Global oil demand - an inexact science
Should you bet the house - or your company's drilling programme - on long-term forecasts for oil demand?
Forecasting long-term oil demand, never easy, is getting harder. Opec's latest games with the oil price, Trump's election in the US, Brexit, new battery developments: politics and surprises can play havoc with the models.
In December, the International Energy Agency (IEA) tweaked its shorter-term forecasts to reflect stronger than expected oil demand in China and Russia. Global oil consumption will have risen in 2016 by 1.4m barrels a day, or 120,000 b/d more than previously thought, and in 2017 demand will rise by 1.3m b/d, or 110,000 b/d more than its earlier projection.
Such adjustments are a regular feature of the IEA's market outlook. If the world's leading energy-market forecaster needs to change its near-term projections, you can see why forecasts for how consumers will behave decades hence can be so varied too. Who, 25 years ago, expected you might be reading this article on your phone?
At least most of the big forecasters-alongside the IEA this includes the US Energy Information Administration (EIA), Opec, and a host of oil majors-have one thing in common. They all think oil demand will keep rising, driven by economic growth in developing economies like India and China, and in segments like petrochemicals, shipping, road transport and aviation.
To deal with the uncertainties inherent in these outlooks, the caveats are now key.
In its World Energy Outlook (WEO), the IEA lists three main scenarios-Current Policies, New Policies and the 450 Scenario-distinguished principally by the world's response to climate change. Only the last of these takes full account of oil demand if all the measures needed to restrict global warming to 2°Celsius above pre-industrial levels are implemented.
Opec also produces a range of outlooks, including one incorporating a relative penetration by electric vehicles (EVs) in the transport segment. The EIA highlights somewhat different variables, producing high, low and medium oil-price scenarios and forecasts based on differing economic-growth assumptions.
Despite the differing methodologies used to come up with their forecasts, the three agencies' long-range conclusions tend to be remarkably similar-at least in their core scenarios.
The IEA's latest WEO demand forecast for liquids (oil plus other liquids), published in November 2016, expects consumption to reach 107.7m b/d in the New Policies scenario-its baseline outlook that assumes commitments made around climate change will be carried out. But under the Current Policies scenario, which assumes no new climate measures are enacted, it sees 120.6m b/d by 2040. Only in the 450 Scenario does liquids demand fall-and sharply, to 82.2m b/d in 2040.
The EIA's middle oil-price reference outlook is comparable to the IEA's current policies one (not presuming any progress in climate policy), and it's almost identical in its conclusion, expecting consumption to reach 120.9m b/d in 2040. The two organisations do share some data, and their oil-price assumptions for 2040 are similar-between $140 and $150 a barrel.
ExxonMobil shares this broad thrust. In January, it produced a figure of 224 quadrillion British thermal units for global oil demand in 2040, a rise of 18% from 2015 and around a third of total primary energy demand in 2040. That comes in somewhere between the IEA's New Policies Scenario, which sees crude oil demand of 103.5m b/d for 2040 (a rise of 12% from 2015), and the 117m b/d in its Current Policies scenario (a 26% rise).
Above all, the major forecasters see the same direction of travel, expecting only a gradual slackening of oil-demand growth over the next 25 years. Fatih Birol, the IEA's executive director, summed this up last November, saying "peak oil demand is not in sight".
Yet some cracks in this consensus are appearing. Simon Henry, Shell's chief financial officer, told analysts in November that the company had "long been of the opinion that demand will peak before supply". He suggested that the peak "may be somewhere between five and 15 years hence, and it will be driven by efficiency and substitution, more than offsetting the new demand for transport".
Even Opec sees the possibility of demand peaking as soon as 2029-though only if all the measures in the Paris climate-change agreement are executed. The Paris agreement came into force last November. But predicting how rapidly its climate mitigation efforts translate into a hit on the oil market was hard enough before Donald Trump took the helm of the world's biggest economy (and energy consumer).
A bias that sees government policy as the guiding force-and not so much behavioural economics-tends to lie behind most oil-demand forecasting too. The IEA's Birol wrote in the WEO, that while "there is no single story about the future of global energy, in practice, government policies will determine where we go from here".
But energy consumption is also determined by personal preference and plain economics. For example, wind and solar energy are already as cheap to install and run as any other power source in some parts of the world, so long as the grid can handle them-and their costs are still coming down. Lower costs could make policy on renewables subsidies much less important drivers, though the financing of infrastructure able to cope with intermittent power is still likely to require government involvement.
Renewable energy is a greater challenge for gas and coal producers, which dominate the power sector. Oil accounts for only around 5% of global power generation, according to the IEA. But adoption of green energy-and improved transmission infrastructure-in the developing world will still dampen demand for diesel for generators and oil-fired power stations.
The real problem for the oil industry-and its forecasters-will come if consumers ditch the internal combustion engine (ICE). Passenger vehicles-the segment most vulnerable to competition from EVs-account for nearly a fifth of global oil demand. Transport as a whole consumes over half the world's oil supply.
Industry titans don't seem overly worried. ConocoPhillips's boss Ryan Lance spoke for many of his peers when he said recently that EVs wouldn't have a big impact on the oil industry for another 50 years.
But that's not a universally shared view.
Bloomberg New Energy Finance produced research in December forecasting that EVs could displace about 8m b/d of demand by 2035, rising to 13m b/d by 2040. That would account for about 15% of the world's current oil demand (or more than Russia's total production). By some estimates, oil demand from passenger vehicles could peak in the early-to-mid 2020s, as EVs make inroads, in turn forcing ICE vehicles to become more efficient.
It's isolating the tipping point-if one is to come-that's so difficult.
The first wave of ICE-powered cars didn't erode the fleet of horse-drawn transport overnight. But in retrospect it certainly looks like that. Within about a generation, in the industrialised world, mass-produced Model T Fords had clogged streets once packed with horses. The pace of change was even faster in the cities, where a policy to get rid of horses, and their dung, speeded up the process. To some, that transition looks awfully akin to today's promotion of clean transport to reduce urban pollution.
The analogy shouldn't be pushed too far. The ICE had more demonstrable advantages over horsepower than EVs have over the ICE. That's one reason for scepticism. BP's chief economist, Spencer Dale, says he doesn't see EVs as a "game changer" for oil demand over the next two decades.
Clean air push
Still, momentum is plain in one of oil's biggest target markets. China-where city smog is deathly-overtook the US in 2015 to become the largest consumer of clean-energy vehicles, including EVs, plug-in hybrids and fuel-cell vehicles. China's plug-in EV fleet is estimated to total more than 0.65m vehicles already, with sales supported by heavy subsidies. The government wants output of 3m EVs by 2025. Chinese EV manufacturers are proliferating, to gobble up a slice of this pie.
To its credit, the IEA tweaked its expectations of EV growth upwards in its 2016 scenarios to reflect this (and did the same for renewables). Look out for more revisions in years to come.
But cars are just one area of uncertainty. Less sexy, but equally significant, are petrochemicals, another bedrock of medium and long-term oil demand. In the reference case for its World Oil Outlook 2016, Opec is counting on petrochemicals demand for oil rising by 3.4m b/d, or 28%, between 2015 and 2040, to 15.3 million b/d. (As ever now, the developing world is the engine of this growth.) But naphtha faces increased competition from gas as a feedstock and from technologies such as coal-to-olefins and methanol-to-olefins in key markets such as China. Oil prices will need to stay low enough to keep naphtha competitive with these rivals.
But what if people want less plastic?
In a recent outlook, McKinsey, a firm of consultants, estimated that petrochemical feedstock would drive 70% demand growth for liquid hydrocarbons until 2035. But the authors warned of a slowdown in the expansion of demand, as mature markets reached saturation point in their appetite for plastics. Also, if global plastic recycling improved from the prevailing 8% rate to 20% in 2035 and plastic packaging use declined by 5%-some fairly modest assumptions-the impact on demand for liquid hydrocarbons from the chemicals industry could be 2.5m b/d lower than thought.
So you pay your money and you make your choice when it comes to oil-demand forecasting. Confirmation bias comes in the package. And time hasn't been kind even to forecasts made just 10 years ago. The IEA's WEO of 2006 expected global oil demand to be 99.3m b/d in 2015. In fact, it came in 94.9m b/d. Few economists foresaw the global financial crisis, so the IEA can't be blamed. Still, if such relatively short-term projections can be so askew, what of those for 25 years hence?
The industry itself has started to show some self-awareness in this regard. The authors of the World Energy Council's outlook, released in October, eschewed heavyweight nomenclature. Their scenarios were Modern Jazz, Unfinished Symphony and Hard Rock, referring to market driven, low-carbon and unsustainable growth models, respectively.
Music involves varying degrees of emotion, unpredictability and imagination, along with underlying rigour and a sense of direction-not a bad description of the forces that will act on the oil industry over the next quarter of a century.