Bryan Caplan  

Posner Gets Behavioral

Imagine a Referendum on the Ho... Switzerland...

Here's something you wouldn't expect Richard Posner to write:

I wonder, too, whether the recent decline in U.S. gasoline consumption doesn't represent to some degree an irrational panic reaction. To take a huge loss on the sale of your SUV in a market that is depressed because so many other people are doing the same thing at the same time is unlikely to be justified by the gains from the improved gas mileage of the car you buy with the modest proceeds of the sale. Likewise, driving a substantial distance to save a few cents a gallon on the gas you buy is unlikely to be worthwhile. A recent article suggests that people fixate on the price of gasoline because unlike most regularly purchased items, such as food, gasoline is purchased separately from other items so that its price is not buried in a bill for multiple items.
Anecdotally, I think he's on to something. I too have noticed people searching harder for bargain gas. But as search theory explains, the whole reason to search for lower gas prices is the variance of price. If gas sells for exactly $10.00 in every station, there's no point searching. And while gas prices have increased, I haven't noticed any increase in the variance of prices. So there's no reason to search more today than five years ago.

Is my impression of a stable variance of gas prices incorrect? Am I mistaken to think that people are really doing more bargain hunting than they used to? Or do we need to give behavioral search theory its due?

Comments and Sharing

COMMENTS (12 to date)
Grobstein writes:

Income effects.

Dan writes:

Interestingly, your theory would predict LESS searching as prices rise, since searching brings the same benefit and is more expensive thanks to the higher price of driving around to find gas stations.

Because of this effect, arbitrage is more difficult, so price variance would probably be higher. But that effect obviously wouldn't dominate.

8 writes:

What Dan said. When the difference was a nickel at $1.20, it might pay to drive further for the $1.15 gas. But $4.00 versus $3.95 is a rounding error, especially since my credit card gives me 1% back. It's anecdotal, but I am seeing a stable variance in nominal prices, albeit a slightly larger range, rather than a stable variance in percentage terms.

Dan Weber writes:

Is the variance a constant (10 cents per gallon), or is it a percentage (5% of the cost of a gallon)?

I've noticed ten cent variances in the past few weeks. I don't remember that before.

Incidentally, I've heard that gas stations in New Jersey, in additional to be required to be full-service, are only allowed to raise prices once per week. Has this caused any gas shortages as people realize the inevitable price increase is coming and fill up early?

conchis writes:

Both the points I would have made have already been noted: depends on the balance between the negative income effect of higher prices (which means that saving money is more valuable) and the increased search cost (which makes savings harder to achieve).

"$4.00 versus $3.95 is a rounding error" is most definitely *not* the way to think about it - but I don't doubt that it's what lots of people do.

Brad Hutchings writes:

Reason Foundation's Out of Control blog had a link to this article (

Lord writes:

My guess is normalized variance is similar but the increase in price has increased the absolute variance. Do people really drive around doing this anymore when they can just check on the internet? I basically know where they are cheaper and fuel up when I need to and happen to be in the area, so although I no longer do so locally, it doesn't cost me anything to do so other than making sure I don't allow it to go too low.

Lord writes:

The other common misunderstanding is averages aren't very average. I know people that the gas savings alone makes the payments on their replace vehicle. Some people drive 60-80k a year. Some only 5-10k. Using 12-15k only means half drive more, sometimes much, much more.

aaron writes:

Wow, Posner writes something down-right sensible.

Of course we are, see this from CNN. People cutting back on mass transit, police, road repair and maintenance.

I'll add that I think people are driving less efficiently and communities are probably not managing traffic lights well, further driving down fuel efficiency.

I plotted gas consumption and vehicle miles driven one year change for Jan 07 to April 08. When the change in gas comsumption is bigger than the change in miles driven, fuel efficiency is declining. There are huge declines in efficiency in Feb 08 and March 08.

Interestingly, I also happened upon report (via futurepundit) that says gas consumption is down 1.7%, the first significant reduction in consumption in 17 years. In August 1990, oil prices spiked upward. We entered a recession in September 1990 until September 1991.

aaron writes:

Brad, anecdotally, I know someone who decided to try E85 in his flex-fuel Tahoe. He got half the miles out that tank compared to normal.

Brad Hutchings writes:

Aaron, That's the other half of the equation. So E85 is about half the price of gas right now. The typical gas mileage percentage I've heard is 60%. Ignoring maintenance, it might be a good deal when gas is over $4, but it gets economically dubious when a gallon is under $3. And if miles per tank (like long trips) is an issue, you might as well have an electric car.

There is new plant somewhere in the LA area (Palmdale?) that just got under construction, where the management claims to be able to create ethanol from waste for about $1.50/gallon wholesale. Workable when gasoline is over $3 wholesale, probably not otherwise... Man, are they proud of that plant this week! It made all the local news, and no reporter questioned the most basic economics!

David writes:

This debate is all the rage in Australia right now, where the government is introducing "Fuel Watch" - a scheme that requires every service station in the country to fix its prices every 24 hours and publish them on the web the day before. It cuts search costs and it should reduce variability but opinion is divided as to whether it will lead to lower, higher or no change to overall prices.

Comments for this entry have been closed
Return to top