Annex B countries are currently considering the creation of a market for GHG emissions on the basis of grandfathered quotas or of quotas delivered in function of voluntary agreements of sectors to given emissions targets (see Chapter 6). This option does not generate revenue, but (contrary to tax exemptions) implies participation of the carbon-intensive industry to climate policy and does not transfer the full burden to households and the rest of industry. However, the welfare impacts are systematically found to be less favourable than under a full revenue-neutral taxation. Jensen (1998) found a welfare loss of 1.4% in Denmark and a permit price of US$110/tCO2, while a uniform tax to meet the same 20% target is only US$40 and the resultant welfare loss is 1.2%. Bye and Nyborg (1999) investigated the effects on welfare (total discounted utility) of both uniform taxes and tradable permits issued freely compared to the current system of tax exemptions. To keep total tax revenues unchanged for the government, the payroll tax is adjusted accordingly. They found that a permit system gives a welfare loss of 0.03% compared to the current system, while with uniform taxes there is a gain of similar size. The main reason is that payroll taxes must increase to maintain the budget balance when carbon taxes are not used. There are similar findings for the USA. Parry et al. (1999) show that the net economic impact (after accounting for environmental benefits but not without climate benefits) of carbon abatement is positive when permits are auctioned, but switches to negative when permits are grandfathered.
Other allocation rules have been tested, but do not improve the result compared with grandfathered permits. For Denmark, Jensen and Rasmussen (1998) examined the aggregate welfare loss (equivalent variation) of an emission target of 80% of 1990 levels from 1999 to 2040; they found 0.1% with a permit auction, 2.0% for grandfathered permits, and 2.1% when the permits are given to firms in the proportion of market shares and sectoral emissions, similar to an output subsidy.
Such results are obtained because, in the case of free delivered permits, the interactions with the tax system occur without the compensating effect of tax-revenue recycled, as in the cases of environmental taxes and auctioned permits. Studies by Parry (1997), Goulder et al. (1997), Parry et al. (1999), and Goulder et al. (1999) show that the costs of quotas or marketable permits are higher if there are prior taxes on the production factors concerned than in if there are no such taxes. Quotas or permits tend indeed to raise the costs of production and the prices of output. This reduces the real return to labour and capital, and thereby exacerbates prior distortions in relevant markets and decrease the overall efficiency of the economy.
Bovenberg and Goulder (2001) found that avoiding adverse impacts on the profits and equity values in fossil fuel industries involves a relatively small efficiency cost for the economy. This arises because CO2 abatement policies have the potential to generate revenues that are very large relative to the potential loss of profit for these industries. By enabling firms to retain only a very small fraction of these potential revenues, the government can protect the firms profit and equity values. Thus, the government needs to grandfather only a small percentage of CO2 emissions permits or, similarly, must exempt only a small fraction of emissions from the base of a carbon tax. This policy involves a small sacrifice of potential government revenue. Such revenue has an efficiency value because it can finance cuts in pre-existing distortionary taxes. These authors also found a very large difference between preserving firms profits and preserving their tax payments. Offsetting producers carbon tax payments on a dollar-for-dollar basis (through cuts in corporate tax rates, for example) substantially overcompensates firms, raising their profit and equity values significantly relative to the situation prior to the environmental regulation. This reflects that producers shift onto consumers most of the burden from a carbon tax. The efficiency costs of such policies are far greater than the costs of policies that do not overcompensate firms.
The costs of meeting the Kyoto targets are very sensitive to the type of recycling used for the revenue of carbon taxes or auctioned permits. In general, however, modelling results show that the sum of the positive revenue-recycling effect and the negative tax-interaction effect of a carbon tax or auctioned emission permits is roughly zero. Thus, in some analyses the sum is positive, while in others it is negative. In economies with an especially distortive tax system (as in several European analyses), the sum may be positive and hence confirm the strong double-dividend hypothesis. In economies with fewer distortions, such as in various models of the US economy, the sum is negative. Another conclusion is that even with no strong double-dividend effect, a country fares considerably better with a revenue-recycling policy than with one that is not revenue recycling, like grandfathered quotas. Analyses of the US economy found that revenue recycling reduces the cost of regulation by about 30%50% for a certain range of targets, while European analyses report cost savings that are even higher than 100%.
However, at this stage insufficient evidence exists either to confirm or to
substantiate these results in the context of developing countries. Studies to
date have concentrated on developed countries and, while these studies are comprehensive
and rigorous, their conclusions may not be directly transferable. It can be
argued that, in developing countries, direct welfare losses typically associated
with specific factor taxes (such as a carbon tax) may have fewer opportunities
for mitigation within the fiscal-reform policy envelope. Nevertheless, the complex
linkages between formal and informal sectors of the economy may show this intuition
to be incorrect; the only existing study for China reviewed here suggests that
this may be the case but further research is needed to confirm this more generally.
Other reports in this collection