The great Victorian economist, Stanley Jevons, published an exhaustive study of the coal industry in 1865. An exemplary exercise in applied economics, comprehensive in its analysis of supply and demand, geology and technology, it proved influential. William Gladstone, then chancellor of the exchequer, devoted much of his Budget speech to praising Jevons's achievement and implementing his recommendations.
Jevons had concluded that Britain would be unable to meet its future needs for coal. These requirements - which, at 80m tonnes per year, constituted a larger share of world energy consumption than America's today - would grow to 500m tonnes in 1911 and could, on unchanged policies, reach 2.5bn tons in 1961.
British coal use did continue to grow for 50 years after Jevons wrote and reached a peak of 300m tonnes in 1913. By 1961 it had fallen to 200m tonnes, twice the level of the 1860s but less than 10 per cent of his estimate.
Jevons dismissed the idea that electricity might meet the needs of industry. But he did conceive that power might be obtained by capturing sunbeams and noted ruefully that England would then be poorly placed to compete effectively on world markets. Jevons recognised that some American inventors believed steam engines could be fuelled by petroleum. He did not, however, consider the possibility that petroleum could be obtained from a source other than coal. What is petroleum, he asked rhetorically, but essence of coal?
The world has changed since 1867, but only in ways that falsify predictions more rapidly. Several agencies collect forecasts for long- term oil prices. It is hard to forecast oil prices, but easy to forecast forecasts of oil prices. Forecasters always expect that prices will revert towards their long-term average within the next year or two and thereafter increase at a steady rate. Consensus forecasts of current oil prices made more than a decade ago range from below $30 to above $100. The estimate depends mainly on whether oil prices were low or high when the forecasts were made.
None of this dismal experience deters policymakers from making projections and basing proposals on them. The Intergovernmental Panel on Climate Change lets its models run for 100 years with the same insouciance as Jevons. The European Commission recently published its long-term assessment of Europe's energy needs. And the British government - hoping, no doubt, that no one will ask how well the predictions and policies contained in past documents have fared - has just published a policy document on energy strategy.
The need for such strategies appears compelling. Energy is fundamental to every human activity. Major energy projects have long lead times. Surely we need to make the best forecasts we can and plan accordingly? But this is to repeat Jevons's mistake.
It is absurd to criticise Jevons because he did not predict the future correctly. He could have anticipated the internal combustion engine only by inventing it. How could he have realised that the deserts of Saudi Arabia concealed essence of coal far more productive and extensive than the coal fields he had surveyed?
A justified criticism of Jevons is not that he did not know things he could not have, but that he did not recognise the limits of his knowledge and made recommendations with unwarranted confidence. But the voices of the wise who know that they do not know are often drowned by the ignorant who do not know they do not know.
The right approach is not to peer more carefully into the crystal ball: there is nothing to see. The availability and future price of fossil fuels is highly uncertain. We can be sure there will be new fuel technologies but we do not know what they will be and some will be far outside the scope of our imagination, as nuclear electricity was beyond the scope of Jevons'.
We should eschew forecasts, acknowledge unresolvable uncertainty and plan accordingly. The right energy policy is one of diversification - developing as many uncorrelated options as possible. That is the argument for a strategy for energy supply I will develop next week.