The Fallacy of Average Gasoline Prices: Why Widely Cited Sources are Often Misleading

By Chuck Bonner

Every news outlet on TV, Internet, and in print tells you about movements and trends in the average price of gasoline. But how is the "average" calculated, and what does it really mean? This article will shed some light on that question, and show you why it matters, even though you never thought about it before.

You hear that news report every day, and you use it to make predictions about the future price of gasoline. There's nothing scientific about it, but you always try to develop your own forecast: If the average price today is "A," and yesterday it was "X," would I be better off buying gas today, or waiting until tomorrow, when it might be lower.

Of course, you know your own off-the-cuff prediction is not all that reliable. You're no expert, but you're trying to save a little money when you can.

But the big problem is that the reported "average" price of gas is almost meaningless. Let's take a look and find out why.

We all know what an "average" is. You take a set of numbers, add them together, and divide by the number of numbers in the set. So, if there are fifty gas stations in a given area, you record the price of a gallon of regular unleaded gasoline at each, add the prices together, and divide by fifty. Simple, right? But is it an accurate reflection of what you'll pay for gasoline?

Absolutely not! This method will not tell you the true average price of a gallon of gas. But more importantly, this false "average" price, analyzed over time in order to make predictions of the future price of gas, will not work.

I'll tell you why this "average" is not real, and I'll tell you a better way to analyze trends in the true price of gas, which you can use to develop a forecasting model that actually works.

Consider those fifty gas stations and their "average" price again. Most of us will go to the gas station with the lowest price. For our purposes, the other 49 gas stations don't matter.

Yet that simple average is the method used by all the major sources of gas price information. Go to the Web sites of the U.S. Department of Energy, the American Automobile Association, or any other well-known report and read about their methodology. Behind it all is a simple, "how many gas stations and what is their price?"

The real question to ask is, how many gallons of gas are sold at each price? This will yield a properly weighted average, because the gas station that sold the most gas has more of an effect on the true average price than the station that sold the least gas.

There are problems with that approach. First, the gas station managers won't give you that information. They're too busy running their businesses, and there is no value to them in giving you this information. Another problem is that the data set simply becomes too complex. Not all gas stations change their prices at the same time, so you have to ask how much gas a given station sold while their price was at one level, then how much they sold after they changed the price, etc.

Not practical.

But there is a way to get a meaningful approximation of the true price of gasoline. Each local market drives it, and it is plain to see in any area.

The most reliable indicator is the price of a gallon of gas at the second-cheapest gas stations. It sounds too simple to be true, and it might not stand up to any deep mathematical analysis, but it is information that is easily obtained, and it works.

It starts to make sense if you think about it. In any given area, there will be one or two gas stations that have extraordinarily low prices for a day or two. They may be selling gas at a loss in order to draw in customers for their convenience store, or to attract a following of what they hope will become loyal return customers. But these cheapest prices are out of the norm, and not a good indication of the real average price of gas.

On the other end of the scale, the more expensive gas stations are automatically driving away most of their potential customers. Yes, they have a core clientele of rabidly loyal customers, but they are not selling much gas.

But if you look at the range of prices in any given area, you will see that most gas stations cluster around one particular price, which is much lower than the most expensive gas stations, and only a little higher than the cheapest ones. That's the price to watch.

These not-quite-cheapest gas stations are the ones that are selling the most gas in their market, and the gas stations with the lowest prices today will go back into the fold tomorrow, leaving another gas station to sell at a loss tomorrow.

So, ignore the news reports of the "average" price of gas in your area, or nationwide. Look at the second-cheapest gas station in your city, and you will have a more realistic indication of the true average price of gasoline.

What do you do with this information? Watch the trends, come up with your own moving average, compare it to the price of gasoline futures on the commodities markets, whatever you like. But whatever your chosen forecasting algorithm, it will probably work better if you ignore the "average" and use the second-cheapest.

Chuck Bonner is the lead analyst at GasPredictor.com His technique, based on the second-cheapest gas stations in a given area, has accurately predicted the trends of gas prices throughout the strange roller-coaster conditions of 2008. Get the latest forecast for gasoline prices at http://www.GasPredictor.com

Article Source: http://www.GasPredictor.com