Miles Driven Revisted: Score One For Cyclical

August 19th, 2011 at 5:51 pm

In my earlier post on this question of the historically unique change in the trend miles driven, I wondered whether this change was cyclical or structural, the former meaning you’d expect things to pretty much snap back when the economy “stabilizes.” 

Here’s a point for cyclical.  I created a simple statistical model to explain the (log change in) miles driven series shown in the link above, based on unemployment and gas prices.  I ran the model through 2007 and predicted the series forward using actual values for unemployment and gas prices.

Sources: National Highway Admin, BLS, my analysis

As you can see, the forecast series (miles_f) hugs the actual pretty closely, meaning movements in joblessness and gas prices broadly explain the movements in series.  Had a structural shift occurred, the forecast would be less accurate.

As I wrote the other day:

“I’ve generally been skeptical of arguments about “the new normal,” thinking that much of what we’re going through is cyclical, not structural, meaning things pretty much revert back to the old normal once we’re growing in earnest again.  But it’s worth tracking signals like this that remind one that at some point, if it goes on long enough, cyclical morphs into structural.”

That last part is worth repeating…it will be a while before the economy “stabilizes”—before unemployment comes down to normal levels and family balance sheets are back in the black.  Relationships between economic variables like these and many others could start to change—and that’s worth our attention.

Print Friendly, PDF & Email

7 comments in reply to "Miles Driven Revisted: Score One For Cyclical"

  1. tom says:

    If you were really able to ‘forget’ what you knew was coming after 2007, and not let it impact your forecast, that is a really, really good model.

  2. pollian says:

    You do incredible work…everday. Glad you’re devoting yourself to influencing public discourse.

    One question about the incredible work. What programs do you use for your data, graphs, and modeling? Krugman is always talking about Fred (at the St. Louis Fed). What’s your go-to toolset?

    • Jared Bernstein says:

      Thanks! I use E-views for econometrics…best times series package on the planet but not cheap. I use FRED too but mostly source data sites: BLS, BEA (NIPA accounts), Census, CBO, Energy Information Agency, Treasury, and a ton of stuff my CBPP colleagues whip up. I think I use FRED mostly for financial data–interest and currency rates, eg.

      The extent of data one can get for free these days–as in all those sites just noted–absolutely blows my mind. I can literally build a data set in a half hour that used to take a week at least. With all these crazy budget cutting going on, we risk undercutting our statistical agencies. Which is fine if you don’t want to know about the impact of all of the cuts you’re presiding over…

      • pollian says:

        Thanks so much for the quick response. It’s not just the availability of free datasets that has made such a difference to my own intellectual development. It’s the willingness of people like you to share their practices of knowledge.

  3. mle detroit says:

    On the other hand, some may be structural:

    “I was in our front yard last week when the Google Street View car went by. Apparently they are updating their database with higher resolution photos…

    “In some countries objections have been raised based on privacy concerns, but here in Madison elected officials are very supportive, for a surprising but logical reason:

    “I’m very glad to hear they’re getting new data,” said Dave Davis, Madison’s geographic information system manager for the engineering department. City staff rely extensively on Street View to show them such things as fire hydrants, traffic signs and sidewalks without having to leave the office, he said, saving trips and time.”

  4. Jim Horak says:

    I caught myself driving down the road trying to find the best gas prices until I almost ran out of gas and than after filling up would get mad when I saw that gas cheaper at the next exit. Yes, you have done it too! I thought it was about time to do something because I was absolutely getting sick of paying over $3.00 a gallon at the pump.
    I decided to do some serious Goggling in my quest for lower gas prices; I came across this unique international discount gas card club. After researching all the benefits, I immediately joined the membership club which is much like AAA Insurance where I enjoy discounts on prescriptions, cellular phones, travel and roadside assistance, plus I now also received weekly discounts on my gas purchases. The money we saved comes in handy for more important things.
    I think we can win the war against high gas prices by sharing my research with our community. You can choose to join me or keep on complaining about high gas prices and see if that works for you. ( Click on my Face Book icon on this web site to friend me.)
    FYI, I also found additional gas saving information; actually you can eliminate your total out of pocket expense for gas. LEGALLY!

  5. Gary Haubold says:

    Hi, great chart.

    One observation: it looks to me as if the model became biased high in late 2010 or at least early 2011.

    Have you updated the model forecast on the out-of-sample miles driven since August 19, 2011? Just wondering if the model is still biased high, or if the bias has gotten higher, which is what I’d expect.

    I’m not sure, but I’m thinking that there’s a demographic effect that’s impacting miles driven along with labor participation rate (which impacts miles driven as well).