Elastic Gas

Posted by Eric de Place

12/21/2007 01:30 PM

When gas prices go up, our consumption goes down.

I don't know about you, but I've been spending my holidays reading papers about the price elasticity of demand for gasoline. Wait, where are you going?

Listen, I'm sure you've heard this before: "It doesn't matter if we tax gas, people won't change their behavior." Or: "Gas prices go up and down, but people keep using just as much."

So here's what's interesting about my Yuletide reading: these claims just are not true -- not at all! Over the long run -- usually meaning longer than a year -- a 10 percent increase in gas prices results in a decline in gas consumption of around 6 or 7 percent. (This is what economists call the "price elasticity of demand," or just "demand elasticity" if they're feeling hip.)

Now, it's certainly not a one-to-one correlation, but the connection definitely exists and it's much stronger than most people think. Interesting, right? I'll get into the details below the jump.

And hey, is that mistletoe?

I guess I should mention right now that enough ink has been spilled on the demand elasticity of gasoline to fill volumes, and probably whole shelves. But don't worry, I'm not getting into every study here -- just a few meta studies that look at the results of hundreds of analyses.

In all of the examples below, the figures refer to a 10% increase in gas prices over at least a year:

In 1992, Phil Goodwin looked (pdf) at a large number of studies and determined that gas use declined by 7.3%.
In 1998, Molly Espey surveyed 101 different international studies and found that demand for gas declined by 5.8%.
A 2002 study by Graham and Glaister, also surveying numerous studies, found that gasoline use declined by 6% to 8%.
In 2003, Goodwin teamed up with some pals for a reunion tour (pdf), this time of 175 elasticity equations published since 1990 (about half were for the U.S. or Canada), and found gasoline consumption dropped by 6.4%.

You're hitting that drink pretty hard. Would you like another? No?

You know who I should introduce you to? Todd Litman. He runs the Victoria Transport Policy Institute and he's done a bang-up job of summarizing gasoline elasticity equations -- especially here. (And here too.) You'll find a range of credible estimates all the way up to more than a 10% decline in gas use (over the long term) when the price goes up by 10%.

The point isn't that we know precisely what effect higher gas prices have on consumption. We don't. Demand elasticity is affected by a huge number of factors, including the volatility of the price and the availability of substitutes for gasoline. (There are actually quite a few substitutes, even in exurban America, but that's for another post.) The point I'm making is that higher gas prices really do change our consumption. That old chestnut "demand for gasoline is impervious to price" is an urban legend I'd like to put to rest.

Before you go, there's a wonderfully short and friendly article on About.com (unfortunately, one of the links goes to the wrong paper). It's a great starting point.

By the way, there are a ton of caveats to this material, and there's something that works at cross-purposes to rising gas prices: rising incomes. When incomes go up, folks tend to spend some portion of that new money on gas (this is "income elasticity" to economists). The figures I've included in this post mostly don't deal with income changes. So my arguments should be thought of as referring to inflation-adjusted price changes.

Oh, leaving already? Okay, well, happy holidays.

Say, what are you doing for New Year's?

Permalink

Comments

Posted by sf 12/21/2007 03:31 PM

I don't have any citations handy, but I have, unfortunately, seen a couple of US only studies that were much more pessimistic, showing about 20% elasticity during the 70's and less than 10% more recently. Studies in Great Britain show a lot higher elasticity because income is lower, gas is much higher to begin with, and public transportation options are better.

Posted by andrew 12/21/2007 04:59 PM

Don't under-estimate your readers. I would bet that quite a few of us find this intrinsically interesting. :-)

Posted by Eric de Place 12/21/2007 06:54 PM

andrew,
That is precisely why I so love the readers of this blog!

sf,
Later on, when I have time to breath, I'm going to do a post about demand elasticity specifically in the U.S. in recent years. I've also seen some of the same evidence you refer to, but I don't believe it's the whole story.

There's also an interesting real-world example here in the PNW: for the last 8 years, total gas consumption has been nearly flat. That is to say, it's been declining so steeply on a per capita basis that it's being help aloft basically by population growth.

Posted by MichelleV.P. 12/21/2007 07:41 PM

For New Year's? Um...totally looking forward to your next post about demand elasticity specifically in the U.S.

...Will there be champagne? :-)

Posted by a-kos 12/21/2007 09:00 PM

Eric, I'm also anxious to read your article on this topic pertaining to the US. For the past few years I've been saying I'd love to see gas at around $5/gallon because this would motivate people to use much less of it.

Posted by Levin 12/22/2007 01:07 AM

The key is that prices have to stay high for a YEAR or more. If they simply spike up for a few months, then neither businesses nor individuals will buy fuel-efficient vehicles or adjust their fuel-consumption habits. We need a state or federal floor on the gas price, so people can plan ahead and make rational choices. Suppose there's a variable-rate tax that sets the gas price floor at a constant $4/gal. Whenever the actual commercial price (before the tax) drops lower than that, the government gets a windfall, which they promptly pump into a climate-friendly investment like renewable energy. If the gas-price-floor rises by 5% or so, every year or two, then we'll be weaned off petroleum before we know it, without suffering the economic chaos of unpredictable price spikes along the way. By making fuel prices predictable, the government could help businesses plan ahead, and budget for fuel efficiency.

Posted by eldan 12/22/2007 11:31 AM

I think Levin has hit the nail on the head here. Gas consumption is highly inelastic over the short term, because it takes time to either change a highly car-dependent lifestyle or replace a car with a more efficient model. But over the long term, people do factor the price of gas into their car choices, and the overall cost of their commutes into decisions about where to live.

I think that one of the reasons this effect has shown up more strongly in Britain is that for many years Britain had a "petrol tax escalator" which increased the gas tax by a predictable amount each budget, causing the price at the pump to go up faster than inflation every year. That provides the sort of predictable, consistent price signal which triggers the long-term elasticity we need to see.

Posted by barry 12/22/2007 08:09 PM

Excellent info, Eric. Thanks. I get this "carbon tax won't change behaviour...it will just be more money for government" argument from others all the time.

Glad to have the references to site!

Posted by David 03/07/2008 08:06 PM

Eric,

Great stuff, and not at all boring to the 0.1% of us who think marginal utility was a great
discovery.

After living through the reaction to spikes in gas prices in the Seventies (yes, I'm that old and actually was working in the US Dept. of Transportation at the time), I concluded that it isn't the magnitude nor the duration of a price increase that affects demand, it is the expectation of future price increases. Do you know anything from the current literature on this point?

As a result of reaching this conclusion (without challenging myself through any rigorous research), I came up with my favorite policy prescription: a gas tax that would automatically increase by 5% each year. Politically impossible, but I still liked it.