Oil price increases and decreases seem to have asymmetric effects

From an interesting 2003 review article by Jones, Leiby, and Paik (pdf):

The energy economics literature has noted the asymmetric responses of petroleum product prices to price changes for well over a decade, as observed by Balke, Brown, and Yücel (1998) in a review of previous studies.  Product prices rise more quickly in response to crude price increases than they decline in response to crude price reductions.  Using weekly data on crude prices and a variety of spot and whole gasoline prices, BBY (1998) find considerable support for asymmetry in the time pattern of downstream price changes to changes in upstream prices, although they find that different specifications of asymmetry yield different results.

Applied to the crude-product relationship, asymmetry has a different meaning than it does in the oil price-GDP relationship. In the crude-product relationship, the asymmetry is in the speed of the response, while in the oil price-GDP relationship, it is in the magnitude of the response. Competition will ensure that the magnitudes of the response of product prices to crude price changes are eventually equal. Otherwise profits in refining and distribution would grow without bound.

Here is a JSTOR link to a somewhat later Balke, Brown, and Yücel paper.  Here is their 2008 paper (pdf) on why the oil price/gdp link has weakened in the United States.  Here is a related 2010 paper (pdf).  Here is a recent James Hamilton blog post on oil gluts.  Here is Scott: “Focus on Q, not P.”


The German expression for this exceedingly obvious reality is that gas prices rise like a rocket, and fall like a feather.

And for the less cynical commenters, do keep in mind that governments also collect more revenue from higher priced petroleum product prices.

The effect may be obvious, but it is still interesting in that it is not clear why this would be so.

A dozen years ago, a friend who drives 100 miles per day asked me why gasoline prices go up immediately on bad news and down slowly on good news. It seemed like I should have been able to figure out why, but I couldn't. Does anybody know?

Prices at the pump fall more slowly because everybody downstream from the well makes more money when crude prices are falling because margins open up.

At the local corner with 2 to 4 stations on it, the market for wholesale gasoline is a world market. The cost increase come to you immediately, but the market you sell in is very local. As prices rise, retail margins on that corner get squeezed as the four managers are reluctant to lose business by raising prices. So for a while they take reduce margins on their gas. Eventually on of the managers will raise the price, the other three breath a sigh of relief and raise their's too. It is easy to see the coordination of raising the price working in everyone's favor on that corner. So prices go up pretty easily and quickly.

It is just as clear why price declines take more time. At that same corner with wholesale prices declining, the managers are now making a better retail margin on their gasoline. And it is still that very local market. So they all watch each other to see who's going to be the first to lower prices. They are all happy to have the improved margins so they will resist lowering the prices as long as possible. Eventually someone on that corner, or at the next intersection, decides he needs/wants more volume and will lower the prices. But each round of price cutting takes longer because of the improved retail margins with falling wholesale prices.

Just put in a call to QuikTrip or RaceTrac for confirmation of my story.

Maybe I'm missing something, but doesn't this tell us what is occurring, but not why it is occurring? Managers respond quickly on the way up but slowly on the way down. That just moves us one step up the ladder.

It's counter-intuitive because it seems like the first-to-raise-prices is hurt on the way up, while the first-to-drop-prices is rewarded on the way down. So it seems like there would be pressure in the other direction. But, and I believe the literature supports this, local gas markets aren't terribly efficient. People are considerably more reluctant to cross a street or make a u-turn to fill up than you would expect.

So it actually works like this: People are reluctant to change gas stations, even in response to prices. So the early price raiser doesn't see much hit to his business and the early price dropper doesn't see much gain. I.e., the pressures are not symmetric. That's why one direction of change happens quickly and the other slowly.

I wonder if any gas station owners read this blog?

I suspect that there are some terms in the contract about pricing and price changes that have some influence here. One of the things that will have some impact I would think is the timing of the various deliveries.

Other factors would certainly be implicit colusion allowed by the global market signals along with profit motives What most seem to focus on) -- and the fact that the gas delivered to the station after the global market price is dropping will likely be the higher prices gas from the earlier price rise.

In short, given the assymetry I'd look a contract terms that make gas a non-fungable item in an intertemporal sense. Purely hypothesis as I don't own a gas station of know the terms of either the franchise or indepentant operations with the refiners and wholesalers.

Why not just assume a simpler and less cynical model where people do nothing unless they really have to. After all, in day to day work people focus on the immediate priorities. A raise in supply costs is an immediate concern while the potential loss of sales resulting from a price increase is a problem for tomorrow. A drop in supply costs means higher margins and any intent gets pushed far down the todo list.

Off-topic, here's a pretty funny James Stewart article on billionaires buying modern art:


At what price per barrel should we try converting the paintings back to oil?

Of course. The US and Saudi Arabia have kept the oil price artificially high for decades.
As everyone is forced to use oil for dollars a higher oil price means more dollars in circulation, so nominal dollar prices are higher.

I'm guessing. The first sentence is true, but the second sentence could be nonsense.

Oil prices were pretty darn low in 1998.

Ail travel price. Tickets go up if oil rises 20% but never go down if oil descends 30%.

The asymmetry may arise from medium term fuel "frozen price" contracts. People is afraid of variability and make contracts to buy at the same price for a year or more. The lower price now only changes the contracts that are being signed now, not all the previous ones.

On the other side, increases are reflected almost immediately on options or future oil contracts.

This doesn't really explain the phenomenon, though. Frozen price contracts should simply add stickiness to prices, whether oil goes up or down.

Consumer sentiment? Risk aversion?

Going back to my air travel example, if you reduce prices for a few months while fuel is cheaper you make customers happy and gain in cabin occupation at the expense of not giving the savings from cheap fuel to investors. You make investors unhappy.

When fuel goes higher a few months later your customers may get upset because you will be rising fare prices. The fare rise may cause market share loss that make investors unhappy once again.

If you leave prices just like before, you don't risk any of the previous problems and make cash. Also, you're in better position to offer fares discounts to kill your competitors that made discounts while fuel was cheaper.

The issue here is that cheap fuel only last a few months. In 2009 the 40 USD per barrel was a few months long. Average 2009 price was around 80 USD per barrel. Who will risk to make discounts for an ephemeral market share gain?

What are you talking about? The real price of oil hovered mostly between $25 and $40 a barrel from 1986 to 2003 with only occasional bursts above and below that.


The inflation adjusted data you linked says we're in a 10 year occasional burst above the 40USD historical average ;)

I've had some scary thoughts that when oil does go back to the $100 per barrel, gas at the pump will be double what it was when previously at that price per barrel, based on this asymmetric pricing effect.

Like the article was saying, that would make it a great time to get into the refining business!

There is a lot of literature on asymmetric pricing.

Among the factors: transaction costs for menu or price changes (it actually costs money to change a price in the grocery store...so overshoot on the list price if you think the price will go up again in the future) and just offer coupons until it does; coordination problems in oligopolistic industries where members have different input costs and input contracts (stagflation)...coordination is difficult going down, and it is also difficult going up, so once a price is set there are coordination costs from moving from it unless there is a formula (rule of thumb) tied to input costs that one can easily follow (again, it is more risky going down unless there is a clear leader and excess capacity); there is also some behavioral econ issues as well: we notice a price increase more than we notice a price decrease if we habitually purchase an item, but less so if we purchase less frequently; another behavioral econ observation is that elasticity effects are greater going up than down (Putler's (1992) egg study--consumers much more responsive to price increases for commodity eggs than to decreases).

This is a very well-known phenomenon. Gas station owners "feather down" prices to take some windfall profits after a rapid decline in the wholesale price of gasoline but raise prices immediately if they anticipate that the overnight cost of replacement will rise.

Should probably also mention that station owners almost never make a profit on their gasoline otherwise. During normal periods many will make a loss. Gas station profits come from their convenience stores.

Do you have a reference for this?

Gas stations seem to have a lot of direct competition. Multiple stations are often adjacent to each other, they display prices prominently, and they are selling a commodity product, for which few consumers prefer a particular brand. I don't see how they can keep prices artificially high, in general.

See my note above...

brickbats and adiabats is correct. I used to finance retail petroleum outlets. In a stable price environment, the retail margin on gasoline hardly covers the cost of equipment, utilities and labor. The possible exceptions are the very large and well run stations (QuikTrip and RaceTrac are the ones I would point to here in the Dallas area). Most of the money is made inside.

I would bet that most Marginal Revolution readers would not believe this because they just buy their gasoline by paying at the pump. The typical MR reader is not going inside to pay for gas inside, get lottery tickets, cigarettes, and either breakfast or lunch. Many people do, however.

Here's a reference, which refers to US EIA documents:

There is always some implicit collusion going on; if you think about it, it's a classic iterated prisoner's dilemma: all gas station owners will realize that wholesale prices of gasoline will decrease at about the same time and that they will be better off if no one defects and suddenly lowers prices by too much. Competition wins out in the end because the prices do bottom out near the breakeven price for each station, but for a short period of time there is significant implicit collusion.

I own several gas stations, and a reasonable rule of thumb is that 3/4 of gross sales are gas, but close to 3/4 of profits are from inside sales. But in some locations gas margins can be high.

I'd point to gas stations alongside major arteries into New York City as a good example. They charge $.50 to $1 per gallon more than streetside outlets because of their location directly along the main commuter routes, with just enough people desperate enough to take the deal. Also TC once cited full service gas stations in places where self-service is the norm, which figure that anyone who goes to a full-service station in a predominantly self-service area can eat the premium with no worries.

Yes, a well-known phenomena, and well explained by consumer search behavior.

See Matthew Lewis, (2011), Asymmetric Price Adjustment and Consumer Search: An Examination of the Retail Gasoline Market. Journal of Economics & Management Strategy, 20: 409–449. doi: 10.1111/j.1530-9134.2011.00293.x

ASTRACT: This article proposes a new explanation for why retail prices respond more quickly to cost increases than cost decreases. I develop a search model that assumes consumers’ expectations of prices are based on prices observed during previous purchases. This model predicts that consumers search less when prices are falling, which results in higher profit margins and a slower price response to cost changes. I then empirically examine patterns of retail gasoline price response and price dispersion to show that this model predicts observed price behavior better than previously suggested explanations.


Yes - this is Kahneman & Tversky stuff. People will go to great lengths to forestall losses (in this context price increases) but are not as assiduous about gains (price decreases in this context).

Very good comment.

Perhaps a different explanation (but maybe not -- could be rolled in to the whole search concept) might be marginal value of the moeny saved. Which prices rise and the buget constraints are biting into expected consumption the marginal valuation one the foregone consumption is rising. When a relative price is falling the marginal value of the dollar saved is declining. The marginal value of the time lost in searching will be related but not entire defined by the prices as it's also reflecting non-prices leisure activities so would seem to always be increasing if more search is required.

One possible mechanism may be as follows: immediately after a price drop, sellers are still sitting on inventory with a high cost (which they have possibly borrowed to acquire). Even though they can replace that inventory with low-cost gasoline, they still wish/need to recover that cost to the extent possible (and if in a low-margin part of the business, which is everyone except for the mineral rights holders, this cost recovery can be critical). So, even though input/wholesale prices have dropped, sellers have an incentive to keep prices high (with the cost of defecting potentially being a loan default and/or going out of business). This, however, is short term as the cost of defecting disappears as inventory clears.

For price increases, however, a similar mechanism creates a rapid price rise: even though you are sitting on low cost inventory, in order to afford to purchase the next cycle of inventory, you need to increase your prices now; if you defect and sell your inventory at low prices, you will not have the capital necessary to fully restock your inventory.

That is true and there are often differences between entities on what type of accounting they use: FIFO or LIFO. In some industries, there are rule of thumb measures as well: even though distribution costs are fixed, the seller pancakes--adds on--a fixed percentage margin based on input costs. So, even though operations costs do not increase, gasoline prices increase and the station marks up the price based on a percentage of wholesale. Then what happens is that competition begins to break out with advertising and promotion--expect to see more Exxon commercials on the quality of their system or gas--or they engage in softer competition--expect to see more cents off offers based on use of a credit card or affiliation with another marketer, like a grocery store.

Don't forget Sam Peltzman's paper, Prices Rise Faster than They Fall, The Journal of Political Economy, Vol. 108, No. 3 (Jun., 2000), pp. 466-502, available here: http://research.chicagobooth.edu/marketing/databases/dominicks/docs/2000-PricesRiseFaster.pdf

+1 Here's a more recent piece from an economist at the Antitrust Division and who also attended Chicago:

After looking into a handful of the articles on your site, I seriously appreciate your
technique of blogging. I saved as a favorite it to my bookmark website list and will be
checking back in the near future. Please visit my website too and let me know what you think.

Comments for this post are closed