Posing the Energy Debate (I)

1. - Back to the 70’s

Those of us who spent the 90’s reading science fiction from the 70’s are used to apocalyptical energy predictions. Though there were others, the master piece of the literature on resource shortage was “Stand on Zanzibar”, which described a global panorama of energy shortage and overpopulation in a world where the United States had become a second-hand potency, plagued with economic decadence and armed racial conflict.

During those days, while King Fahd was getting ready to govern the believers in New York’s “Studio 54” and Jimmy Carter would go back to the promise of “peace for our time” with the Iranians, the MIT published the famous report on “The limits of growth” which asserted that the great banquet of natural resources was coming to an end and that in our Cobb-Douglas function of production, apart from labor and capital, the “land” factor had to be included. A factor, that since the old days of David Ricardo nobody had taken seriously it as a serious growth limiter.

Well, the 90’s, were the NASDAQ and “New Economy” age, and oil fell to less than 10 dollars the barrel. But let’s not be driven by the nostalgia: our best times are yet to come.

2. – Energy cojnuncture: the Asian growth

From the economical standpoint, the Asian take-off has been the most important event of the last thirty years. No one must believe that the present day economical earthquake coming from Asia is an isolated or recent phenomenon. As in every growth process, multiple causes can be identified, but if I had to choose a few, I’d say it has to do with the improvement in the agricultural productivity that followed the Green Revolution, the strong accumulation of human capital in the region resulting from the domestic culture of saving, and the political stability the area has enjoyed since the Sino-American (and anti-Soviet) pact that followed Kissinger’s visit to China, and which started to impact the economy since Deng Xiaoping’s presidency.

From the mid 90’s, the signs of Asian growth have become evident for the rest of the world, and China is becoming a regional economical power, able to offer an enormous amount of cheap labor. The consequences of this massive work force entering the world labor market are evident: salaries have gotten depreciated in relation to capital and resources (and non-qualified labor in relation to qualified one: read K. Rogoff’s article). In welfare terms, however, the net positive effect cannot be denied: millions of Asian workers have moved from quasi-Malthusian misery to mere capitalist exploitation. That is, they have greatly improved their standard of living. As usual, the migratory flows indicate the welfare gradient, and the greatest rural exodus in history has just started in China, and to a lesser degree, in India.

In energy terms, China’s industry is extremely inefficient, which is normal, as they can compensate for the energy waste with their cheap labor, and still be competitive. China’s growth has therefore been very intensive on energy consumption (oil barrels/GDP). However, as for the moment, the increase on oil prices has had little effect on the western economy. Inflation has not shot up, and 2005 has been the greatest world growth year ever recorded.

China is, apart from an oil consumer, a producer of cheap goods and services, so its net effect on world economy is being clearly positive. It is true that gasoline is more expensive, but you can buy a Wal-Mart DVD in 30 dollars.

Why has the world economy reacted differently in 2002-2007 that in 1973 to the increase in oil prices? Because there was no oil crises in 2002-2007. In 1973 there was a fall in production while the 2002-2007 price increase was exclusively caused by the strong demand. It would be enough to cross a pair of curves to understand that, as oil supply demand is ruled by economic growth, the price increase can only slow down the growth: it cannot invert it. It can’t happen that an increase in the added demand produces a reduction of the GDP (by definition!). In fact, it is worth to mention that one positive effect from the increase in oil prices is an increase in the pressure towards efficiency, which means, for China, that certain wasteful production techniques (or enterprises) will have to leave the market, and for the United States, that the popularity of the SUV is going down.

3. - Oil peak: land vs. capital

How about long term? There is no agreement as to oil exhaustion time. There are two schools of thought radically opposed: the school of “capital”, and the school of "land". The “school of capital” asserts that the world supply of oil is perfectly elastic, and responds to the amount of resources invested in its production. To put it in another way, the elasticity of the production to the price is very high. In favor of this point of view, one can mention the abundance of non-conventional oil, and the fact that wide areas of central Asia have not been conveniently exploited. Not the oceans either. However, the extraction techniques for this non-conventional oil go from the available at 18$ a barrel (and so the Canadian bituminous oil starts to appear in the list of proven reserves) to science fiction. Enhanced recovery techniques are also a clear hope for increased oil production based on investment.

The “school of land” thinks that the main determinant in the energy production at world level is the physical shortage of resources. The modern formulation of the school of the physical shortage is owed to Hubbert, a geophysicist that in 1965 predicted that the production of oil in the United States would peak around 1970. The prediction was a success. His methodology has been tried at world level ever since, with a series of resounding failures.

Why was Hubbert able to predict the U.S. peak and his successors failed to predict the world peak? Well, as in every simulation exercise, the value of the results depends on the verification of the initial hypothesis. The first verified hypothesis for the U.S. was that of a wide geophysical exploration: by 1956, the North American territory had been explored almost completely. However, there are still relatively little explored areas around the world (vg. in Central Asia). On the other hand, Hubbert’s model supposed a constant oil demand, that is, that the incentive mechanism to save oil, and the speculative operations that guarantee the smoothness of energy consumption would not be activated.


It has happened like this because the United States are only a fraction of world production, and as consequence, the shortage of American oil, has hardly affected world oil price. The United States are not an isolated system in oil consumption but “a small open economy”.

To put it another way, the use of Hubbert’s methodology to analyze the world shortage oil of supply is contradictory, because one of the sunk hypothesis in the Hubbert’s curve is precisely the (world) abundance of oil.

Many of Hubbert followers, of course, cite the inelasticity of short-term oil supply demand as a proof of the avidity of world economy for energy. The elasticity of short and long-terms world supply demand is radically different. Short-term elasticity is determined by the economy fixed capital structure. But if prices are constantly high, it affects the decisions on capital renovation. When energy prices are high, consumers and entrepreneurs invest on efficient machinery and less energy intense techniques of production. This causes the elasticity of long-term energy consumption to be much higher than the short-term one. (Only compare the energy consumption of Japanese and European automobiles to Americans’: the difference is a result of the high taxes on gasoline and diesel in Europe and Japan).

In general, peak estimations belong to an area of science that is highly emotional and subjected to strong interests. It is enough just to watch the range of predictions (see graph) to realize that nobody is making science out of this topic. From all the prediction sets showed there, the most visually believable (and the most consistent with the trends in prices), is that of BP 2003, which asserts that an oil production peak can be predicted between 2015 and 2020. The prediction is alarming but not catastrophic.

4. - From oil peak to combined fossil peak

However, fossil energy, apart from oil, also includes natural gas and coal. Let us concentrate on natural gas. Natural gas has been the surprise of the last twenty years. For a long time, it was considered a byproduct of oil production, but for the last twenty years, there have been multiple discoveries of autonomous gas wells. Today, the proven reserves of Russia are enormous (honoring Putin’s phrase, “Russia holds the energy resources of the Christian world”) and the exploitation of Siberian natural gas is far from the peak of reserves discovery. Farther is its production peak. No one knows how much gas there is in Russia, let alone the rest of the world. The hypothesis of “good geophysical exploration” that underlies the Hubbert’s curve is less true in the case of natural gas.

In absolute terms, to set out a gas curve that reaches its peak before 2035 seems somehow excessive. Not only geologists think that the reserves are large; in addition, markets and enterprises have betted strong on the gas economy: it is reasonable to believe they know what they are doing. In any case, the estimations of Hubbert’s curve are still at the onset. What is true is that as we move towards an oil-gas economy, Hubbert’s curves on oil are each time more incomplete.

It is necessary to construct an oil-gas Hubbert’s curve to answer the matter of world fossil fuel peak. The first problem encountered at the time of constructing the combined curve is to choose the units that must appear in the Y axis. It was clear that in the case of oil, the number of barrels was a reasonable unit, but in the case of the combined peak, the most natural units would be of energy (Giga Jules or BTU’s). In all instances, to construct that combined curve it is essential to have carefully studied the degree of gas-oil substitutability. For example, to the technology of gasoline-electric hybrid vehicles, you have to add that of the natural gas cars, ships and trains. That degree of substitutability depends strongly on the relative price of both resources.

Without a long-term model of substitutability between gas and oil, no combined curve is possible. Once that model is available, that combined curve will be relevant to measure the world energy problem.

Of course, the entire world energy problem strongly depends on the belief of global warming, because there is a practically unanimous agreement that coal reserves are almost unlimited. Coal can be used not only for electricity generation, but also for oil synthesis through coal liquefaction. Synthetic oil, that was massively used by the Germans in II World War, and South Africans in the apartheid-embargo years, is still not widely used simply because oil is cheaper (it is obvious: in the end, it is a similar product that exist directly in nature). Counting on coal, there’s no energy crisis (though maybe it’s possible to speak about the end of cheap energy), as the gas-oil-coal combined peak won’t be reached before 2100.

5. - The nuclear option: on the installation costs, operation and dismantling

Let us think about the worst scenarios: the (gas-oil) combined peak is reached in 2030, and from then on, the fall down begins. Let’s also suppose that by then, there is enough consensus on the Kyoto Agreement, and that the coal option has to be discarded.

There will still be nuclear energy as primary source of production, and hydrogen as vector for transportation (well, even this is doubtful: there are already very good prototypes of electric automobile. An electro-nuclear economy is more likely than a nuclear-hydrogen one). Let’s not discard that the reserves of bituminous oil that are energetically unviable, can be exploited through the use of nuclear energy (nuclear-oil), or that synthetic oil is extracted from coal with the nuclear energy feeding the operations.

Let us start with the case of the United States. Using these 2005 data from BP, the reader can one can see that all the oil and gas consumed by the U.S. is equivalent to some 1.840 millions of tons of oil, while the 100 American nuclear power plants produce the equivalent of 186 millions of tons of oil. Therefore, assuming perfect substitutability, it turns out that to compensate for the disappearance of gas and oil the U.S would need 1000 nuclear power plants. Let’s think of 1.200 to cover the growing demand and imperfect substitution. For a billion dollars a plant, (that would be the price if the current regulatory sabotage were abandoned) it would be 1.200 billion. Can the American economy cope with the 1.200 billion cost for the nuclear power plants in twenty-five years? To answer such question, one is tempted to look straight into what percentage of the GDP those 1.200 billion stand for (10 % of the GDP in 2004).

However, we have to admit that the production of power plants does not have the same characteristics and inputs as cotton production. That’s why I am going to focus on the value-added produced by a group of industries with costs and inputs structure similar to those of power plant production.

In the BEA web page, one can find the breakdown of the U.S. GDP in the value added by industry accounts . The sum of the following list of industry accounts got to a value-added of 1, 414 billion in 2004: metals, metal products, machinery, electronics, motor vehicles, chemicals, transportation and constructions. The group of listed industries consumes inputs similar to those of the power plants construction. An economy able to generate a value-added of 1.414 billions of dollars per year in the mentioned sectors (12% of GDP) can generate value-added of 1.200 billion dollars in the construction of power plants in a 25 years span. For instance, in twenty-five years, it would be enough with the equivalent of 1200/ (1414*25) = 3.4% from the production of those industries that are intensive on energy and physical resources (and a hardly noticeable fraction of the total GDP).

At world level, the number of power plants necessary to forget about gas and oil, assuming the correct energy vectors, would be (given that the gas and oil consumed are equivalent to 6.193 millions of tons of oil, and given that the 441 power plants in operation generate the energy equivalent to 627 millions of oil tons) 4.361 power plants, which I am going to round up to 6.000. With a billion per power plant, we would be talking about 6.000 billions. The world output in 2005 was 44.454 billions of dollars, and using the same proportion as in the last case (these industries were 12% of US GDP), I am going to suppose that 12% from World output belongs to these industries of high energy consumption/high capital intensity.

So we have that the world produces heavy industry/high energy content goods for 5.334 billions of dollars a year. Then for the next 25 years, it would be enough to consume 6000/(5334*25) =4,5% output from the industries that are similar to the production of power plants, to maintain the current energy production to nuclearize all oil+gas consumption. To put it in another way, it would be enough to consume (equivalently) a 4,5% less, or produce a 4,5% more (ceteris paribus) of metals, metal products, machinery, electronics, motor vehicles, chemicals, transportation and construction for the next 25 years, to build the necessary nuclear power plants.

In the above calculations, I have stated many catastrophic hypotheses (for instance: I have ignored all the gas and oil on the decreasing side of the possible Hubbert curve, and I have assumed that new power plants will be as productive as the first ones, and not more). What the above approach suggests is that to demonstrate that the nuclear option is not viable, one has to prove that:

Price (5000 nuclear)>>5000* price (1 nuclear) [I]

The only way to prove that is by finding an input whose shortage makes the construction or operation of the plants unviable. But the common inputs used by the sectors already mentioned are not limiting inputs because those sectors use them about as much (per dollar of value-added) as the power plants. That is: labor, energy and construction materials, which are inputs of these sectors, can’t be a problem for the nuclear construction. If they were, they had to be for the listed sectors that are also very intensive on such inputs.

Only a limited input, characteristic to the nuclear industry, could make the [I] inequality be verified; that input limiting nuclearization can be only the fissile material: that is, uranium. Deffeyes & MacGregor paper of 1980 (Scientific American, January 1.980) proves that there is enough uranium, in high enough concentrations to supply every conceivable fleet of nuclear reactors for at least one century. On top of that, if breeder reactors, instead of (presently used) light water reactors were used, the amount of available uranium would be able to feed human needs for at least five centuries. I will devote an extra post to this.

Finally, on the dismantling costs, did the builders of Stonehenge or the Burgos Cathedral ever worry about dismantling costs? Once the uranium has been eliminated and the plant turned off, the contention building itself, potentially full of concrete, is a perfect coffin for the reactor. If power plants aren’t dangerous during operation, they have to be much less after they are turned off.

6. - The nuclear option: on waste treatment

Therefore, all is left is to deal with is the environmental side of nuclear energy. Those who think that 500 power plants could generate waste and radioactivity able to poison the world, forget that we have already made experiments on the limits of the biosphere’s radioactive tolerance. From 1945 till late in the 90’s, more than 2000 nuclear tests (528 atmospheric ones) have been made— some of them of several gigatons. I won’t say these tests were harmless, but if the equivalent of a small nuclear war has not had severe effects on the planet, much less can be expected from the consequences of some solid radioactive waste correctly stored.

In terms of space, nuclear waste is extremely compact, though some of it has an average life of 20.000 years. For example, all the waste produced by the 100 American power plants since its construction occupy the equivalent of football pitch surface by five meters high. It is by all measures, a small amount, and producing a similar amount every ten years, can’t be an unsurpassable problem. Let’s not forget that the same mountains from Nevada that will shelter the American nuclear waste were already able to contain the underground nuclear tests, which were much more traumatic events.

Even France that produces 78% of its electricity in such a way can deal with the waste without suffering an ecological disaster. Nuclear waste, as opposed to atmospheric pollution, does not have international externalities. So, if France can treat its waste, the rest of the world can imitate the French strategy without major harm. There is no ‘crowding’ effect in the treatment of nuclear waste.

7. - Conclusions

I think that the energy concerns are justified. The thousands of millions of dollars that will have to be invested on plants for the synthesis of crude oil, new equipments based on gas and electricity, and nuclear power plants, it’s going to be money we won’t be able to spend on vacationing, bigger houses and new versions of PSP. It means more work and less welfare. Those of us who assert that this economic system is efficient, cannot at the same time sustain that the disappearance of a valued resource does not have repercussions on welfare.

In the same way, those who sell the idea of crude oil exhaustion cannot assert, at the same time, that nuclear energy is not profitable on the grounds of capital costs. And the most extremist “oil peakers” who predict an economical re-adjustment with more than 4 000 millions dead, cannot say, at the same time, that coal puts us into the dangers of a climate change: anyway, I think we all prefer an uncertain climate change to the certainty of universal famine.

The western world no longer lives out of spending energy: almost all the increase on energy demand of the next forty years is going to happen in Asia and maybe Latin America. Those who besides helping themselves, intend to help Third World countries should be especially enthusiastic towards any energy program (as nuclearization) that liberates the largest quantity of energy resources for the poorer countries.

[1] A breeder reactor uses slow neutrons of fission to produce more fissile material than it consumes. I will leave it as an assignment for the reader to consider how that is possible, despite the Laws of Thermodynamics preventing the existence of perpetual motion machines. I will consider it on my next energy post…