What Economic Models Can Tell You

Nat KeohaneThis post is by Nat Keohane, Ph.D., director of economic policy and analysis at Environmental Defense Fund.

Last Friday, the U.S. Chamber of Commerce sponsored a panel discussion on the economic impact of climate change legislation. I was on the panel, along with Bob Shackleton of Congressional Budget Office, Francisco de la Chesnaye of Environmental Protection Agency, Margo Thorning of American Council for Capital Formation, and Anne Smith of CRA International.

In responding to questions about my recent study of how climate legislation will impact the economy, I made two key points:

  • It’s important to look at a range of numbers and modeling results, using averages and medians, and not rely on one number.
  • Economic models are designed to compare different policies, not to predict the future. It’s the difference between model predictions that counts, not the absolute numbers.

This second point is very important, and often misunderstood. The predictions of a particular model for two different policies can tell you which is better, but cannot tell you the precise dollar impact of each policy. Beware the economic-model-as-crystal-ball fallacy!

This entry was posted in Economics. Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. johnmashey
    Posted May 12, 2008 at 1:23 am | Permalink

    Have you had a chance to consider my questions on the original report?

    I read IPCC & Stern, so I certainly accept your first comment, and the general idea of the second, i.e., scenario planning not forecasting.

    My concern, especially as I’ve been studying the MIT reports, is trying to understand the assumptions with the Reference case: which says:

    Table 5: Crude oil ($/bbl) 2002-2006 = $40

    Figure 5 (a) shows Reference case as reaching about 2.25X current (i.e., about $90) around 2050. The other scenarios yield 1.25X – 1.5X, i.e., $50-$60/bbl. Given that the current price is around $120 (well, somewhat less in 2005$), i.e., already off the chart by 2008, I’m having trouble understanding this.

    Looking at the Reference case (page 66), I divide
    a) Net crude oil imports in billion 2005$
    b) The Petroleum product prices, rising from 1.0 in 2005 to 2.25 in 2050.

    to get
    c) The relative percentage of 2005 imports:
    2005 100%
    2010 96%
    2015 93%
    2020 90%
    2025 85%
    2030 87%
    2035 98%
    2040 106%
    2045 94%
    2050 83%
    I don’t yet understand the gyrations, but even more IF the peak Oil folks are even close to right, we’re at/close to peak right now, and world oil production in 2050 will have fallen to around 50% of current, as in Oil poster, or The Oil Drum Update. The exact numbers and dates don’t matter of course, but still:

    a) US import declines through 2025 might track world production decline, depending on the curve, but then, somehow, it bounces back equal or above current, unless my spreadsheet is broken.

    b) Still, in an era of rapidly-increasing demand from China & India, and increasing use of subsidized oil in many exporting countries, it’s is puzzling to see how the US *increases* its consumption of world oil production going forward.

    c) Then, the primary energy use (EJ) of oil rises from 42.0 in 2005 to 68.8 in 2040 (pre-shale oil), i.e., 1.6X larger. US oil production peaked in 1970, so this substantial increase in primary energy from oil must come from a big boost in imports. (?)

    Some of the scenarios have lower oil usage, but they seem based on policy choices, not on petroleum geology, and where the money goes is different.

    Anyway, maybe I’m just missing something obvious, and I wouldn’t mind if Peak Oil+Gas were mentioned only to dismiss them as issues to be covered, but I don’t see any mention, which makes me nervous.

  2. Posted May 13, 2008 at 4:44 pm | Permalink

    Nat just posted a reply to your previous comment: