I paid $2.24 per gallon this morning to fill up my Elantra. It’s bizarre to see a total bill for a fill-up under $25. The truth is that gas prices have decreased remarkably over the past month or two, despite a major hurricane in the gulf, a full-scale (and global) economic crisis, and steadily decreasing supplies of crude since August. Naturally, anyone who thinks about this for more than a few minutes is going to at least wonder if there’s any possibility that the gas price bonanza has anything to do with presidential politics.

The reasoning is obvious – higher gas prices hit people in the pocketbook in a direct way. It makes them pessimistic on the economy, and resentful of Big Oil and their GOP allies, so voters will retaliate by voting for the Democrat. Lower gas prices, however, make people feel good about the economy, optimistic about life, and generally disposed to reward the People in Charge for doing such a great job – leading to votes for the Republican. This sort of reminds me of the Shoe Event Horizon.

For some people, it’s simply accepted as given that there is some manner of conspiracy afoot, usually involving Bush, Cheney, Halliburton, the Saudis, and whatnot. It’s not entirely unreasonable, given that Prince Bandar allegedly (according to Bob Woodward) assured President Bush that the Saudis would attempt to reduce oil prices prior to the 2004 election. Somewhat more credible is the passive conspiracy theory, whereby the oil companies simply make an independent decision to forgo short term profit and “invest” in the Republican campaign by voluntarily lowering prices. This sort of contribution would also nicely circumvent campaign finance laws since no direct donation is taking place.

It’s certainly an appealing theory, and with an election only three days away, paying the lowest price for gas in years lends the speculation more credence. But what does the actual data show? There’s a public DOE database of oil prices over time (note: averaged nationwide), and other bloggers have used it to attempt to see whether prices actually did decrease prior to presidential elections or not:

I decided to tabulate the price behavior of gasoline stretching back
over the past three presidential elections. I chose to track the price
from the beginning of summer driving season – Memorial Day – until the
first part of November when the elections take place.

The results are shown below:

[table – see original post]

Personally, I think one would be hard-pressed to find a pattern
there. The biggest price drop happened in a non-election year, albeit
it was an anomaly caused by 9/11. Of the thirteen years recorded,
gasoline prices fell between Memorial Day and November during nine of
the years. This is what I generally tell people: Prices fall for
seasonal reasons, and do so even when there are no elections. The
reason prices fall is that demand for gasoline falls after the summer.
The price generally peaks in early summer, and following Labor Day in
early September the price falls. (The details of why this generally
occurs was explained in The Transition to Winter Gasoline).

Of the presidential election years, the price fell in 1996 when
President Clinton was running for reelection, was essentially unchanged
in 2000 and 2004 when President Bush ran against Al Gore and then John
Kerry, and will almost certainly fall this year as oil prices pull back
from their record highs.

In fact, if you take out the major anomalies on the graph – the
slowdown caused by the 9/11 attacks, and the 2005 run-up of price in
the wake of Hurricane Katrina, followed by easing in 2006 as refineries
recovered, the truth is that gas prices usually don’t change
dramatically between May and November – election year or not.

However, as one commenter to that post points out, the total price change between Memorial Day to November obscures the weekly price fluctuations that might give more information about trends. That commenter used the same source to extract the weekly data for the presidential election years only and came up with the following interesting graph:

Note that the 2008 data is on a separate axis (albeit same scale) because the prices are so much higher. It’s clear from this data that for the previous three elections (96 Dole-Clinton, 00 Bush-Gore, and 04 Bush-Kerry), gas prices tended to drop over the summer, but also recovered just prior to the election, which i the opposite of what you’d want to manipulate – you’d expect if there were indeed collusion that the price woudl stay high all summer and then suddenly drop as you get closer to the election to give voters a feeling of real change just prior to voting day.

Which, however, is exactly what the graph does for 2008 (a spike thanks to Hurricane Ike aside). The data is not complete because the last few weeks are missing (the analysis above dates to early October). But from the EIA database, the week of 10/27 had an average price (for standard, non-reformulated unleaded gasloline) of $2.59/gallon, which is off the scale on the chart above for 2008.

Now, keep in mind that for the conspiracy theory to be true, the price would rebound immediately after the election. After all, once the votes are cast, why sacrifice the (theoretical) profit? So, again looking at historical data, what happenned to the price after the elections?

Summary – gas prices followed the same general trends after the election as they did prior to it. In other words, if there was indeed any conspiracy afoot, it was remarkably incompetent in its execution.

However, the data from 2008 is a major outlier, and we aren’t at the election yet let alone two months out. So, the big question is, what will the graph of post-election gas prices look like for this year? If the prices stay down for a while, then the suspicious behavior of the prices pre-election can be justified as normal. But if the price of gas rebounds this year, especially if that rebound occurs immediately post-election over the spane of a week or two, then I think that it will make a compelling argument that this year, for whatever reason, there was indeed something afoot.

So, was there a gas price conspiracy? As far as previous elections, it seems unlikely. The jury is still out for 2008 however. we will have to wait and see.