The dispute over the future size of oil revenues that erupted last week is not helpful.

Following the Office for Budget Responsibility's (OBR) publication of its latest Fiscal Sustainability Report and the Scottish Government's paper on oil and gas, the argument has generated more heat than light.

The SNP claim the remaining oil reserves are larger and the revenues greater than OBR have assumed. They also claim the UK Government has always underplayed the importance of North Sea oil and there is a suggestion that this has been done for political reasons.

Past estimates of the amount of oil and gas resources have certainly proved to be too low, as has been the expected length of life of the industry. But this is not deliberate; it has been based largely on data the companies themselves have provided. These figures are naturally conservative because companies only include what they know about, not what may or may not yet be discovered. That the industry is now expected to last much longer than previously expected is excellent news, but not something that could have been predicted in the 1970s or 1980s.

Oil revenues, however, depend not just on the amount of oil or gas remaining, but on prices and the costs of extracting the oil. Prices are notoriously volatile and, in conjunction with the costs of extraction, determine what is commercially viable to produce.

No-one in the early 1970s would have predicted that the international oil price would have quadrupled by the end of the decade. Nor would anyone in 1980 have foreseen the drop in oil prices that took place in the middle of that decade. Costs, too, can change as technological advances make it possible to extract oil and gas from smaller fields and from greater water depths, such as those of the fields west of Shetland, that were previously considered uneconomic.

This means that forecasting oil revenues is a pretty hopeless exercise. I doubt if the OBR would even have attempted it, if it had not been that they had to have figures to put into their long run forecast of the UK's fiscal position. Of all the components in future UK budgets, revenues from the North Sea are undoubtedly the most difficult to predict with any hope of being right. In such a situation the natural instinct would be to err on the side of being conservative.

What is known is that the output of North Sea oil peaked in 1999 and gas in 2000. Both are now at less than half of the levels of production achieved then. The best estimates are those made by Professor Alex Kemp of Aberdeen University, the official historian of North Sea oil and gas. He expects the decline in the volume of output to slow; it might even be flat or reversed for a short time as the major investment presently being undertaken comes to fruition. But the trend is inevitably for continued decline until the resource is substantially depleted by about 2040.

Revenues, even if output temporarily rises, may continue to fall as companies set off against profits the costs from their present high level of investment. As decommissioning of the older structures in the North Sea gathers pace, that will involve huge costs and that too will be set against profits, thereby reducing tax revenue.

All of these factors will have influenced the OBR in deciding to reduce their estimates of future revenue. Kemp's analysis shows that some 90% of the oil is in what under international rules would probably be Scottish waters, if Scotland were independent. That would be subject to negotiation as it has never been agreed. But on that basis, Scotland's share of the total oil revenues over the last few years would have varied between £5 billion and £12bn a year, which gives some indication of their volatility.

These figures compare with the early 1980s when the total oil revenue for the UK reached a peak of around £28bn, if calculated at constant 2009-10 prices.

For the future the OBR predicts that revenues, which were 0.7% of UK GDP in 2011-12 and fell to 0.4% in 2012-13, will fall to 0.2% in 2017-18 and reach 0.03% in subsequent decades.

That may be too low but the truth is nobody knows. Forecasting oil revenues are no more reliable than the long range weather forecast.

Forecasters get that wrong because they cannot predict where the jetstream will be.

Those who forecast oil revenues do not know what to assume for prices, exchange rates, costs and technological progress, even if they have some idea of what oil is still under the sea.