The rate of change in the price of gas is rising

In my last post I was a bit surprised by the increased spikeyness in gasoline in what I had dubbed the "new trend" starting around 2001. One way to look at spikeyness is to look at changes in prices over time (in this case, the rate of change from week to week).

This picture is the week to week change for the national average price of a gallon of gasoline from 1994-present. As with the prior two posts, the chart is built using data available at www.eia.gov.

In this chart, data points above zero mean the gas prices went up that week. Data points below zero mean that gas prices went down that week. Data points around zero mean gas prices were largely unchanged that week. With that in mind, some observations:

1. From 1994 to 1998 gas prices were "mostly stable". You can see year over year trends hovering around the zero mark. While large spikes upwards occur, these are eventually offset by more prolonged but lower downward spikes afterwards (prices rise fast, but lower slowly).

2. Peak prices (highest spike points) often occur in April. So, I suppose there's a reason why the urban legend that started me off on this road happened across three Aprils (1997, 2011 and 2012).

3. To my surprise, the new trend I talked about didn't start following the 2001 recession. It began in 1999. Starting in 1999, most of the data points are above zero. This suggests that the price of gas began - and continued - rising.

4. The prolonged downward trends in gas prices are generally associated with recessions (2000-2001 and 2007-2009).

5. Also, to my surprise, the "new trend" isn't linear, it's curvilinear. Beginning in 1999 you hardly ever see prices hovering around the zero mark. There are short stretches of unchanged prices (April to August 2010 for instance), but in general, a trend line for price change would be somewhere around the 1% mark. Basically, since 1999, the rate of price increase has itself been increasing by about 1% a week. This has been punctuated by brief (but large) spikes and longer (but lower) price drops. 1% a week doesn't sound like much, but that's a pretty substantive annual change...

6. We actually had a pretty nice downturn in prices between April 2011 and December 2011. I'm not sure what's happening there, but that's a prolonged downward price movement during an economic recovery (albeit a slow one).