Matthew Yglesias: Gee...

« Trading With Australia | Main | Not To Worry »

Gee...

...it strikes me as a bit misleading to construct a chart plotting change in GDP versus time and then discuss it as thought it were a chart of GDP versus time. The way Schultz has done it, the second Clinton administration looks like a period of plateau, albeit at a decent level, whereas mapping the second set of data points would reveal that it was, in fact, an era of rapid improvement in living standards. One could go on...

August 4, 2004 | Permalink

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8345160fd69e200d834567b4669e2

Listed below are links to weblogs that reference Gee...:

» Why Oh Why Are We Ruled by These Liars? (George Shultz Needs a Better Staff Edition) from Brad DeLong's Semi-Daily Journal (2004)
Matthew Yglesias reads the op-ed published under the name "George Schultz" in the New York Times this morning and is an unhappy camper: matthew: Gee...: ...it strikes me as a bit misleading to construct a chart plotting change in GDP versus time and th... [Read More]

Tracked on Aug 4, 2004 12:30:20 PM

Tracked on Aug 4, 2004 12:48:58 PM

» George Shultz: Propaganda Mouthpiece from Kautilyan
Delong and Yglesias beat me to the punch in pointing out the ways George Shultz's chart of Clinton vs. Bush economic performance is misleading. One point they didn't mention: Shultz saysThe employment picture has been a little puzzling since the [Read More]

Tracked on Aug 4, 2004 1:48:59 PM

» Right Wing Media Lies: The Times Pisses off a Mathematician from BOPnews
On Wednesday, the New York Times had the unfortunate lack of judgement to print this op-ed by former Treasury Secretary George Shultz. The piece uses a 2+2=10 type argument to claim that Clinton inherited prosperity and bequeathed recession, and that... [Read More]

Tracked on Aug 7, 2004 11:29:55 PM

Tracked on Aug 8, 2004 2:42:06 AM

Comments

Standard statics dirty trick. Sometimes it is clearly malevolent, sometimes it is just plain lack of understanding. What is that old adage, "Never attribute to malice what can be adequately explained by stupidity.".

Posted by: mdmhvonpa | Aug 4, 2004 9:35:44 AM

mdmhvonpa,
I think this can fairly be attributed to malice. In a world that cared about intellectual honesty know one would pay attention to Shultz again for a period 20 years.
The New York Times should run a correction.

Posted by: theCoach | Aug 4, 2004 9:45:12 AM

not only that, but it's in comparison to the same quarter of the year before.

now, I'm no paul krugman, but I fail to see how this is in any way meaningful.

Posted by: praktike | Aug 4, 2004 9:56:24 AM

now, I'm no paul krugman, but I fail to see how this is in any way meaningful.

Because that is what the people who matter are concerned with. They are not concerned with gdp growth in the abstract, but, rather, they are concerned with comparing this quarter's gdp growth with ladt quarters.

That is why the NY Times last week can have headlines that state:

"Economic Growth Worse Than Expected"

when what really happened is the GDP fell from 4.1% growth in one quarter to a paltry (smirk) 3.0-3.1% in the next quarter.

There are lots of misleading graphs out there, but this is not one of them.

Of course, in the fever swamps that are the comments section at any left-leaning site, it might be more fun to chalk it all up to malice and the evil George Schultz.

Posted by: John Cole | Aug 4, 2004 10:06:02 AM

I was going to agree with John Cole that I did not find the charts totally misleading and actually useful as one measurement among many, but my pathological Bush-hatred and psychotic leftism rendered me mute.

Posted by: bob mcmanus | Aug 4, 2004 10:13:30 AM

John Cole: as usual, you're wrong.

The chart shows the percent change in GDP from the same quarter of the prior year.

Posted by: praktike | Aug 4, 2004 10:19:54 AM

I don't think plots showing the rate of change in GDP are misleading-- provided people know they are plots of the rate of change in GDP!

Similarly, I would find no problems with showing plots of interest rates, nor would I have problems with plots showing the rate of change in the money supply. All are rates of change in some sense.

The real problem with statistics is that the writer can often assume that 95% of his readers know nearly nothing about statistics.

I do wish Schultz's plot had a big, dark heavy line indicating ZERO. It's easy for the reader to ignore the significance of that level on his plot (and many plots-- but in his case, it's particularly important.)

Possibly, the line should be red.

Posted by: lucia | Aug 4, 2004 10:20:09 AM

Give me record surpluses and allow me to run record deficits, and I'll give you even better GDP growth.

Just stick it to working class children down the road (unearned wealth is OK!).

How about promising each tax break (for campaign donors) will create millions of jobs?

Posted by: MattB | Aug 4, 2004 10:22:09 AM

Phew, I was beginning to wonder if anyone had noticed this excrement.

Seems to me that the lower growth rates toward the end of Clinton KINDA-SORTA support the idea that GWB got in at the wrong time, but hardly support GS' specific argument, much less the impression he attempts to give.

I wonder what's more harmful - American ignorance of our military history (apropos Matt's other post) or our tendency to directly link a president to the economic performance during his term? (Not that they can't both be harmful, obviously)

Posted by: T: Central | Aug 4, 2004 10:26:43 AM

Anyone want to call this chart what it is? Anyone want to discuss the significance of the zero levels (as lucia alludes to). And thus this chart is useful, but only when compared as well to the rate of change of change of the gdp as well as the level of gdp itself.

And whether Schultz, former Engineer from Bechtel himself knows all of this.

Posted by: derivative | Aug 4, 2004 10:30:42 AM

Got my voice back. No minion of the Dark Lord can break my spirit.

a) I have always said that Clinton did not necessarily improve the economy. Is any President responsible for the economy? Only Republicans when it improves, I guess. And Democrats when it declines. But Clinton did no harm (altho Greenspan should have put on the brakes a little) and Clinton used what he had efficiently. Neither of which you can say for the current dude.

b) On the topic of lies, damned lies, etc my Dallas Morning News this morning has business section headline:"Is a Job Surge in the Works?". Substance is that the July numbers may look very good. Course it really is because of a seasonal anomaly, and that this is a year in which July has five official weeks, but heck, when it comes to economic questions, the important factor is whether you love America and Bush or are a terrorist, I guess.

Posted by: bob mcmanus | Aug 4, 2004 10:32:02 AM

You forgot to mention that the chart shows all of 2001 as a recession year, following the "that recession started under Clinton" party line, even though the NBER says the recession only lasted from March to November, and the chart cites NBER as one of its sources.

Posted by: Seth Gordon | Aug 4, 2004 10:40:50 AM

We really need mandatory calculus classes in high school. I'm guessing that many Americans would call those graphs totally interchangable.

Posted by: Kriston | Aug 4, 2004 10:43:50 AM

People -- listen to what I am saying.

The chart compares, say, Q1 of 2001 with Q1 of 2000. So if GDP growth was 4% in Q1 of 2001 and 3% in Q1 of 2000, it shows up as a 1% in the chart.

Posted by: praktike | Aug 4, 2004 10:48:10 AM

And while we're on the subject of John Cole, first quarter GDP growth, John, was 3.9%; second quarter growth was 3.0% and was "weaker than expected" (even i, a pessimist on the economy, expected 3.2%); and we've now had three consecutive quarters in which GDP growth has fallen, which suggests that we've turned a downhill corner (although i won't be shocked if Q3 GDP growth is better than Q2, partly because Alan Greenspan, who has earlier access to information than you or i, thinks it will be. On the other hand, if it isn't, then once again, "GDP growth worse than expected" will be an accurate title).

And those fevered left-wing swamps on Wall Street have noticed....

Posted by: howard | Aug 4, 2004 10:54:43 AM

I wanted to let you all know about a new petition website we just started. We're trying to build it up as much as possible to let the press know how many working people like us are against George W. Bush. I hope you stop by www.peopleagainstw.org and give us some feedback. thanks!

Posted by: sam | Aug 4, 2004 10:54:50 AM

misleading to construct a chart plotting change in GDP versus time and then discuss it as thought it were a chart of GDP versus time.

It would be helpful if Matthew would point out exactly where he alleges that the op-ed discusses "GDP versus time", as opposed to "change in GDP versus time".

Indeed, the article discussions the recession, and recessions are determined in reference to change in GDP -- i.e., negative growth in GDP.

So, unless Matthew comes up with at least one piece of evidence to support his claim, I'm chalking this post up to more left-wing ignorance.

Posted by: Al | Aug 4, 2004 10:56:49 AM

I think there is another subtle deceptive trick in the chart which is to only show the last three years of the first Bush term, so that it looks as if he too inherited an economic downturn.

(I suppose one could argue that Schultz wouldn't want to imply that because if Bush I did inherit a recession he would have inherited it from Reagan, but I don't think most people reading the chart are going to think back that far.)

Posted by: cs | Aug 4, 2004 11:01:28 AM

What I found the most deceptive was not the use of the change in GDP -- which is, after all, the way it is usually reported in the press, but instead the plotting of "change in employment" which is not at *all* what is usually plotted.

Yes, the chart correctly shows that the rate of *growth* of employment dropped towards the end of the Clinton administration, but in fact employment did not start *contracting* until the GW Bush administration. And indeed, there has been a "recovery", as measured by GDP growth, during the latter part of this administration, but only quite recently has the economy started *adding* jobs as compared to *losing* them. I think that a plot of total *employment* -- especially as a percentage of population -- is much more instructive here.

Even the plots of GDP growth show an impressively long period of sustained growth, that only started to fall by the end of the Clinton administration. I don't believe that there has been such a previous long record of growth, though the 1983-1990 period comes close.

Posted by: Alex R | Aug 4, 2004 11:05:13 AM

The chart compares, say, Q1 of 2001 with Q1 of 2000. So if GDP growth was 4% in Q1 of 2001 and 3% in Q1 of 2000, it shows up as a 1% in the chart.

Uh, no!

Or are you REALLY saying that GDP growth was (say) 0% in mid-1996, then 4% in mid-1997 (because the chart gives a figure of approximately 4% at mid-1997, so praktike is saying that mid-1997 must have a 4% higher growth rate than mid-1996), then 8% in mid-1998 (because, under praktike's "logic", the mid-1998 growth rate must be 4% higher in mid-1998 than in mid-1997 if the chart shows 4% at mid-1998), then 12% in mid-1999, then 16% in mid-2000!!!

Think before posting.

Posted by: Al | Aug 4, 2004 11:09:56 AM

Well, Al, Matthew might have been referencing the fact that the point of Schultz's article is that despite what you think, the Clinton years were bad for you economically and Bush I and Bush II were good. He might have drawn that conclusion from the misleading statement that Clinton "inherited prosperity" (obviously in the broad sense this is true; there hasn't been a president in a very long time who hasn't "inherited prosperity" since we are a very prosperous country) and the equally misleading statement that Clinton "bequeathed a recession" (obviously in the broad sense this is true, since recessions are part of the market ecology of capitalism and therefore inevitable, but still).

In the real world, we had the longest peacetime expansion in history under Clinton, created 20M jobs, and saw real wages improve at all levels of the economy, but you'd never know that from schultz's setup.

Otherwise, read Praktile's comment.

Posted by: howard | Aug 4, 2004 11:18:00 AM

from the legend: "Change in GDP: percentage change from same quarter prior year."

Maybe I'm describing it wrong, but it does seem like a very odd way to measure trends.

Posted by: praktike | Aug 4, 2004 11:18:57 AM

ah, so what the chart is showing is that the level of GDP was X% higher in q1 of 2001 than it was in q1 of 2000.

ignore my previous explanation.

Posted by: praktike | Aug 4, 2004 11:21:22 AM

No, Howard.

What Shultz says is that economies have momentum. And the the momentum was upwards when Clinton took office, and downwards when Clinton left office. Shultz doesn't say much about that long period of time between when Clinton took office and when he left it.

Posted by: Al | Aug 4, 2004 11:32:59 AM

One fallacy of Schultz's construction lies in his mention of "momentum". He implies
that because of momentum, the negative slope
at the end of Clinton's term was bound to continue into Bush's term.

Of course this would be true if the chart showed GDP, but since it shows the first
derivative of GDP, any "momentum" would be reflected as a tendency towards a flat line,
not a continuation of a downward (or upward) line.

If Schultz feels that the economy has some inherent trait that tends towards a constant
acceleration of GDP (positive or negative), then he is welcome to argue the point. I
don't think that many economists would agree.

At any rate, I would say "yes", equating momentum with constant acceleration is
misleading.

- Punkerdubh

Posted by: punkerdubh | Aug 4, 2004 11:52:29 AM