Sunday, February 27, 2011

The New York Times Editorial Board Writes

In a recent report, economists at Goldman Sachs estimated that the House cuts would reduce economic growth by 1.5 percentage points to 2 percentage points in the second and third quarters of 2011. That would devastate employment. As a rule of thumb, each percentage point drop in growth means a loss of 1.2 million jobs.

They confuse "growth of" and "the annualized rate of growth of". Since the time interval is two quarters, this is a factor of two error. Also their Okun's law coefficient (the rule of thumb) is roughly 12/7 times the standard coefficient estimated with historical data. Thus an overall errof by a factor of 24/7. Not Sour Sixteen, but not a very good job for colleagues of Paul Krugman.

It is a good thing that the NY Times editorial board is trying to predict the effects the Republicans proposed cuts would have. But they didn't succeed.

The Goldman Sachs team forecast a decline in the annual growth rate of 1.5% over the next two quarters (not 1.5 % to 2% by the way). That means they forecast that third quarter 2011 GDP would be 0.75% lower if the Republicans get their way.

The employment level is related to the GDP level (via the rule of thumb called Okun's law) so related to a decline in the annual growth rate times the duration of that decline (in this case they only forecast out two quarters).

It isn't and can't be true that "each percentage point drop in [the annualised] growth [rate] means a loss of 1.2 million jobs" Also it isn't true that a 1% reduction of GDP means a loss of 1.2 million jobs. The rule of thumb is employment declines roughly 0.5 % of the labor force for every 1% decline in GDP. That means a 1% decline in GDP (not the growth rate the level) corresponds to a loss of around 700 thousand jobs not 1.2 million.

The forecast decline in employment if the Republicans get their way is around 350,000 jobs not 2.4 million as suggested by the New York Times.

The correct number shows that the Republicans are irresponsible and insane and must be stopped, but there is no reason to print arithmetic mistakes.

The editorial board should get on the phone to their colleague Paul Krugman who needs to give them a good talking to.


Brendan said...

Hope you sent this as a letter to the editor.

Robert said...

Well you can hope. I don't know how to send letters to the editor. Following your advice I will try.

Brendan said...

Good to hear.

If you haven't already figured it out, here or here are places to start.

Robert said...

I can't figure out how to e-mail them letters.

Brendan said...

Links in an earlier comment, probably still in your moderation queue. Or possibly your spam queue, due to the links.

(Go to your Blogger dashboard and click the Comments link.)

carlo said...

... I confess I suspected
you were not quite good in contacting editors .... :)


Robert said...

Dear Dr Dr Ciccarelli PhD

I can hardly believe it myself, but I have never written a letter to an editor of a newspaper. Back in the day you needed stamps and envelopes and stuff, which are just too complicated for me. Now I just comment on blogs.

carlo said...

please keep doing it. I learn more from your blog than from most macro textbooks.