Worst News of the Day

The market at Intrade is predicting a probability of a U.S. Depression (10% or greater decline in GDP) in 2009 to be 56%.  I am shocked.  The probability is up significantly over the past two weeks.  Why?

Addendum: MR Readers are fast!  Within two minutes Marc points us to this post from Donald Luskin who says that it is a contract error.  Here is the key clause from the contract:

For expiry purposes a depression is defined
as a cumulative decline in GDP of more than 10.0% over four consecutive
quarters [that sounds ok, AT]. This is calculated [here is the error] by adding together the published
(annualized) GDP figures (as detailed below). If these annualised
figures add up to more than -10.0% over four consecutive quarters then
the contract will expire at 100.

As Luskin notes

The problem is that if you add four quarterly
change-figures that are already each annualized, you will get a far
larger cumulative result than the actual change over a four-quarter
period. Suppose there are four successive quarters each showing an
annualized 2.5% decline in GPD. Intrade will add those together and get
10%. But over the year, the annual decline in GDP will acually be 2.5%.

Thanks readers, I was about to open an Intrade contract to bet heavily against!

Hat tip to Tim Groseclose.

Comments

http://www.poorandstupid.com/2009_01_11_chronArchive.asp#857823201775567383

Re: Marc's response

So if that interpretation is correct, then people are actually betting on a 2.5% contraction? Or something like that?

Because the definition in the contract uses annualized sum over four consecutive quarters. Specifically, the sum of last four numbers from Table 1.1.1 published by BEA. So even a 3% annual GDP decline would cause the contract to expire at 100. This has happened 3 or 4 times since the depression.

This issue was cleared up about a week ago. Hence, the price increase.

There are separate contracts for annual GDP decline (such as 10%) for US and other economies.

*Specifically, the sum of last four numbers of the first row from Table 1.1.1 published by BEA.

Relying on Luskin for an explanation is like relying on a hooker for a rationalization for fooling around.

What hellblazer said

Alex and Don,
Thanks for the clarification. I was thinking of betting against too.
D.

perhaps for this contract we wouldn't want to assume risk-neutrality!

"Suppose there are four successive quarters each showing an annualized 2.5% decline in GPD. Intrade will add those together and get 10%. But over the year, the annual decline in GDP will acually be 2.5%."

I saw an error just like this in an economics book recently! I would have written to the publisher but it was published in 1957 and out of print, and also it was about an economy which doesn't exist anymore, so they might not have cared all that much.

Are the quarterly numbers the percentage change from the previous quarter, annualized, or are they the actual change from the same quarter a year ago?

In neither case can four such changes be added together to get a year's result, but in the latter case it should be an annual number.

-dk

It actually only calls for a 2.5% reduction over the year. There was a recent contract clarification

Share admin thans a lot for sharing a very successful and wonderful.

You are posting comments too quickly. Slow down.

Very useful in terms of sharing people. Thank you.

You are posting comments too quickly. Slow down.

Fine was a work and full-time. Thank you.

Honestly, looking at all the chickens without heads running around Washington and New York these days leads one to the most pessimistic forcasts...

Comments for this post are closed