Why Does Gasoline Cost So Much?
Gas prices are a product of supply and demand. This column attributes recent gas price increases to stagnant oil supplies and growing global demand from emerging Asian economies – not speculators. Additional shocks to the US refining capacity further tightened gas prices in the US. These forces will likely keep gas prices high for the foreseeable future.
At the end of 2007, both gasoline and crude oil prices (adjusted for inflation) were at levels last seen in 1981 and they continued to climb throughout much of 2008. While Europe has been cushioned in part from these developments, as the dollar depreciated against the euro, the fundamental forces that drove up US gasoline prices have done the same in Europe.
With retail gasoline prices in the US persistently above $4 per gallon, the determinants of gasoline prices is no longer an esoteric topic best left to industry insiders. The debate has moved into the mainstream. Congressional committees as well as media pundits have advanced explanations and proposed policy changes to stem or reverse the increase in gasoline prices.
Why did this surge occur? To answer this, it is important to distinguish between:
the price of gasoline and other motor fuels, and
the price of crude oil in global markets.
A distinction often ignored in discussions of higher energy prices. In a recent Vox column, Francesco Lippi discussed the price of crude. My column focuses on the US gasoline market, which is an interesting case for understanding the underlying market forces because of the availability of high quality data for extended periods.
Supply and demand shock in the US gasoline market
Recent research has stressed the importance of identifying the demand and supply shocks underlying unpredictable shifts in the price of imported crude oil.1 The same distinction between demand and supply shocks obviously applies to the retail gasoline market.
While crude oil is the main input in the production of motor gasoline, the US retail price of gasoline is also affected by gasoline-specific demand and supply factors, such as the ability of US refiners to process crude oil. Refinery fires, changes in the regulatory environment or refinery outages caused by hurricanes, for example, may cause increases in the retail price of gasoline that are not driven by events in the crude oil market.
But the US market does not operate in isolation. We can only hope to understand the evolution of US retail gasoline prices if we also account for the global demand and supply shocks that drive the price of crude oil.
In recent research, I propose such a model where global and US demand and supply shocks co-exist and there is feedback between the markets. Estimating it on US data, my analysis helps us understand the forces been behind the surge in US gasoline prices since 2002.2 The model allows for five distinct demand and supply shocks:
global crude oil supply shocks;
shocks to the demand for industrial commodities (including crude oil) that reflect fluctuations in global real activity;
demand shocks that are specific to the crude oil market;
domestic gasoline supply shocks (such as refinery outages); and
domestic gasoline demand shocks (reflecting changes in the degree of urbanisation, driving habits, tastes, etc.).
Given the importance of crude oil imports, it is natural to focus first on developments in global crude oil markets as the likely cause of the gasoline price increases in recent years.
Crude price hikes: it’s the demand shocks not the supply shocks
The model results show that unpredictable global oil supply disruptions (such as unanticipated production cutbacks by OPEC or disruptions of oil production in Venezuela or Nigeria, for example) played no role in the recent build-up of gasoline prices. Rather the bulk of the increase in US gasoline prices since 2002 has been associated with a series of unanticipated increases in global demand for crude oil (along with other industrial commodities).
Where did the extra crude demand come from? It can be shown that this sustained growth of demand was driven not so much by unusually high economic growth in OECD countries – although some of these economies experienced robust growth – but by additional demand from emerging Asia.
In addition to fluctuations in oil demand driven by global real activity, the model allows for demand shocks that are specific to the oil market. Economic theory suggests that growing uncertainty about future oil supply shortfalls boosts demand as oil companies attempt to build inventories to protect themselves from being caught short.3 As uncertainty shifts may be sudden, large and persistent, such “precautionary” demand shocks are a natural suspect in explaining the recent surge in oil (and gasoline) prices.
Precautionary inventory holding was not to blame
Such shocks played an important role in the oil price shocks of 1979 and 1990. My estimates, however, show that such shocks have not played a major role in recent years.
This evidence is important. There has been a lot of public discussion in recent years about the alleged role of speculators in crude oil markets and how speculators may be reigned in by legislative changes and regulatory supervision.4
A speculator in this context refers to someone who buys crude oil now, stores it, and hopes to sell it at a higher price later in anticipation of future shortages. If there were speculation in global crude oil markets, the resulting extra demand in the model would by construction be reflected in an increase in the component of the price of oil driven by precautionary demand.
My model estimates clearly tell us that there is no such increase, suggesting that speculation is not behind the recent gasoline price increases. The same conclusion is reached when looking at petroleum inventory data. If speculation were important, we would expect to see a noticeable increase in OECD oil inventories. The data show no such increase.
Finally, this conclusion is also consistent with the fact that most industrial commodity prices – not only the price of oil – have been growing at rapid rates in recent years. This pattern is consistent with a general increase in the global demand for industrial commodities but not with explanations that are specific to the crude oil market.
Having reviewed the role of shocks in the global crude oil market, I now turn to the role played by US domestic gasoline demand and supply shocks.
The US gas market: shocks to US refining capacity matter
A striking result is that US domestic gasoline demand shocks have had virtually no impact on the US retail price of gasoline in recent years. This is not surprising, as US consumption of gasoline has been moving very slowly and predictably. Indeed, that evidence is fully consistent with the notion that increased demand for crude oil and refined products comes from emerging Asia rather than the US.
In contrast, US domestic gasoline supply disruptions have played a role – in particular following Hurricanes Rita and Katrina in 2005. The primary effect of these exogenous events was not the reduction in US crude oil production (which was negligible on a world scale), but the reduction of crude oil refining capacity in the Gulf of Mexico.
Given that other US refineries were already operating close to capacity at the time, this event constituted a major unanticipated reduction of the supply of gasoline in the US, which would be expected to raise the price of gasoline sharply (while lowering slightly the price of crude oil, as demand for oil imports falls). The model indeed shows a sharp increase of US gasoline prices driven by adverse refinery shocks in late 2005. Only half a year later, the price seems to have stabilised again, although there is evidence of intermittent unanticipated refining shortages in 2006 and 2007 as well.
In short, the primary reason for the recent surge in gasoline prices has been growing global demand driven by developments abroad, given fairly inelastic global oil supplies. This evidence helps us understand why gasoline prices have risen to record levels, but what about the future?
Forecasting the future
Predicting gasoline prices is an all but impossible task even at horizons as short as one year. In recent co-authored work (Alquist and Kilian, 2008), I have shown that simple no-change forecasts are the most accurate forecasts of the price of crude oil in practice. In other words, the change in the price of oil is unpredictable.
Given the close relationship between global crude oil prices and domestic retail gasoline prices, the same result is likely to apply to US gasoline prices. The problem with forecasting the change in gasoline prices is not so much that we do not understand its economic determinants, but that it is difficult to predict the future evolution of these determinants. The fact that many media pundits disagree on the future course of gasoline prices reflects to a large extent the fact that they disagree on the future evolution of these determinants. It is this uncertainty that renders the no-change forecast of gasoline prices a good approximation.
Conjectures on the future price of US gasoline
It is instructive to speculate about the future evolution of the determinants of the price of gasoline. Abstracting from unpredictable refinery outages, the future evolution of US gasoline prices will depend primarily on developments in the global crude oil market.
Although past oil price increases have been followed by substantial increases in crude oil production with a delay of a few years, there is reason to be sceptical of the idea that substantial increases in oil production will be forthcoming in the foreseeable future. This is not a problem of the geological scarcity of oil. The problems are:
Oil exploration had been neglected, as the price of oil fell in the late 1990s.
The political environment in many oil-producing countries discourages oil companies from making the much-needed large-scale investments. In particular the threat of expropriation of successful investments in many countries prevents investments from taking place at the needed pace.
Additional crude oil likely to be available in the short run is heavy crude oil that US refineries are ill equipped to process. Building new refineries in the US takes many years.
Thus, a fair presumption is that the crude oil market will remain supply-constrained for the next few years.
With supply constrained, demand growth is the key
With crude oil production remaining flat or increasing only slowly, the price of crude oil and hence the price of gasoline will depend first and foremost on the extent to which countries in emerging Asia will continue to grow. Clearly, the current expansion in Asia will not continue unabated – all the more so as rising energy prices will leave their mark abroad while the US economy is already slowing. While a decline in demand seems inevitable, what is not clear is how soon that decline will occur and by how much global demand for industrial commodities will slow.
If past global expansions are a guide, global demand will recede only gradually. This is a direct implication of the model underlying this analysis. This suggests that US gasoline prices will remain high for the time being. Barring a major economic collapse in emerging Asia, prices will stabilise only as the world economy learns to economise on the use of oil and gasoline and as the supply of crude oil expands. Both corrective forces will take time to gain momentum.
In addition, there is reason to be concerned that oil-market specific developments, which for the most part have played no role since 2002, could become more important in the future.
Oil-market specific developments
Sharp shifts in the precautionary demand for oil reflecting uncertainty about political developments in the Middle East tend to occur only when demand for crude oil exceeds supply. When they do occur, they tend to cause sharp increases in the price of crude oil, as in 1990/91, for example. Under the current conditions, the world economy is particularly vulnerable to threats of military conflict in the Middle East. A good example is Iran’s threat to close the Straits of Hormuz if Iran were to be attacked by Israel. Such developments could potentially cause US gasoline price movements that dwarf the effect of sustained strong global demand for industrial commodities.
In short, based on this interpretation of the evidence, the price of crude oil and hence the price of gasoline is likely to stay high for the foreseeable future, but there is considerable uncertainty in either direction.
Lutz Kilian is Associate Professor at the Department of Economics at the University of Michigan and a Research Fellow at CEPR.
1 Also see Kilian, L. (2008b), “Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks in the Crude Oil Market,” forthcoming: American Economic Review.
2 This article is based in large part on results in Kilian (2008a), “Why Does Gasoline Cost so Much? A Joint Model of the Global Crude Oil Market and the US Retail Gasoline Market,” CEPR Discussion Paper No. 6919.
3 See Alquist, R., and L. Kilian (2008), “What Do We Learn from the Price of Crude Oil Futures,” mimeo.
4 See the Vox column by Guillermo Calvo, and the reply by Paul Krugman.