Network News

X My Profile
View More Activity
Posted at 9:50 AM ET, 12/ 7/2007

The Fact Checker Fact Checks The Post

By Michael Dobbs


The leaders of Mexico, U.S., and Canada celebrate the NAFTA agreement.

"The impact of NAFTA seems to have been both larger and more positive in Mexico than in the United States. Mexico's gross domestic product, now more than $875 billion, has more than quadrupled since 1987."


--Washington Post editorial, December 3, 2007.


"If anyone still thought that the Washington Post editorial board could discuss trade in a rational manner, today's editorial on NAFTA proved them wrong. How on earth does the Post get that Mexico's GDP quadrupled when it actually only grew by 67.6 percent?"


--Blog by economist Dean Baker, The American Prospect, December 3, 2007.


A reader, Alan Abramowitz, asked me to fact check this one, and I am happy to oblige. Headlined "Democratic Candidates Resort to Far Fetched Denunciations of the North American Free Trade Agreement," the Post editorial was part of the "Ideas Primary" series looking at the issues raised in the presidential campaign, so it is fair game for this column.


The Baker post was picked up by Paul Krugman, a Princeton economist who writes a column for the New York Times. In his blog, Krugman blasted the Post editorial as "embarrassing," but made no effort to establish the facts behind the stats. The Post editorial board has also failed to explain, at least in print, the source of its statistics. So what is the truth? Did the Mexican economy quadruple in size over the last two decades, or did it grow by only 67 percent?


The Facts

First a disclaimer. Washington Post editorials are the responsibility of the editorial board of the newspaper, which has nothing to do with the news side. I have spoken with the principal author of the editorial, but I am not going to identify that person, as Post editorials are the expression of editorial board policy. This is what I found out:

The editorial writer got the information about the growth in the Mexican economy from a Mexican embassy slide show distributed by the Peterson Institute for International Economics, a Washington D.C. thinktank. The presentation is authored by Antonio Ortiz-Mena, a well-known Mexican economist. Slide Six shows an increase in Mexican Gross Domestic Product from $200 billion dollars in 1987 to $875 billion (estimated) in 2007.

The slide does not provide any further information about the source of the data, and does not say whether the dollars are current or constant, i.e. inflation adjusted, which is the more relevant method for measuring economic growth. I turned to the data tables of the International Monetary Fund, a widely accepted source of information about comparative economic growth rates. They show that the Mexican economy grew from $148 billion in 1987 to $886 billion in 2007, an increase of 337 percent. These are current dollars. If you adjusted these figures for inflation, you might get a result similar to the figures used by The Post.

UPDATE: Actually, as a couple of readers have pointed out, this seems a stretch. According to the Consumer Price Index, the cost of living has increased 79 percent between 1987 and 2007. $148 billion in 1987 is the equivalent of $265 billion in 2007. I will ask the Mexican embassy how they arrived at their 1987 $200 billion GDP figure, and report back.

There is, however, a 64 billion peso question: why use dollars to measure the growth rate of the Mexican economy? If exchange rates are volatile, growth rates can be all over the map. Imagine what would happen if you used dollars instead of Euros to measure the growth of the European economy. Measured in dollar terms, the European economy has been booming recently, simply because the dollar has fallen sharply vis a vis the Euro.

It turns out there are several other ways of looking at the same statistics.

Dean Baker, an economist with the Center for Economic and Policy Research, used the IMF peso figures, adjusted for inflation, to calculate the growth in the Mexican economy. I have been over the figures with him and he now concedes that he made a couple of mistakes. First, he used the American dollar sign when he meant pesos. Oops! Second, he had a mistaken figure for 1987. The real figures, as reported by the IMF, are 1,029 billion pesos in 1987 and 1,894 billion pesos in 2007, which works out at a growth rate of around 83 per cent. That is higher than the 67 per cent originally reported by Baker, but still much lower than the quadrupling claimed by the Post editorial.

To help me adjudicate this dispute, I turned to Paul Blustein, a former international economics reporter for the Post, now with the Brookings Institution, where he is writing a book about international trade. He said he was "sorry to go against my old alma mater," but he came down on the side of the critics. His explanation:

Constant dollars can be a good way of looking at a country's economy, but when there have been huge moves in that country's currency against the dollar, it is better to rely on the local currency. This would have raised a red flag with me. I don't think any economy in the world has quadrupled in twenty years. That would be an amazingly fast rate of growth. I doubt that even the Chinese economy has done that."

Overall, the Mexican economy has grown at more or less a similar rate to the U.S. economy, according to an IMF chart available here. It has done a little better some years, but it has also had sharp setbacks, as in 1995, after the peso crashed.

I also turned to the IMF, which suggested a third way of looking at the statistics. According to spokeswoman Angela Gaviria-Baptiste, IMF economists favor a measure called PPP, or purchasing power parity, which allows for international comparison by using a common basket of goods. The IMF adjusts for population growth by using per-capita PPP rates. According to the IMF data, the purchasing power of the average Mexican has risen by around 125 per cent between 1987 and 2007.

So take your pick. Depending on the statistics you use, Mexican economic growth over the last two decades has been either 337 percent, 125 percent, or 83 percent. By consulting the IMF website, you can find even more figures. International economists, please weigh in!

The Pinocchio Test

This is a case study of how statistics can be used to support virtually any argument. Without specifying the source, they can mislead more than they can clarify. At the very least, the Post editorial board should have been much clearer about the source of the statistics, and explain why dollars are the appropriate measure for the growth of a peso-based economy. The claim of a quadrupling in the size of the Mexican economy over two decades is misleading, and should have raised some eyebrows. Two Pinocchios for the Post.

The critics were also sloppy in their use of statistics, but at least they pointed out the source. One Pinocchio for them.

(About our rating scale.)

By Michael Dobbs  | December 7, 2007; 9:50 AM ET
Categories:  1 Pinocchio, 2 Pinocchios, Economy, MSM Watch, Other Foreign Policy  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   Del.icio.us   StumbleUpon   Technorati   Google Buzz   Previous: Clinton vs Obama: The 'Fact Check' War
Next: Most Revealing Fibs: Hillary Clinton

Comments

My math says an increase from $200b to $875b would only be a gain of 337.5%. True, $875b is over 400% OF $200b, but the increase would only be $675b.

Posted by: Marc | December 7, 2007 10:54 AM | Report abuse

Ooops! A miscalculation by the Fact Checker! I have corrected it.

Posted by: The Fact Checker | December 7, 2007 10:59 AM | Report abuse

Just a point about math. Your statement that an increase in GDP from $200 billion to $875 billion is "more than a 400 percent increase" is incorrect. It's actually an increase of 337.5 percent (675 divided by 200).

Posted by: Steve | December 7, 2007 10:59 AM | Report abuse

"In his blog, Krugman blasted the Post editorial as "embarrassing," but made no effort to establish the facts behind the stats."

I hope the Post gives you a remedial reading course for the holidays because in Paul Krugman's blog post he agreed with what Dean Baker wrote at The American Prospect. He further felt no need to "establish the facts behind the stats" because, as Krugman said, he was "more disturbed by the Post's characterization of Hillary Clinton's fairly careful remarks..." Krugman then went on to address what bothered him about what the Post had to say in that regard.

Two Pinocchio's for the Fact Checker.

Posted by: cab91 | December 7, 2007 12:13 PM | Report abuse

Why are we assuming that Mexico's increase in GDP is only due to NAFTA?


.

Posted by: jeffboste | December 7, 2007 12:34 PM | Report abuse

"...why use dollars to measure the growth rate of the Mexican economy?"

Because by using a common standard (the dollar) comparisons are understood much more easily. It's Apples-to-Apples. Not only with the U.S., but interchangably among many other economies.

Using the dollar equivalent of the Euro, or any other currency, doesn't change the underlying mathmatical results. A percent is a percent, whether or not the measure is the Peso, the Dollar or the Eueo. Having percentages as the sole marker would work nicely; except people always want the raw data behind the percentage.

Posted by: Nor'Easter | December 7, 2007 3:43 PM | Report abuse

You claim that the IMF figures "show that the Mexican economy grew from $148 billion in 1987 to $886 billion in 2007. These are current dollars. If you adjusted these figures for inflation, you might get a result similar to the figures used by the Post."

How would adjusting these figures for inflation yield the Post's figures or anything close to a quadrupling? Currency choice aside, isn't the major error that the Post did NOT adjust for inflation?

Posted by: Ben Zipperer | December 7, 2007 7:45 PM | Report abuse

Zipperer wrote:
-----
How would adjusting these figures for inflation yield the Post's figures or anything close to a quadrupling? Currency choice aside, isn't the major error that the Post did NOT adjust for inflation?
-----

Yep. I'm guessing the fact checker needs to go back to school for a while...

Posted by: Mark | December 7, 2007 9:46 PM | Report abuse

Can you please respond to Dean's comment on his blog about how it is not appropriate to "take your pick" on how to measure growth? I don't see the Post reporting economic growth three different ways ever. Why would it have been appropriate to do in this one case where it is part of the basis for making an specific editorial argument? The false claim of quadruple growth is more than 2 pinocchios - it powerfully undermines any reasonable belief the public could have had in the editorial writers' ability to analyze facts. When viewed along with the many other frequent errors made by the Post as pointed out by Dean Baker, it should result in somebody getting fired.

Posted by: mlarsen23 | December 8, 2007 3:03 PM | Report abuse

You must be kidding! The only--repeat only--way to measure GDP growth over time is adjusting for inflation. This response by a so-called "fact checker" is totally ridiculous. There is simply no way that the 337 percent figure, unadjusted for inflation, should be used. By that standard, when German inflation after World War I had Germans pushing wheelbarrows full of currency around, Germany was having spectacular economic growth. Is there no one at the Post who has taken a beginning economics course? Clearly, the so-called "fact checker" has not. Pathetic.

Posted by: macheath | December 8, 2007 5:43 PM | Report abuse

2008 Presidential Election Weekly Poll

http://www.votenic.com

The Only Poll That Matters.
Results Posted Every Tuesday Evening.

Posted by: votenic | December 8, 2007 6:21 PM | Report abuse

"The editorial writer got the information about the growth in the Mexican economy from a Mexican embassy slide show...."

Perhaps I'm missing something, but the source that the Editorial Writer used appears to me to indicate lazy and shoddy research. A Mexican Embassy slide show ??? I wouldn't let freshmen and sophomore history students use that source. That's embarrassing journalism.

Posted by: Florida Dan | December 8, 2007 10:55 PM | Report abuse

"The editorial writer got the information about the growth in the Mexican economy from a Mexican embassy slide show...."

Perhaps I'm missing something, but the source that the Editorial Writer used appears to me to indicate lazy and shoddy research. A Mexican Embassy slide show ??? I wouldn't let freshmen and sophomore history students use that source. That's embarrassing journalism.

Posted by: Florida Dan | December 8, 2007 10:57 PM | Report abuse

I appreciate the Fact Checker's willingness to look at the accuracy of the facts cited in a Post editorial, but there's another issue here that the F.C. did not address. The Post editorial cited data on the growth of Mexican GDP since 1987. But NAFTA didn't take effect until 1994. In fact, even measuring growth in real dollars, Mexico's GDP grew at a slightly higher rate before NAFTA than after NAFTA. According to World Bank data , Mexican GDP grew at an average annual rate of 3.16 percent in real dollars for the seven years from 1987 through 1993 compared with an average of 3.06 percent for the 14 years from 1994 through 2007 (2007 based on preliminary estimates, I suppose). So there is no evidence using the Post's own measure of growth that NAFTA has a positive impact on the growth of Mexican GDP.

Posted by: Alan Abramowitz | December 9, 2007 1:15 PM | Report abuse

there were a bunch of other comments on here. Were they deleted?

Posted by: macheath | December 10, 2007 1:19 PM | Report abuse

oops--nevermind

Posted by: macheath | December 10, 2007 1:20 PM | Report abuse

i'm sorry, but, this is nuts:

"At the very least, the Post editorial board should have been much clearer about the source of the statistics, and explain why dollars are the appropriate measure for the growth of a peso-based economy. The claim of a quadrupling in the size of the Mexican economy over two decades is misleading, and should have raised some eyebrows."

Wrong. The problem is that the Post used numbers that didn't adjust for inflation. Failing to do isn't "misleading", it's wrong. When the Post reports GDP growth in the US economy every quarter they report the inflation-adjusted number, not the nominal number. Why is this different? Or hard to understand?

And, this is really embarrassing - the IMF numbers you quote (125% growth in GDP per capita) are...not adjusted for inflation, either!

Which makes this sentence:

"According to the IMF data, the purchasing power of the average Mexican has risen by around 125 per cent between 1987 and 2007."

also flat wrong. Nominal GDP measured in PPP current dollars increased 125%, not "purchasing power". For purchasing power you'd have to, again, adjust for inflation.

Please, please try to get this right.

Posted by: josh bivens | December 10, 2007 4:43 PM | Report abuse

Josh, you are wrong on this one, not the fact-checker. PPP figures cannot be adjusted any further, as they represent a basket of goods. A crude example of a PPP indeed is the Big Mac index, popularized by the Economist. Say a Mexican can buy one big mac in 1987 but 2.25 big mags in 2007. That is a 125 per cent increase in the Big Mac index, and in the Mexican's purchasing power. it makes no sense to adjust that for inflation. that is the whole point of PPP indexes.

Posted by: economist | December 10, 2007 5:31 PM | Report abuse

What is unfortunate here, and is a frequent crutch of the media, is the reliance on big numbers to make headlines. It's exciting to say that Mexico's economy grew at 337% or 125% or 83% or whatever. But, we're talking about a 20 year time span. To quote such large numbers and stretch them out over such a large timeframe gives an extremely misguided impression. No one here is pointing out that that the annualized rate of 83% growth over 20 years is only 4.15% per year, which would generally be considered only modest growth. In comparison, China and India have, from various reports, seen 8-12% annual growth in their GDP in the last few years, which is considered very robust. As Mr. Blustein alludes, the idea that any country would have grown by 337% over 20 years, an annual rate of 16.85%, is absurd and the Post editorial board should have at least realized that. But, more importantly, they seem to be throwing fuel on the anti-NAFTA fire by entitling the editorial "Trade Distortions" and combining the growth and the time factors to create misleading numbers and a misleading headline.

Posted by: Richard Baker | December 11, 2007 12:35 PM | Report abuse

I have an email saying that Michael Dobbs was once a "good reporter". What has happened to him to lead him to produce this piece of misleading tripe?

Posted by: Brad DeLong | December 11, 2007 12:41 PM | Report abuse

Students are failed for not adjusting for inflation.

Why aren't Post editors (and "fact-checkers") held to the same standard?

Posted by: Econ Teacher | December 11, 2007 6:13 PM | Report abuse

This is really horrendous. It is plainly wrong to use current (non-inflation adjusted) rather than constant (inflation adjusted) currency to calculate growth rates. No economist--indeed no one who understood Econ 1--would say otherwise.

Also, either the IMF spokesperson is incompetent or (more likely) this post is entirely misrepresenting what she said. Current dollar PPP is used to compare levels of GDP per capita across countries but would never be used to calculate GDP growth rates.

Posted by: Gabriel | December 11, 2007 6:23 PM | Report abuse

Econ Teacher makes a good point. If this had been provided as the answer to a question about growth rates in an introductory economics course, it would have received an F, no question.

Posted by: Ph.D. Economist | December 11, 2007 6:26 PM | Report abuse

The Pinocchio Test [the revised and corrected version]

I have revised and extended FC's evaluation [putting in what must have been in an honest FC's thoughts.]

This is a case study of how statistics can be used to support virtually any argument. [In the hands of a liar who takes care to present the statistical 'evidence' ] Without specifying the source, they can mislead more than they can clarify. [This might be why there is a best seller "How to lie with statistics"] At the very least, [had] the Post editorial board [been at all interested in presenting a factual arguement, they] should have been much clearer about the source of the statistics, and explain [that the only reason] why dollars are the appropriate measure for the growth of a peso-based economy [is to get the result they wanted to claim 400 percent growth]. The claim of a quadrupling in the size of the Mexican economy over two decades is [so terribly] misleading [that no one who values facts over bloviating would consider the Post editorial board to have anything to say since Katherine Graham died (bless her soul)], and [the editorial was so blatantly wrong that people who bring reason to their reading of the Post wonder what can be done. To say that it ] should have raised some eyebrows [is really letting the boss' Editorial Board off easy. While this editorial properly deserve four Pinnochio, my desire to remain employed by Donald Graham requires that I award only]. Two Pinocchios for the Post.

Posted by: Anonymous | December 11, 2007 8:32 PM | Report abuse

economist

w/ all respect - that's incorrect. PPP figures can indeed be adjusted for the passage of time.

they generally shouldn't be, for very esoteric reasons, and, the best measure of a country's growth through time is inflation-adjusted measures of their domestic currency.

But, if one insists on using PPP dollars, they do indeed need to be adjusted for inflation if you're comparing two different points in time.

PPP indices are essentially a way to adjust for price differences across countries, not across time. An example, using your Big Mac numbers:

in 1987 a Big Mac costs $1 in the US, and, 10 pesos in Mexico, so, the PPP rate between the countries is 10 (10 pesos to 1 dollar).

in 2004, a Big Mac costs $2 in the US, and, 20 pesos in Mexico, so, the PPP rate between the countries is still 10 (20 pesos to 2 dollars).

Now, say that average incomes in Mexico are 5,000 pesos in *both* 1987 and 2004. Is the average Mexican worse off in terms of purchasing power (over Big Macs) in 2004 compared to 1987? Of course they are. In 1987 they could buy 500 Big Macs (5,000 peso salary divided 10 peso price of Big Macs) and in 2004 they can only buy 250 Big Macs (5,000 peso salary divided by 20 peso price of Big Macs).

All the PPP measure has told us is that Big Macs haven't gotten any more or less expensive in Mexico between 1987 and 2004 *relative to their cost in the US*.

But, they have absolutely gotten more expensive in peso terms, and, hence one needs to inflation-adjust any measure that uses Big Macs as a numeraire.

Posted by: josh bivens | December 12, 2007 2:38 PM | Report abuse

For those very few non-economists reading all the comments, the key issue is that to be meaningful, economic growth rates must be done using inflation-adjusted, aka "constant" dollars (or other currency). This is perhaps the single most elementary point for understanding economic data, and yet the Post's "Fact-Checker" proclaims himself agnostic as to whether this is true or not.

The choice of national currency vs. PPP-dollars for the calculation is a more subtle point and of much less consequence, although as Dean and others have noted use of PPP dollars for growth rates is not standard practice.

Posted by: Ph.D. Economist | December 13, 2007 2:36 PM | Report abuse

Dear PhD economist,
I don't know why you think I am "agnostic" on the question of whether growth rates should be adjusted for inflation, when I say in the post that constant, inflation-adjusted dollars are "more relevant" to this argument. ( I thought it was so obvious that growth rates should be adjusted for inflation that i didn't belabor the point.) I also described the editorial's claim of a quadrupling in the Mexican GDP as misleading, saying it should have raised eyebrows. That does not sound like agnosticism to me. When I said that you can use statistics to support any argument you like, that was not intended as a justification for the quadrupling claim.
A more valid criticism may be that I misunderstood how PPP rates are calculated, as Josh Bivens has explained. The reason I cited the PPP growth rate is that the IMF suggested that this was the most appropriate rate for international comparisons. I will leave it to you economists to debate whether this is in fact the case. The IMF does not have a constant PPP series: they only provide the data in current international dollars. An IMF economist told me that the PPP series could not be adjusted further. (There is no doubt a technical reason for that which is beyond me.)

Posted by: The Fact Checker | December 13, 2007 4:42 PM | Report abuse

Dear Fact Checker,

Many readers believe you do not understand that the only way to convey meaningful economic growth rates for a country over time is to adjust for inflation.

Readers believe this because of two points:

(1) When the Post failed to adjust GDP in 1987/2007 for inflation in its article, you did not state that the Post's growth rate of well over 300% is completely incorrect. Instead, you called the rate "misleading" and went on to imply that nominal growth rates are just as meaningful as real growth rates: "take your pick. Depending on the statistics you use, Mexican economic growth over the last two decades has been either 337 percent, 125 percent, or 83 percent.... the Post editorial board should have been much clearer about the source of the statistics, and explain why dollars are the appropriate measure for the growth of a peso-based economy."

(2) You wrote

"[IMF tables] show that the Mexican economy grew from $148 billion in 1987 to $886 billion in 2007, an increase of 337 percent. These are current dollars. If you adjusted these figures for inflation, you might get a result similar to the figures used by The Post."

This claim is false. Adjusting these figures for inflation does not yield the quadrupling by the Post. Even the IMF tables you cite show that in constant pesos Mexican GDP increased by 84%.


The major error remains in print and online: the Post tremendously exaggerated the growth rate of Mexico during the NAFTA years by not adjusting for inflation.

Last year the Post made the virtually the same kind of error (see page A1, April 17, 2006), claiming that Mexican GDP increased seven-fold over 1993-2005. The Post corrected this error 43 days later.

This time, must we wait another month for a correction?

Posted by: Ben Zipperer | December 13, 2007 7:40 PM | Report abuse

fun search test!

Posted by: Anonymous | January 3, 2008 1:32 PM | Report abuse

the annualized rate of 83% growth over 20 years is only 4.15% per year,
Posted by: Richard Baker | December 11, 2007 12:35 PM

wrong -- it's 3.07% a year (and 83% is not "annualized")

1.83 to the 1/20 power

Posted by: idiot | January 15, 2008 6:42 AM | Report abuse

The comments to this entry are closed.

 
 
RSS Feed
Subscribe to The Post

© 2011 The Washington Post Company