The Pass-through of Oil Prices to Gasoline Prices
Changes in the price of gasoline, particularly in the last few years, have been closely watched by consumers. Gasoline expenditure is a substantial part of the average household's total consumption expenditure, ranging from 2 percent to 5 percent since the late 1970s. Moreover, the share of household expenditure that must be devoted to gasoline is affected by changes in the relative price of gasoline, tending to rise when gas prices spike, because it is hard to adjust the quantity of gasoline consumed, especially in the short run. Economists say that the demand for gasoline has a low price-elasticity of substitution. In other words, changes in gasoline prices have a strong impact on the consumption of other goods and services as well as of gasoline.
The single most important factor affecting the price of gasoline is the price of crude oil, which accounts for roughly half of the price of a gallon of gas at the pump. About 45 percent of the oil refined in the world today winds up as gasoline, which makes it the primary product of the downstream oil industry. The remainder of each barrel of oil yields byproducts like jet fuel, kerosene, heating oil, and diesel. After the cost of oil, a substantial share of the pump price comes from federal, state, and sometimes local taxes, refining margins (that is, the costs and return of refiners), the retailer's markup, and distribution and marketing margins.
The ratio between the price of a gallon of average-grade gasoline at the pump and the price of crude oil per gallon - which we refer to as the gas-oil ratio - trended up from the late 1980s until 1999 but has been trending down recently.
Most of the change in the ratio over time has been caused by the effect of taxes. An excise tax is a given tax per unit, which means that when the gasoline price goes up the tax rate falls and vice versa. Notice that the highest tax rate, 60.7 percent, occurred when oil prices were at their record low of $12.01 per barrel in February 1999.
However, taxes cannot account for the entire path of the gas-oil ratio and, in particular, for the recent downward trend. This trend might be explained by a compression of the margins in the downstream oil industry when oil prices are high. In particular, refiners - who contribute 10 percent-30 percent to the gasoline price - have been shrinking their margins on gasoline products, reflecting both a higher oil-gasoline transformation rate (thanks to technological progress) and lower returns on gasoline.
Lower returns could be due to higher competition in the industry or to the fact that companies have been trying to absorb some of the recent upward trend in oil prices rather than pass it entirely on to consumers. By doing so they hope to prevent a substantial change in the future demand for gasoline (its long-run price elasticity is higher than the short-run elasticity because consumers have more time to adjust to any price change). However, this tactic will be sustainable only if the trend in the price of crude oil reverses.
A glance at figure 3 above will confirm that the price of gas and the price of crude are highly correlated. But to determine how much of an oil price increase is passed on to the gasoline consumer, we need to look at the "oil-price pass-through," which refers to the effect of changes in oil prices on changes in gas prices.
When we calculate the effect of contemporaneous and past changes in oil prices and gas taxes on the retail price of gasoline from 1986 to today, we find that, on average, less than half of an oil price change is passed to consumers. The time that it takes is relatively short; it passes through within the same month of the oil-price increase or in the month month after; on the other hand, changes in excise taxes do not appear to have a significant effect on the price of gasoline.
A casual observation of gas price changes over time suggests a remarkable change in the volatility of gasoline price after 1998 (the sample variance triples after 1999). Because the two periods differ so markedly, we redo the calculation of the pass-through, this time splitting the sample into two subsamples, one pre-1999 and one post-1998. In the earlier subsample, the pass-through is much lower and slower, amounting to about 30 percent over the course of two months. Furthermore, changes in taxes have a significant effect before 1999. In the more recent period, there has been a dramatic increase in the pass-through. After 1998, about 72 percent of a change in the price of oil passes through to gasoline consumers within a month's time. If one looks at the pass-through before the effects of taxes are added to the calculation, the pass-through amounts to about 96 percent!
What the results from splitting the samples suggest is that the higher volatility of the gas price series could be attributed to a higher pass-through from oil prices. In fact, even if oil prices have not shown any particular increase in their volatility, the transmission of crude price fluctuations to gasoline prices has changed. The downstream oil industry is no longer smoothing fluctuations in the price of crude for U.S. households and it is no longer guaranteeing relatively stable gasoline prices. This could be due to more compressed margins within the industry (a hypothesis not really supported by the data) or by limits in refineries' capacity (capacity utilization has been averaging above 90 percent in the past 10 years).
Finally, we recalculate the pass-through to see if positive and negative changes in oil prices affect gas prices the same way, and whether their pass-throughs have changed over time. We observe that in the pre-1999 sample, a decline in the price of oil had no effect on gas prices within one month, while in more recent years, a price decline has a strong pass-through of almost 50 percent within the same amount of time. In fact, after 1998, 95 percent of a decline in the price of oil passes through to gasoline prices within a month - this figure is about 100 percent after taking taxes into account. However, we also find that after 1998, the pass-through of oil price changes is often erased at the pump after about five months. This is true especially for oil price declines, which show a strong reversion effect at about five months. Oil price increases also show a somewhat weaker reversion effect. In other words, an initial reduction in gasoline prices due to a reduction in the price of crude oil will not last for long, unless the oil price reduction is sustained.