the sectoral implications of revenue-raising
policies have led to four types of responses being studied:
Lessons from the few modelling exercises suggest that the efficiency cost for the whole economy of offsetting the sectoral impacts of carbon taxes through tax exemptions are very high. Böhringer and Rutherford (1997) show for Germany that exemptions to energy- and export-intensive industries increase the costs of meeting a 30% CO2 reduction target by more than 20%. Jensen (1998) has similar findings for Denmark with respect to a unilateral reduction of CO2 emissions by 20% (Jensen, 1998). To exempt six production sectors that emit 15% of Denmarks total emissions implies significantly greater welfare costs (equivalent variation) than full taxation to meet the same abatement target. Namely, welfare loss of 1.9% and a carbon tax on the non-exempted sectors of US$70/tCO2, against a welfare loss of 1.2% and a carbon price of US$40/tCO2 in the no-exemption case (uniform taxes). A similar result is found in Hill (1999) for Sweden: the welfare costs of using exemptions are more than 2.5 times higher than in the uniform carbon tax case for a 10% emission decrease. The high costs of tax exemption are also confirmed by a US study (Babiker et al., 2000).
Tax differentiation is studied in a CGE model for Sweden in Bergman (1995), who compares its effect with a uniform tax for given emission targets. The tax rate applicable to the industrial firms is set to one-quarter of the tax rate for non-industrial firms and households. The GDP loss increases slightly compared to the uniform tax, but it is still quite small. However, the purchasing power of the aggregated incomes of labour and capital is significantly reduced. Consequently, tax differentiation does not seem to have as much of an adverse effect as full tax exemption. The reason is that all sectors pay a carbon tax when taxes are differentiated, while this is not the case for tax exemptions. Thus, the burden on sectors that pay the highest carbon tax is not that large, and hence results in lower welfare losses.
Böhringer and Rutherford (1997) as well as Hill (1999) envisage labour subsidies used to keep a given employment target. They conclude that compared to tax exemptions for energy- and export-intensive industries a uniform carbon tax cum wage subsidy achieves an identical level of national emission reduction and employment at a fraction of the costs.
A second option is a special case of voluntary agreements. In most of the literature,
voluntary agreements result from negotiations on emission levels between public
authorities and firms adversely impacted by environment policies. Carraro and
Galeotti (1995) examined another form of voluntary agreement for European countries:
firms receive financial benefits if they have engaged in environmental research
and development (R&D) spending. This option is justified because economic
tools may be inefficient in reaching the optimal R&D level, even in a pure
and perfect market competition (Laffont and Tirole, 1993). According to this
study, a strong double dividend could occur in all European countries except
Belgium and the UK, even if the impact on employment is weak. One of the reasons
for this double dividend is the technical progress induced by this policy.
Other reports in this collection