Forecasting economics: a classic case

Here‘s a classic case of extrapolation problems.  The Wall Street Journal polled a bunch of experts, asking them to predict American GDP and unemployment for the next two years.

The 54 economists see a 2.3% expansion for 2012, compared with an expected rate of 1.7% for 2011. Although an improvement, that isn’t fast enough to tame unemployment. On average, they see that rate still at 8.5% in December 2012, compared with the 8.6% reported in November this year.

Those predictions are based on a series of educated guesses about certain key factors.  Folks seem to have a consensus on these.


Most respondents — 88% — say Congress and President Barack Obama will come to an agreement on a payroll tax cut before the end of 2011. If policymakers fail to reach a deal and rates increase, it could crimp consumer spending and overall growth.

Item: “Just 15% of respondents say the government will agree on a significant deficit-reduction plan next year. If Washington manages to come to a deal, it could improve confidence and add to growth next year.”

88%, 85%: serious majorities.  The kind of thing a Delphi process would show, too.

But some of these factor assessments aren’t anywhere near consensus.  That makes those results more problematic.

Consider: “56% expect expanded unemployment benefits to be extended for 2012”.

Or “nearly half — 48% — predict some form of policy [quantitative?] easing by the Federal Reserve.”


some 60% of respondents assume that Europe will come up with an orderly resolution to its crisis during the next year. If policymakers fail to defuse the crisis and it spreads to the banking system and across the Atlantic, it could pose a significant downside risk for the U.S. In such a scenario, the relatively slow 2.3% forecast growth rate could look wildly optimistic.

That’s a serious, nearly even split among respondents.  Perhaps the thing to do would be to generate different scenarios based on each of these splits.  i.e., one where the Fed does policy easing, and one where it doesn’t.  With three of those splits, in my quick reading, we would see up to eight (?) scenarios.
unemployment benefits extended yes, Fed policy easing yes, orderly Euro resolution no

unemployment benefits extended yes, Fed policy easing no, orderly Euro resolution yes

unemployment benefits extended no, Fed policy easing yes, orderly Euro resolution yes

unemployment benefits extended yes, Fed policy easing no, orderly Euro resolution no


Leave a comment

1 Comment

  1. skjandrews

     /  December 12, 2011

    It is amazing to the the possible complexity that could emerge with just a few variables. On the other hand, the basic, neoclassical understanding of how a change in each of these variables WILL affect the others is completely suspect at this point.

    Point of fact: how could the fed do any more easing that it already is? At this point it seems like anything the fed did would be symbolic at best. In this sense, it is hard to parse the material economic effects of the change from the ideological or religious signal it provides to its adherents.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: