4. Long-term Oil Demand is Being Overstated
For the purposes of this paper, the term ‘energy establishment’ applies to the IEA, the OPEC Secretariat, the US Energy Information Administration (EIA) and numerous other institutions in the business of forecasting energy supply, demand and market developments. It includes the large IOCs that regularly produce energy forecasts. The prevailing view presented by this establishment is that future oil demand will continue to be strong. Figure 7 illustrates a typical forecast from the IEA. Other members of the energy establishment present similar views.
Figure 7: IEA estimates of oil consumption by sector
This view of future oil demand underestimates two factors that could significantly cut oil demand growth. Both are linked to prices. The first is the fact that, as Figure 8 shows, between 1998 and 2011 the price of crude oil rose in real (i.e. inflation-adjusted) terms, and that – as is explained below – this will take time before it begins to affect future demand.
Figure 8: Real oil prices 1990–2017 (Brent)
There is a tendency among energy analysts to forget the time lag between price movements and changes in oil demand. Oil demand is what is termed a ‘derived demand’. No one wants a barrel of petrol or a bag of coal. Energy consumers want energy services – light, heat and work. To secure these services requires the use of energy-consuming appliances. A three-stage process therefore affects energy consumption. It involves a series of choices that ultimately determine oil demand. The first is whether to buy the energy-using appliance or facility. While the price of the appliance matters, of greater importance is the income of the consumer. As incomes rise, more appliances are purchased.
The second choice is what type of appliance to buy. Here there are two issues: which fuel should power the appliance, and whether an efficient or inefficient appliance should be preferred. The choice of fuel is determined by the technology. A jet aircraft requires the use of jet fuel, but boilers to generate steam can depend on a variety of fuels. A key determinant of the choice of fuel, depending on the technology available, is the current price of fuels. However, expectations regarding future energy prices are probably an even more important factor. As to efficiency, at least historically, more fuel-efficient appliances tend to command higher purchase prices. So this presents a trade-off between the initial price of the appliance versus any savings on running costs that may result from higher fuel efficiency. Once these choices have been made, the appliance stock is fixed for a significant period (as it takes time to change the appliance stock to any extent). For example, the car parc normally takes around 15 years to change, and the housing stock well over 50 years.
Once the appliance stock is fixed, the third and final choice in determining demand for fuel is the capacity utilization of the appliance. Here there is an important conceptual distinction between ‘conservation’ and ‘deprivation’. Doing more or less the same thing with lower capacity utilization constitutes conservation, and can be viewed as a desirable action. However, to save energy through zero capacity utilization – e.g. turning off all the lights and sitting in the dark – is ‘deprivation’. This might be seen as undesirable. In the short run, when the appliance stock is fixed, and if no deprivation is to occur, it takes time for higher oil prices to reduce fuel demand to any significant degree. Thus, the higher prices experienced between 2004 and mid-2014 will take time to reduce oil demand significantly, but they surely will. The experience following the oil price shocks of the 1970s provides a classic example of this phenomenon of lagged demand response, as briefly outlined in Box 1.
Box 1: The impact of the oil price shocks of the 1970s on OECD oil demand, and how the response is lagged
Figure 9 shows the real crude oil price and OECD oil consumption. The underlying story is as follows. OECD oil consumption was rising rapidly post-1968 in response to two drivers. First, the period after 1970 saw a major economic boom in the OECD, with very high rates of GDP growth.
Figure 9: OECD oil consumption and Brent oil prices, 1968–85
Second, oil was increasingly used for power generation as the expectations were that the fall in the oil price at the end of 1960s was expected to continue. Hence (in an example of the three-stage consumption decision described above), generators chose to run new power stations on oil when, as a result of the economic boom, electricity demand was growing very strongly. Then, in 1973–74, the price of oil quadrupled. Immediately, oil consumption fell as the capacity utilization of the oil-using appliance stock fell. In many cases, this change represented ‘deprivation’ – consumers were unable to afford to use oil for essential energy services. At the same time, oil consumers and manufacturers of oil-using appliances began to respond by trying to improve oil-use efficiency. This was most apparent in automotive manufacturing. However, turnover in the car parc takes time, so the effects were not immediately apparent. Then, in 1979–80, there was a second oil price shock, which was associated with the Iranian revolution. Again, the capacity utilization of oil-using appliances fell. However, this time, the more oil-efficient appliances were beginning to enter the appliance stock at scale, and so overall oil consumption fell to a much greater extent. This provides a classic example of how changes in oil consumption are lagged relative to increases in oil prices.
The more interesting question is not when ‘peak demand’ will occur, but what will happen afterwards. Will oil consumption gently decline or will there be a collapse? This will be of crucial importance to any assessment of the geopolitical implications of the energy transition.
A second and more recent influence that is set to undermine oil demand growth – likely causing demand forecasts to be revised downwards again – is also related to price. Since mid-2014, the price of crude oil has collapsed in both real and nominal terms. However, ‘OECD disease’, already discussed, means that these lower crude prices have not yet translated into the lower product prices that drive movements in oil consumption.
The possibility that future oil demand is being overstated has started a discussion about the concept of ‘peak oil demand’ (Lynch, 2018; Dale and Fattouh, 2018). The original concept of ‘peak oil’ referred to the idea of a peak in supply occurring as a result of reserves being depleted. The concept was seriously flawed, since it took no account of technology, costs or prices. It was finally effectively buried by the shale technology revolution that dramatically increased recoverable reserves. However, it has now been replaced by the idea of ‘peak demand’. There has been much discussion over when global oil demand will peak – indeed, it appears to have already done so in the OECD, having hit its highest level in 2005.
In some ways, this is the wrong question. The more interesting question is not when the peak will occur, but what will happen afterwards. Will oil consumption gently decline or will there be a collapse? This will be of crucial importance to any assessment of the geopolitical implications of the energy transition. The OECD experience of gentle decline may not necessarily be a good indicator of what to expect in emerging economies. This is because energy sector developments in emerging economies in the coming decades will be subject to structural disruptive forces, rather than to a more predictable interaction between improving energy intensity trends and saturation of car ownership.
Technology is already arguably bringing the date of the peak nearer. The role of EVs has already been discussed. As outlined in Section 3, other technological developments may also reduce oil demand in the future.