Friday, May 18, 2007

Affordability of Gasoline 2: Combustion Boogaloo

About a year ago, I posted a graph showing the affordability of gasoline as a percentage of household income, broken out by income quintile. The main point was that, although the price of gasoline has certainly risen dramatically in recent years, it still does not consume as much of our income as it did at its height in the early 1980s. The reason is that household income has grown since then, even for the poorest Americans.

To construct that graph, I arbitrarily assumed each household would consume 1000 gallons each year. I based that assumption on nothing in particular, because my only purpose was to show the pattern of change. I could have chosen 500 or 2000 gallons and the picture would have looked the same, just with a different vertical scale.

But as someone pointed out in the comments, people don’t really want gasoline per se; what they want is travel. And this matters, because fuel economy has risen substantially over the time period in question, meaning that people can travel the same distance with fewer gallons. So I have now reconstructed the graph, using more recent data, on the assumption of 25,000 miles of travel rather than 1000 gallons of gas.

(Income figures are from here, gas prices from here, and fuel economy figures from here. I estimated mean-income-by-quintile for 2006 and 2007 by assuming the growth rate of the prior five years would continue, I estimated the 2007 gas price using the first quarter data, and I assumed fuel economy remained constant after 2005.)

The assumption of 25,000 miles is still somewhat arbitrary; I based it loosely on the number of miles driven per household in 2001, which was 23,100. Again, the arbitrariness isn’t terribly important if our purpose is to show the pattern. And the pattern is clear: even for the poorest Americans, it’s much cheaper to travel now than it was in the early 1980s.

One factor this analysis doesn’t account for is the number of miles traveled per year. As it turns out, this number has risen steadily (see the fuel economy link). However, it’s difficult to tell how much of the increase is due to a greater need to travel (because of longer commutes, for instance), and how much is driven by a greater desire to travel (maybe because of the lower cost of gasoline relative to income!). So I decided it was sensible to focus on the cost-relative-to-income of a fixed travel distance.

Another factor not included is how travel distances and fuel economy could differ by income quintile. The lower quintiles might have fewer cars and not travel as much (or use public transport more). On the other hand, they are likely to have older cars with worse fuel economy.

Nevertheless, I think the picture is useful for perspective. We’ve recently heard about gasoline prices hitting a new record. That is, of course, a record for nominal prices, not inflation-adjusted prices. But that correction is not enough, because even if the inflation-adjusted price hit a record high, it would still overstate the effect of gas prices on our pocketbooks relative to historical highs.


BigZwey said...

I agree wholeheartedly that this approach is the most helpful when discussing 'rising' gas prices. However, it is not helpful to compare the current situation primarily to the late 70's or the 80's. The former was a terrible time for energy and world politics, the latter was a cultural bubble. I would like to see the graph extended to the 50's, or even the 20's. It would give a better sense of the long term changes.

BigZwey said...

I just noticed your use of the term 'historical highs'. I don't think this is a useful measure of the goodness or badness of the present situation. For instance, present mortgage interest rates are far below historical highs of 15% (I think), yet rates above only 10% would be quite painful for us now. Compare to post 9/11 rates of 4%, we are in pain now, but that comparison isn't helpful either since those were abnormally low rates. So the pertinent factor can't be historical highs or lows. Rather, perhaps recent trends are important, or perhaps comparisons to historical averages. Please explore gas prices relative to these benchmarks. Even your graph shows that fuel cost has become half again more painful in the last decade, as a % of income. I'm still curious to see your graph stretch back to the 50's.

Anonymous said...

The "nice analysis" on gasoline/travel costs appears to have overlooked the increased burden of debt on households. According to 2006 data ca. 14 percent of disposable income is used to finance debt. If we take that into account (i.e., multiply disposable incomes by 0.86)I am not sure that we would get the same upbeat results.

Upward ticks in gasoline prices in our times are far from being harmless little side-developments. Reasons for this include (but are not restricted to) the stagnation of most real incomes and the growth in income inequalities since the early 1980s.

I wish Panglossian economics proved to be right after all, but I have serious doubts that it will.

Glen Whitman said...

Anon -- I'm not sure why you think the debt should matter. People are choosing to incur more debts, probably because credit is more easily available. That doesn't mean we should just ignore the income used to pay off those debts, any more than we should ignore the income used to purchase Kleenex.

As for your claim that real incomes have stagnated for many, you're flat-out mistaken. Real income has grown for every income quintile over the time period in question -- in fact, that's precisely why I get the results shown in the graph.

You're right that inequality has increased, but how is that relevant to the question? The graph looks at the quintiles separately. If you want to know what's happened to the poorest people, without being distracted by what's happened to the rich, just look at the top series in the graph. Increasing inequality does not mean that the poor have gotten poorer, only that their real income hasn't increased as quickly as that of the rich.