Pekka Pirilä on Energy

My previous posts have discussed the difficulty of identifying proper goals. The discussion applies rather to concrete act like investment decisions than to policy decisions like carbon tax, cap & trade or feed-in tariffs for renewable electricity. These policy decisions have the apparent advantage that they are supposed to lead private actors to make the needed concrete decisions. It’s certainly true that public authorities are not capable of making well all the numerous decisions that are needed for significant results, but accepting this fact does not solve the real problems.

The world is facing two related developments: running out of high quality resources of oil and gas, and adding CO2 to the atmosphere. The most straightforward answer to the first problem, switching gradually to coal and perhaps shale oil, is at the same time the worst path concerning the second problem. The oil crisis of 1970′s led to a rapid increase in funding of alternative energy technologies. It’s certainly possible to find some signs of success from that research, but in my general judgment it has produced pitifully little. We have learned that it’s very difficult to develop solutions that have a major effect on the scale of the overall energy use. Some of the most promising alternatives, most notably new technological alternatives of nuclear energy production, have been out of favor, but mostly the lack of success reflects rather the difficulty of the task than lack of trying.

The economics of wind power has improved slowly, and has led to competiveness under most favorable conditions. In a worldwide assessment the wind power cannot, however, produce enough to solve more than a small fraction of the energy problems. Solar electricity is significantly behind. Again the operating environment has a major effect, and acceptable economics could perhaps be reached rather soon in areas of best sunshine, when that is combined with a load peaking during sunny hours and a lack of most economic alternatives. Expanding from that to acceptable costs on a wider basis appears still to be a very distant goal. Overly generous feed-in tariffs have resulted in a large solar capacity in Germany, but at a cost that is far too high – by a factor exceeding five. This is in my judgment the most serious existing example of excessive support for renewable energy. Some research done in Germany agrees (see, e.g., Frondel, M., N. Ritter, C. M. Schmidt and C. Vance (2010), Economic Impacts from the Promotion of Renewable Energy Technologies: The German Experience. Energy Policy 38 : 4048-4056. and Ruhr Economic Papers #156), but the opinion is certainly not shared by everybody. While the problem of wind and solar power is mainly in economic efficiency, the net environmental value is seriously contested for biofuels. Here it’s questionable, whether the support has given any positive results at all, or whether the money spent is used to deteriorate also the state of the environment.

The above examples are related to an attempt to use economic incentives to produce more than the state of knowledge allows. A Pigovian tax that corresponds to actual externalities is economically justified, but an economic incentive scaled to produce strong effects disregarding the economic efficiency may be much more distortive than doing nothing at all. This seems to be the case for numerous European policy choices. Support decisions should always have well defined goals worth the cost of the support. Such goals may include direct effect on energy balance or speeding up of the development of technologies that have good potential of reaching a positive contribution in the future, but many of the policies are not justified by either argument – or by their combination.

Pointing out failures of present policies is not enough, but neither is it useful to continue these policies in spite of their failures. Presently I do not see better alternatives than emphasizing research on a wide range of alternatives, and also research on policies for efficient advancement of technological change and on economic incentives of environmental and climate policies. We know far too little on, how different incentives actually influence decision making. The lack of proper knowledge leads at least in Europe continuously to choosing policies that seem nice and make us feel that we have done something rather than policies that influence efficiently the future development leading to the results envisioned and avoiding unnecessary and excessive costs or collateral damage to environment and social structures.

The justification of climate policies is in the mitigation that they provide against the perceived threats of global warming. Making wise decisions is extremely difficult, because both the severity of the threat and the value of the proposed policies are highly uncertain. An accurate uncontroversial quantitative cost-benefit analysis is certainly beyond our capabilities. Still, every rational policy choice implies that we have judged one alternative to be superior to the others. We have expressed a view on the ordering of the alternatives, if not on the quantitative difference between them. If we wish to do that as well as possible and using as much of the available information as practical, we need some way of comparing costs and benefits of the policy to its alternatives. The issues are important and we cannot afford making seriously suboptimal decisions.

Let’s look at the ways our decisions may affect the human well-being in near and distant future, here and in other parts of the world. If we accept the views of the Stern Review and of those with similar thoughts, the future must be considered over several hundreds of years as their conclusions are dominated by those periods with little weight given to the next century. But is that possible? (It’s not.) Is that necessary? (Perhaps.)

The Stern Review presents a quantitative cost-benefit analysis (CBA) comparing the discounted damages from climate change to the costs of mitigation. The costs of damages are very high, because they are summed over a very long period. Having a social discount rate of 0.1% corresponds for a typical time span of 1000 years. Changing that to 0.2% would halve the time span and changing it to 0.05% would double it. Choosing the value as 0.1% is totally arbitrary; there is certainly no good argument to support that value rather than any other between 0% and 0.5%. How sensitive the final cost estimates are to this value depends on the other choices done in the calculation. The Stern Review happened (sic.) to make those choices in the way that the methods used to handle uncertainties became also important. Emphasizing the worst outcomes in the spirit of risk aversion (see my previous post) and making the uncertainties large with a possibility for very serious consequences turned out to be important in their analysis. Again this contained several parameters, which cannot be determined with any accuracy from our present knowledge. Other choices made in the calculation have received less emphasis, but they are equally important in determining the final results.

The conclusion on the approach of the Stern Review is that it’s not valid, if it’s premises are accepted. With these premises, the result is really undefined as it’s determined by arbitrary choices of the analysis. The analysis assumes that we can calculate the consequences of near term decisions to very distant future. How impossible that is can be envisioned by thinking, how decisions of the 19th or 18th century influence the present. Some decisions of those periods may have a great influence, but, how can we decide, what would be the alternative counterfactual history, and how much worse or better it’s results would have been. Great inventions and scientific discoveries may have the strongest influence, but were they finally dependent on individual decisions, or did these decisions have only minor influences on their timing? The decisions that led to World Wars might have been more influential, but even for them the counterfactual history is impossible to define.

Thinking on the past reveals that most individual acts have a temporary influence. Even innovations are mostly based on the actual state of knowledge. The brightest individuals speed up the process, and so do also policies that support science and innovative development. Almost every concrete act has a finite time span of influence, but most discoveries would have occurred irrespectively of them. Over years people make new decisions, new innovations, and new investments, which reduce the significance of earlier acts and make the old investments obsolete. For the constructs of the human societies we must realize that their influence vanes. Many decisions have influence on short periods from days to a few years, some extend their influence to decades. For many present issues a history of hundreds or thousands or years can be identified, but even in these cases the influence of any single original decision had hardly significant influence for long, when the comparison is done to potential counterfactual histories. It’s mostly a fallacy to say that we are here with the present state of the world, because of some specific decision of distant history, and even in those very few cases, we don’t know, how the alternative would be different.

It may be argued that influencing the global resources and the global environment is different. There is some truth in that, as fossil fuels can be burned only once and using them soon means less fossil fuels for the later future. Similarly carbon dioxide added to the atmosphere today disappears from there slowly, but the case is not quite the same as with fossil fuels, because the carbon cycle will continue forever, and can thus also be influenced significantly forever. The dynamics of the carbon cycle and the climate are complicated. They lead to long delays both in the warming phase and in the cooling phase that must follow, if we first use the fossil fuels running out of them rapidly. After all we have already used most of the easiest oil resources and a large fraction of the best gas resources. The emissions from oil and gas are bound to turn to decline rather soon with a maximum not far above the present level. The future of coal use is more difficult to forecast, but the practical limits are strong even for coal. What will be ultimately decided about the extent of the use of coal will be important for the climate, but again the important issue is the ultimate future, not immediate acts.

We see that the human influence on climate will in the long run be determined mainly by how coal will be used (the oil shale may potentially add significantly, some other fuels like peat are comparable, but their resources are much smaller). The use of coal is driven by the social value of energy and by the availability of alternative socially acceptable paths. The immediately available means of reducing the social value of utilization of coal are limited, not by the number of alternatives, but by their quantitative potential.

The main alternatives of continued and even growing use of coal can be divided to

• alternative energy production technologies (renewable and nuclear, carbon sequestration and storage is also a comparable solution),
• more efficient technologies for energy use, and
• changes in the consumption patterns.

The most commonly presented goals requiring a limitation of the CO2 concentration in the atmosphere to 450 ppm or less are very difficult to reach, even 550 ppm may fall to the same category. We really don’t know, what kind of combination of the main approaches could be achievable. The solution must offer space for the social and economic development of emerging economies like China and India. It cannot present politically unrealistic requirements for the industrial economies either. The present optimism on possibilities expressed by many European states and organizations is not based on solid arguments.

There is an alternative way of looking at the future, which is not as directly contradicted by the impossibility of knowing much about the distant future, and which recognizes better the importance of future decisions by the future decision makers based on the better knowledge of the time. I have emphasized that in my previous posts. It is expressed by the Brundtland Commission as the Principle of Sustainable Development, it is discussed by Partha Dasgupta in his research and books, and it has been formulated as a mathematical theory of Dynamic Programming by Richard Bellman and others. One principal observation behind all these formulations is that our present decisions have to types of consequences:

• they influence the immediate well-being and
• they influence the state of the world at a nearby future point of time.

This future state of the world is the basis for decisions done at that future time, and these decisions influence in turn following near term well-being and the state of the world at a moment a bit later in the future. This approach does not mean that the distant future would be irrelevant, but it tells that the influence comes through intermediate steps and that the decision makers active at each step modify the influence.

Is this alternative approach useful in practice? There is plentiful of evidence on the value of Dynamic Programming as a tool for analyzing more limited problems that are influenced by successive decision making. The most familiar problem for me concerns electricity production in a system with a lot of hydropower and reservoirs, as we have in the Nordic power system. The power system is mathematically easy to describe and the results are both excellent and valuable in practice. The worldwide human well-being over generations is certainly not easy to describe mathematically, but the approach seems to me be far superior to alternatives – at least, if we are willing to give much weight to future generations in the analysis. Using a high social discount rate might allow for the more straightforward approach of William Nordhaus, but choices approaching those of the Stern Review makes these methods practically worthless.

I have argued that the future decisions reduce the weight of later times similarly with discounting, but the mechanism is different from the combination expressed in the Ramsey rule discussed in my previous post. I believe that the role of later decisions and adaptation as a factor influencing discount rate is actually taken into account in a simplistic way by many economists, although I cannot present direct evidence on that.  Later decisions do not affect all consequences equally, and taking all likely forms of adaptation properly into account remains a major problem in the analysis. This may have some connection with the idea of a declining discount rate, as the faster adaptation affects shorter periods and the slow adaptation longer ones. The willingness to express the discount rate by the Ramsey rule appears to be limited largely to environmental and development economists, while other economists leave the determination of the discount rate commonly for the markets, which are likely to foresee the importance of all processes that make an investment obsolete. Thus Nordhaus might actually include these effects in his DICE model calculations although he has tried to explain his chosen rate by the Ramsey rule (and had difficulties in that attempt).

The approach that I support here will at the minimum offer a framework for discussing the multitude of issues involved and through that also some basis for deciding, which issues should be taken explicitly into account at the minimum and which might be left out without distorting the conclusions severely. Dasgupta’s book Human Well-Being and the Natural Environment (Oxford University Press, 2001) is perhaps the best starting point, although I have many reservations on the details of his conclusions.