How to Think About Forecasting L-T Oil Consumption
The challenge of energy company managers is planning their long-term future. That depends on future oil demand. Studying the economy’s oil intensity and price elasticities suggests less consumption in the future.
Current industry and investor focus on the oil business is attempting to assess how quickly demand will return. That is key to lifting oil prices and restoring producer profitability. Presumably, once producers believe they can drill, complete and produce oil profitably, they will resume work, which is critical to the survival of the oilfield service industry. Understanding the state of demand is why forecasters, analysts and industry executives closely watch mobility data and other measures of economic activity that use oil.
After feeling sanguine about a rapid recovery in activity, as the various states reopened their economies, that faith has been shaken by the recent surge in Covid-19 cases in a number of states forcing their governors to pause or rollback the easing of restrictions. Texas traffic data is reflecting declines after many weeks of rising congestion. While watching these measures and adjusting current activity, oil companies operate on much longer trends than those for the next 30-90 days. In fact, oil companies are really wondering about oil demand in 2021 and beyond, as projecting the supply needed will help them set future production targets and new drilling and completion work.
Oil company capital spending this year has been slashed, and may be subject to further reductions, that provides optimists and pessimists, depending on their respective views on future oil prices, with ammunition for their forecasts. Higher prices will come due to the lack of capital spending now and rebound in demand. Given the steep output declines of shale oil wells, the optimists see supply falling short of demand. The days of $100 a barrel oil cannot be far off under their thinking.
The only way this scenario fails to unfold is if there is a surge in supply, likely from the high-grading of new wells, plus the ability to extract more oil per well. On the other hand, the high oil price scenario could be derailed by a lack of oil demand. There are numerous ways the demand rebound might not happen: a second wave of Covid-19 forces lockdowns of state economies; the current economic recession becomes much worse as the closure of businesses, high unemployment and weak income growth restrains consumer spending; commuting traffic fails to return, as workers and companies embrace working at home; schools and universities teach predominantly online; air traffic fails to recover; and global trade falls. While many forecasters weigh these issues in their demand models, most models incorporate only minor lifestyle and work-related changes. Assuming a greater impact pushes the pre-Covid-19 recovery out into 2022, or maybe later, rather than 2021. In fact, Citibank’s commodity group sees oil demand as having peaked and never to be seen again.
While watching these measures and adjusting current activity, oil companies operate on much longer trends than those for the next 30-90 days
Exhibit 1. Expect Lower Oil Consumption Through 2050 Source: EIA, PPHB
We have previously presented the Energy Information Administration’s (EIA) long-term oil consumption forecast, along with modifications we anticipate based on permanent demand shifts due to the pandemic. We see a demand shortfall in 2030 of 2.9 million barrels per day (mmb/d), or 3%. That is the equivalent of losing the output from Kuwait or the United Arab Emirates, not inconsequential.
Expect Lower Oil Consumption Through 2050
Here is a recent forecast from the energy classification and consulting firm DNV GL. Their view is consistent with ours, and for similar reasons. We acknowledge the other scenarios calling for little structural change, but these models tend to be more short-term focused with less concern about long-term energy intensity shifts and slowing global economic growth. The vast sums of debt added to virtually every major economy due to pandemic response will slow future growth, consistent with historical experience. In fact, after virtually every significant economic shock, the future growth rate for oil has been lower. That said, we are always willing to change our view based on updated information.
Exhibit 2. How Long-Term Oil Demand Will Contract Source: DNV GL
How Long-Term Oil Demand Will Contract
DNV GL wrote about the assumptions underlying their latest forecast. Rather than paraphrase them, we elected to report them as written.
“Our energy forecast is predicated on IMF’s longer outbreak scenario, where World GDP will shrink 6% in 2020. The lingering effects of the pandemic will take the wind out of the sails of the world economy for many years – reducing World GDP in 2050 by 9%, relative to pre-pandemic forecasts. Even with slower growth, however, by mid-century the world economy will still be twice its size today. In contrast, energy demand will not grow. In 2050, it will be about the same as it is today, in spite of a larger population and world economy. This is largely due to significant improvements in energy intensity, but also due to the effects of COVID-19.
“Before the pandemic, we predicted total global energy demand in 2050 at 456 exajoules (EJ), (Global energy demand using the latest historical figures was at 424 EJ in 2018.) Our modelling now shows that the pandemic will reduce energy demand through to 2050 by 8%, resulting in energy demand in 2050 at almost exactly the level it was in 2018.
“Improvements in energy intensity will remain the most important factor in reducing energy demand in the coming decades, and the contraction due to COVID-19 comes on top of this. That is as a result of the brakes applied to economic activity generally by the pandemic, as well as some specific sectoral impacts. Lasting changes linked to COVID-19 are mainly behavioral in nature and include the impact of the pandemic on the transport sector, especially aviation, but also on less office work and changed commuting habits, which will result in transport energy use never again reaching 2019 levels. Demand for manufactured goods globally will need almost four years to recover to 2019 levels, and the energy-intensive iron & steel industry, impacted inter alia by lower demand for new office space, may never reach its pre-pandemic heights.”
Oil companies need a sense of what consumption levels will be in 2030 and beyond, since they are responsible for developing the supply to meet that demand. Understanding long-term oil consumption helps companies plan their future drilling and completion activity. In that regard, a recent webinar hosted by the United States Association for Energy Economists, featuring Dr. Marie Fagan, the Chief Economist at London Economics International, LLC., provided some interesting insight. Dr. Fagan based her presentation on a paper she prepared in November 2018 for the Columbia University School of International and Public Affairs, Center on Global Energy Policy. The paper, titled “Oil demand: Up the down staircase,” was an econometric study of income and price elasticities of demand for crude oil and key refined products.
Besides listening to the webinar, we read Dr. Fagan’s paper. As the analysis ended in 2016, we have updated it, except for the econometric work, for which we do not have access to Dr. Fagan’s model. In considering the paper’s conclusions, we doubt the elasticities calculated would be materially different if three additional years of data were included, at least for crude oil. While there is a higher possibility that the elasticities for refined products – gasoline and diesel – might change, we doubt the changes would be material.
When we consider the history of oil consumption for 1965-2019, we foresee a steadily rising trend, albeit with different rates of increase. We also note that oil consumption declined whenever oil prices, in real terms, rose sharply. The most notable times were in the 1970s, at the time of the First Oil Shock, and then again at the time of the Great Recession in 2008-2009. We also see less prominent declines whenever the oil price spiked. Dr. Fagan’s paper was designed to examine why these relationships existed and if they changed over time. Detecting any relationship changes could help in forecasting future oil consumption.
Exhibit 3. Price Shocks Translate Into Demand Shocks Source: BP, PPHB
Price Shocks Translate Into Demand Shocks
The start of the analysis of crude oil and refined product demand involves understanding oil intensity in the economy. That measure is developed by calculating the amount of oil used to produce a million dollars of economic activity in real terms. We updated Dr. Fagan’s analysis and calculated oil intensity since 1965. We used the World Bank’s global gross domestic product (GDP) estimates in 2010 dollars as well as BP plc’s estimates of world oil consumption. The resulting chart shows a long-term declining trend in oil intensity, which has significant implications for oil’s future use.
Exhibit 4. Since 1980s Oil Intensity Has Been Declining Source: World Bank, BP, PPHB
Since 1980s Oil Intensity Has Been Declining
Oil intensity rose from 1965 to 1972, and then declined due to the 1973 oil price shock. As the economy adjusted to higher oil prices, oil intensity rose marginally until the next oil shock associated with the Iranian Revolution. Since then, oil intensity has declined, albeit at differing paces over 1979-2019. As a result of this analysis, Dr. Fagan reached her first two conclusions. Those were:
“The decades-long upward movement in crude oil demand has been undermined by a ‘descending staircase’ of declining oil-intensity of economic activity.”
“This finding implies that we should expect oil demand growth to continue to lag behind economic growth as key developing countries’ population and economic growth rates slow and they transition away from an oil intensive economy to a service economy, and the world experiences periodic oil price spikes due to geopolitical events that destroy oil demand, which is not fully restored when prices recede.”
In a world in which global economic growth is moderating, it is unlikely world oil consumption will grow rapidly. The long-term history of oil use versus GDP growth shows annual rates of increase slowing dramatically since the end of the 1960s and early 1970s. At that time, oil and GDP were growing at 6%-8% rates. In recent years, growth has been more in the 1%-3% range. The question we have is whether these low growth rates will slow further, or begin rising?
There are numerous reasons to expect slower growth, primarily due to aging demographics in developed economies, and now also being experienced in China. Population predictions show India’s surpassing China in a few decades, but in the interim, China’s population is aging. At the same time, the global birth rate is falling, to the point it risks dropping below the replacement rate. In addition to demographics, productivity is slowing, which hampers economic growth.
Exhibit 5. Era Of Slower Growth Means Less Oil Demand Source: IMF, EIA, PPHB
Era Of Slower Growth Means Less Oil Demand
Add to that the potential for a reversal in globalization, lowering trade volumes. Huge government debts, besides growing corporate and personal debt, makes it harder for economic growth to be stimulated by future government stimulus.
We see the significance of the slowing growth rate of economies when we examine the history since 1973 of oil intensity of the developed economies (OECD) versus developing economies (Non-OECD). OECD economies have been steadily reducing their oil intensity since 1973, in contrast with Non-OECD economies where oil intensity rose from 1973 to 1979, then stayed level through 1984, before starting to decline.
In 2019, the world’s oil intensity was 1.16. The OECD rate was 0.86, whereas the Non-OECD rate was nearly double at 1.66. If these declining trends continue, they doom the prospects of oil demand growth ramping up significantly.
After analyzing oil intensity, Dr. Fagan used her econometric model to assess the income and price elasticities of crude oil and refined products. She had created populations of OECD and Non-OECD countries to conduct her analysis. A key aspect of the analysis was to divide the history into two eras: 1977-1996 and 1997-2016.
Exhibit 7. Changing Oil Income And Price Elasticities Source: Dr. Fagan
Changing Oil Income And Price Elasticities
The econometric analysis showed that income elasticities of crude oil demand were higher in the 1977-1996 period than the 1997-2016 period for both OECD and Non-OECD economies. For OECD economies, the long-term income elasticity of demand was 0.50 in 1977-1996, meaning a 1% change in GDP corresponded to a 0.50% change in oil demand in the same direction. But for the second period, 1997-2016, the income elasticity declined, but only to 0.46. The income elasticity for Non-OECD countries was much higher in the early period at 0.72, but it fell in the later period to 0.47, not much different than the elasticity of OECD countries. This means the relationship between economic activity and oil consumption did not break down completely, but it did weaken.
The price elasticities provide another interesting perspective about future oil growth. For the 1977-1996 period, price elasticity was similar for both OECD and Non-OECD economies. In 1997-2016, both groups of economies saw lower price elasticity measures. The Non-OECD countries actually had a tiny and not statistically significant price elasticity for this later period.
Dr. Fagan believes the decline in long-term price elasticity may reflect the response to an oil price spike driven by supply shocks versus price hikes due to fast-growing demand. The first period included the oil price shocks of the 1970s, while the latter period included the rapid growth of the China-driven oil boom in the early 2000s. As Dr. Fagan wrote, “The price volatility in the earlier period was seen as a crisis—the ‘energy crisis,’ and may have led consumers to expect a long period of high prices, or of physical shortages of oil. Also, opportunities to substitute other fuels for oil may have been more abundant in the 1970s and 1980s, as oil was more widely used for heating and electric power generation in many locations. Both these conditions would lead to higher (absolute) price elasticities of demand.”
In her view, the oil price increases of 2001-2008 did not create fear of oil shortages, and did not create a global recession. Additionally, there may have been fewer opportunities for efficiency gains and fuel-switching. Both conditions might explain why the more recent period experienced lower absolute price elasticities of oil demand.
The elasticity relationships have important implications for predicting future oil consumption. As we showed earlier, the oil intensity of Non-OECD countries is much higher than the oil intensity of OECD countries. Therefore, the fact that the income elasticity of Non-OECD countries is nearly half of what it was in the earlier years suggests that even as these countries grow, their need for oil will not be as great as it was before. How should forecasters look at possible scenarios for future oil consumption?
Exhibit 8. What Is Future For Oil Consumption? Source: BP, PPHB
What Is Future For Oil Consumption?
If we focus on how the relationship existed in the 1970s, we see oil consumption exceeding 120 million barrels per day (mmb/d) by 2030. On the other hand, if we look at how oil consumption was growing prior to the Great Recession of 2009, oil consumption reaches close to 119 mmb/d in 2030. Will oil consumption only grow in line with the recent history prior to the 2020 downturn? If so, then we are looking at 2030 oil consumption close to 115 mmb/d. However, the shocks the global oil markets have experienced this year due to Covid-19 and the Russia-Saudi Arabia oil war, raise the question of whether we are going to resume growth from a much lower base and at a lower rate. We have presented several possible trajectories that keep oil consumption below 100 mmb/d in 2030.
We should consider the moves by Royal Dutch Shell and BP plc to write down the value of some of their oil and gas assets in response to lowering their long-term oil price forecast to $55 per barrel, down from their prior $70 forecasts. This is indicative of a view of a lower oil consumption growth rate. Those moves, likely to be replicated by other oil companies, signal that fewer reserves will be developed because they will not be necessary. The industry is beginning to embrace the “lower for longer” mantra that emerged in 2015.
The trend of global oil intensity, and especially the convergence of the Non-OECD and OECD measures, is an indicator for lower future oil consumption. This is reinforced by the lower income and price elasticities. The industry is just beginning to come to grips with this outlook. The restructuring of companies – streamlining operations, selling marginal producing assets, as well as reconfiguring, selling and closing refining plants signals management expectations of a different world for their businesses.