Solutions to Climate Change":
the theme of carbon taxation :
page for "Smart Solutions . . "
of economic models
evaluation of "Smart Solutions . . . "
Copenhagen Consensus on Climate
"Smart Solutions to Climate Change. Comparing costs and benefits", edited by Bjørn Lomborg 2010.
Comments on the theme of carbon taxation
(that is, placing a tax to the government on each ton of CO2 emitted)
Heavily biased analysis of carbon taxation in chapter 2
A crucial chapter is chapter 2 written by Richard Tol. This chapter is based on analyses performed with the FUND model, which, according to Tol, is so complicated that except for a few specialists, people should not try to use it or understand it (but just accept the results). Tol´s conclusion is that it does not pay off to levy any substantial tax on carbon dioxide emissions. This conclusion is subsequently accepted by all five prominent economists and adopted as a main conclusion in Lomborg´s final chapter. Tol and Lomborg have for a long time worked closely together, and it is quite possible that they have arranged by common agreement that this should be the conclusion.
However, upon closer inspection, this conclusion is biased and very weekly founded.
First, much criticism may be raised against the FUND model that is developed and used by Tol. As described on the special page on economic models, it produces net benefits for the rich countries up to a temperature rise of 2° C. In this respect, it differs from other widely used models. At the same time, damages in poor countries are given low weight, and damages in the far future are given much less weight than benefits in the near future. The cut-off date of the year 2100 means that high taxes shortly before this are counted as costs, whereas the resulting benefits after 2100 are left out. All this together means that the negative effects of climate change are downplayed very much.
Tol admits that others may criticize his choices of discount rate and lack of equity. But he replies (SSCC p. 95) that in general, people do not care much about other people in poor countries or about the future. Therefore, he thinks, it is justified to build such narrow-mindedness into the computer model. In evaluating the outputs of the model, we should therefore remember that the modeller has defined in advance that narrow-mindedness is the right approach.
Furthermore, the gradual improvements in energy efficiency and carbon intensity of energy are set exogenously and determined in advance by Richard Tol. So it is not possible in any easy way to increase the stimulation of climate-friendly technologies. Once the model run is started, you are forced to stick with fossil fuels as the main energy source to a degree which is exogenously determined by Richard Tol. You may introduce a so-called policy to reduce consumption of fossil fuels, but this policy is simply that you spend more money to reduce the carbon inensity and increase the effience of energy use. This money is then counted as a loss from the overall economy. So if you want to reduce the use of fossil fuels, you will have to pay for it - you cannot just choose an alternative strategy with more emphasis on alternative energy sources. The economic benefit from the carbon tax seems to be counted simply as an economic loss, as if the revenue disappeared out of the economy. There is no possibility to invest the revenue in developing alternative energy sources. And indeed, Tol has no favourable attitude to these energy sources. He writes for instance (p. 86): "Wind power has expanded rapidly on the back of generous subsidies, but its unpredictable nature prevents it from even [ever?] attaining a dominant position in the market". This is strange reading for me as a person who, like Lomborg, lives in a country where wind energy supplies about 20 % of our electricity demand (link).
In his chapter, Tol postulates that the social cost of carbon (SCC, i.e. the sum of future damages discounted to form a net present value per ton of carbon emitted), is about $2 per ton of carbon, or $ 0.55 per ton of CO2. He gives no documentation for this estimate. However,in the beginning of the analysis paper, Tol presents a metaanalysis of estimates of SCC, and he finds that the median values are at $14/tC or higher, probably more than $20 or $30. This is under those conditions that produce low values; under other conditions, e.g. with respect to discount rates, the values become higher. In another context, Tol has argued that a proper value of SCC is probably about $23-25/tC (see here on Lomborg-errors, comments to page 36 bottom). In an email, I have asked Tol how come that in the modeling part of the paper, he has a SCC of 2 $/tC, whereas in the first part of the paper, he presents SCC estimates that are at least one order of size larger than this. His answer was as follows:
"$2/tC is the SCC in the FUND model, and this was used for the analysis for CCC09. The meta-analysis indeed suggests a SCC that is considerably higher."
That is, the output of the model yields values for the cost of carbon that are at least one size order smaller than any likely estimates in the literature. There is no explanation for this discrepancy, and this makes the output of the model of very doubtful value.
The same crucial point has been pointed out by Roberto Roson (SSCC p. 111) and by Frank Jotzo (SSCC p. 287), who calls the value of $2 per ton an `extreme outlier in the literature´. And like me, Jotzo finds no explanation or justification for the low estimate.
So we see that the FUND model has been calibrated in such a way that costs and benefits just balance if we impose a carbon tax of $2 per ton, which is an extreme outlier relative to all other published values, which are higher. Then, of course, if Tol in his model run imposes a carbon tax that is higher than $2 per ton, the result will be that the costs of such a tax are larger than the benefits. $2 per ton is a very small tax, much lower than the present energy taxes e.g. here in Denmark, and the result is therefore that taxation of fossil fuel use must remain at an extremely low level. But the justification of this estimate does not exist. It does not fit in with other published values, and Tol has been unable (or unwilling) to provide an explanation when asked about it. In spite of this, when the five prominent economists rank all climate solutions, they accept Tol´s outlier estimates at face value without asking any critical questions. There is relevant criticism in the book, but this criticism is hidden for instance in Jotzo´s text on p. 287 where nobody notices it. So the low rank of carbon taxation on p. 381-382 in SSCC is based on completely unfounded or undocumented estimates which divert from even the low range of other estimates by a factor of about 10.
Other issues concerning carbon taxation
The main chapter on technological innovation (chapter 7 by Galiana and Green) has a carbon tax as a central theme. Their main thesis is that a `price signal´ from higher prices on fossil fuels will not be sufficient as an incentive to boost research and development in alternative energy sources, unless the price is set very much above the social cost of carbon. Instead, such research and development should be organized through some type of funding mediated by the government, and the funds should originate from a moderate carbon tax. They suggest a tax of $ 5 per ton of CO2 (note that this is $ 13.8 per ton of C, or seven times the tax suggested by Richard Tol as corresponding to his extremely low estimate of the social cost of carbon!). The tax should increase gradually over time in a pre-described way (doubling every ten years). The revenue should then go to the funds for research and development. Only in this way do we have a good chance to induce major technological breakthroughs.
This idea is commented upon in perspective paper 7.1 by V. Bosetti. She agrees that research and development has to be an essential part of any climate policy, and that this has to be coupled with a carbon policy, i.e. carbon taxation. But contrary to Galiana and Green, she says that carbon taxation does give an important price signal that stimulates technological innovation. The link between environmental policy and induced innovation has been found in a large number of studies which is evident from extensive studies and literature reviews. Research and development alone can maximally reduce cumulative CO2 emissions by 13 - 16 % relative to a baseline scenario, whereas even a mild stabilization target (550 ppm CO2) would require that emissions be halved. To attain that, a strong carbon price signal is indispensable.
David Popp, too, has this view on research and development subsidies as a climate policy tool. He writes (SSCC p. 373): "Compared to a combined policy using both optimal carbon taxes and research and development subsidies, a policy using only the optimal research and development subsidy attains just 11 % of the welfare gains of the combined policy. In contrast, a policy using only the carbon tax achieves 95 % of the welfare gains of the combined policy."
When the five prominent economists made their overall conclusion, they did not engage in any critical discussions of the above issues. They accepted the conclusions from Richard Tol´s chapter 2 at face value.
Vernon L. Smith writes that as `clearly demonstrated in chapter 2 and Perspective papers 2.1 and 2.2´, carbon taxation inflicts a cost in sacrificed human betterment and poverty reduction that would be prohibitive. Evidently, he has not understood - or not wanted to understand - the bias of chapter 2, and he is not aware that perspective paper 2.2 is a harsh criticism of that chapter. He does not think that a carbon tax is necessary to stimulate research and development and he postulates that rising energy prices by themselves will induce innovations that will increase energy efficiency, neglecting that several authors argue strongly that this will be vastly insufficient.
Finn E. Kydland writes that he likes the energy research and development solution very much, and he likes the use of revenue from a carbon tax. He therefore accepts the carbon tax of $ 0.5 per ton of CO2, which is the highest tax rate that is acceptable according to Richard Tol. He fails to see the discrepancy between the results of Tol and the results of others, and does not comment on the fact that Tol´s low tax rate is seven times later than what is proposed by Galiana and Green. On top of this he fails to notice that several others (Bosetti and Popp) argue that to attain our goals, "a strong carbon price signal is indispensable", i.e. a carbon tax even higher than proposed by Galiana and Green.
In spite of the moderate support to a carbon tax from Finn E. Kydland, the overall ranking puts all carbon taxes dealt with by Richard Tol at the very bottom of the ranking list. This has led to official protests from two of the experts contributing to the book, Claudia Kemfert and Roger Pielke. They have raised a fairly harsh criticism of the conference on this web site. They write among many other things: "The two of us disagree about how best to price carbon . . . However, we strongly agree that putting a price on carbon is a necessary component in any portfolio of policies designed to decarbonize the global economy."
Thus, even though several experts (Bosetti, Popp, Pielke, Kemfert) state that considerable carbon taxation is ´indispensible´ or `necessary´ , and even though they refer to extensive literature confirming this view, Lomborg manages to have the five prominent economists ignore this completely. As a result, these economists accept to designate carbon taxation of the type treated by Tol as `very poor solutions´. In that way, Lomborg succeds to remove carbon taxation from the portfolio of useful options.