21st January, 2008
The government published, at the same time as the Heathrow consultation, new forecasts for passenger demand in Nov 2007 (‘UK air passenger demand and carbon dioxide forecasts’).
The forecasts are for the UK as a whole and for the big London airports. They revise the previous demand forecasts (published in 2000) and the carbon dioxide forecasts published in 2004. With minor changes, these earlier forecasts formed the basis of the White Paper on aviation released in Dec 2003, forecasting up to 2030, with an extension to 2050 using a simple methodology. A summary of the method by which the forecasts are derived is given in a separate document (which also gives our views/critique in slightly extended fashion). The resulting forecasts are:
These are ‘constrained’ forecasts, which allow for the fact that some demand would be suppressed due to anticipated lack of capacity. However, the ‘unconstrained’ demand in 2030 is 495m, just 3% higher. For all practical purposes, therefore, a ‘predict and provide’ approach for aviation at national level is implied.
These figures are the government’s best estimate or ‘central’ case. A series of sensitivity tests were performed in which a wide range of parameters were varied to give a ‘low’ and ‘high’ forecasts. At 2030, the low forecast is 450 (ie 6% lower than 480) and the high forecast is 505 (ie 5% higher).
1. The mathematical model has not been studied in detail by AEF, but we consider the general approach to be reasonable. The underlying assumptions, namely that demand is basically a function of the growth of the economy and the change in prices of flying, seem sound and have been used by other forecasters.
2. The figures for income and price elasticity have been based on an analysis of the literature, reported in the paper, and we do not have evidence to confirm or dispute the results. However, the change of price elasticity since the previous forecasts from about -1 to -0.44 must engender some suspicion. The use of -1 in the NGOs request for the DfT to rerun ‘SPASM’ [note] with increased prices to reflect a tax on aircraft equal to the tax on petrol, showed that the demand for air travel would be reduced greatly and the runways proposed in the White Paper would be unnecessary. Since that time, the government has always claimed that changes in estimated price would have little effect on demand, thereby showing the expansion policy was robust. The reduction of price elasticity to -0.44 makes the forecast more robust in this respect. (However, the figure masks big differences: the elasticity for business travel is thought to be close to zero; for short-haul leisure it may be around -1.)
3. For the purpose of the forecasts, oil prices (in real terms) are assumed to fall from around $100 per barrel, the price at the time of writing, to about $65 per barrel in 2006 to $53 per barrel in 2030, with most of the decline occurring by 2012. Given the soaring demand from India and China, and generally uncertainty about supply (eg ‘peak oil), the assumptions seem totally unrealistic.
4. DfT has assumed that a ‘cost of carbon’ is built into air fares from 2010. This cost is intended to reflect the cost of damage due to global warming and the forecasts use £93 per tonne (at 2007). AEF has severe misgivings about this figure. The Stern report gave a range of figures, and there are strong grounds for supporting a cost as high as £280 per tonne. The use of £93 appears conservative and helps to boost the net economic benefits of airport expansion.
5. No realistic costs for impacts other than climate, such as noise and air pollution, are factored into the forecasts. This, together with the low climate costs means that the price of flying is artificially low, given the stated objective of getting airlines to meet their full social and environmental cost. Hence, the demand is much higher than it would be if these costs were included.
6. There is an implicit assumption that there will be no direct constraints placed on air travel because of climate change or for other environmental and social reasons (other than minimal constraints imposed by airport capacity).
7. The sensitivity tests give a narrow ‘band of uncertainty’ – only 5 or 6% – on either side of the central forecast. Given the intrinsic uncertainties in forecasting and the long time period involved, this band is patently too narrow. One reason is that the high and low forecasts do not allow for more than one significant parameter to vary simultaneously (e.g. lower economic growth combined with a higher oil prices). The narrow band gives a misleading impression of the robustness of the forecasts.
8. Beyond 2030, the forecasts of CO2 are extremely dubious. Passenger growth is forecast to slow: ” .. capacity constraints begin to bite again, so that growth in passenger demand slows.” It appears the government is hypothesising that its ‘predict and provide’ policies, confirmed in the 2003 White Paper, will be continued for many years and then abandoned. No justification is given for this startling policy about-face. It is probably no co-incidence that long-term climate targets are set for 2050 but the aviation policies only extend to 2030.
A booklet has been written by Brendon Sewill for Airport Watch called Fallible Forecasts (15pp; 240kbytes). It critiques the DfT passenger forecasts and also the forecasts of CO2 emissions. It is complementary to the AEF paper, being more accessible and published in paper form. Hard copies are available on request from AEF.
Note: SPASM is the DfT’s passenger allocation model.