If you’re familiar with “Trading on Time,” I have a quick question. At the bottom of page eleven, in defining the remoteness variable, the subscript for Distance is lj. I suspect this is just a typo that should have been kj. Is this what others who have read the paper concluded? D is used for both distance and a vector of dummy variables at different points in the paper, so I may just be confused.

Dan Drezner points to a NYT article that contains an important point regarding our growing trade deficit with China:

But often these days, “made in China” is mostly made elsewhere — by multinational companies in Japan, South Korea, Taiwan and the United States that are using China as the final assembly station in their vast global production networks.

Analysts say this evolving global supply chain, which usually tags goods at their final assembly stop, is increasingly distorting global trade figures and has the effect of turning China into a bigger trade threat than it may actually be. That kind of distortion is likely to appear again on Feb. 10, when the Commerce Department announces the American trade deficit with China. By many estimates, it swelled to a record $200 billion last year. [NYT]

I apologize for nit-picking, but I think advocates of liberalization hurt their cause when they overstate their case. Johan Norberg writes:

I still meet anti-capitalists who claim that liberalisation and market reform destroy economies and increase poverty. Why don´t they take a look at this graph from the new Index of Economic Freedom 2006. Per capita growth in the countries that liberalised the most since 1995 has been almost three times higher than in those where economic freedom declined.

Let’s stop pretending that trade liberalization skeptics are ridiculous for not hailing the truth of this and similar bar graphs. Anyone that took AP Statistics in high school can produce a sufficient number of objections to render this evidence inconclusive: correlation isn’t causation, the causal relation may be bidirectional or run the other way, lurking variables may be omitted, etc.

In Norberg’s defense, in spite of titling his post “freer and richer,” he actual formulates a weaker claim — liberalization does not destroy economies and increase poverty. It appears true that improvements in a country’s economic freedom score do not negatively affect growth so much as to overwhelm other engines of growth. But “liberalization isn’t harmful” is a far cry from “liberalization is helpful.” Moreover, the bar graph describes per capita GDP growth, not poverty reduction. There is plenty of evidence that economic growth reduces poverty, but I think it’s a bit sloppy to conflate the two.

Peter Gallagher’s latest post is the third or fourth article that I have read this week to allude to Gareth Evans‘ quip that the Asia Pacific Economic Cooperation is “four adjective in search of a noun.” Regardless of opinions about APEC policies, analysts ought to stop repeating this grammatical error.

Asia is a noun. Cooperation is also a noun. Asia Pacific Economic Cooperation might be a clumsy or meaningless title (perhaps it should have been Asia Pacific Economic Cooperative?), but two of the four words are nouns.

Moreover, APEC hyphenates its name, so it reads “Asia-Pacific Economic Cooperation.” In that case there are three words, and one of them is a noun.

In a generally well-written piece criticizing the Bush administration’s fondness for bilateral trade agreements, Bruce Bartlett writes:

A 2003 study by the Congressional Budget Office found the economic potential of bilateral agreements very limited. It noted NAFTA, one of the largest such agreements, had virtually no effect on the U.S. trade balance with Mexico even after eight years. However, the study also noted there might be important noneconomic reasons to support free trade agreements. [WaTi]

The balance of trade is not a measure of economic well-being (though it can signal problems in the economy). It’s an accounting figure that must balance vis-a-vis the capital account (or in the case of capital immobility, balance to zero itself). A more appropriate measure of the economic potential of bilateral agreements is the trade volume, though it too is not a measure of welfare.

To illustrate via the most extreme example possible, imagine a world of two countries with immobile capital. Under both autarky and free trade, each nation’s trade balance would be zero. Clearly there would be welfare differences between these two policy-worlds. Volume, though not a welfare measure, is a more relevant statistic than balance.

[This post originally appeared on another blog written by Jonathan Dingel. It has been imported into Trade Diversion. I apologize for the hyperbolic rhetoric of my former years.]

Tyler Cowen at Marginal Revolution recently unknowingly resurrected an old fallacy. He writes that, of the world’s one hundred largest economic entities, “[f]ifty-one are corporations, and General Motors comes in at number twenty-three, just ahead of Denmark (the data are from 2000, Wal-Mart should be higher than listed, among other changes).” The caveat Cowen offers (“To be sure, these comparisons are problematic. Yearly sales are not strictly comparable to gross domestic product”) is woefully insufficient.

Jagdish Bhagwati tackles this fallacy in his new book, In Defense of Globalization.

This dramatic statistic is misleading, however, as the two sets of data are not comparable… So when we compares sales volumes, which are gross values, with GDP, which is value added, we are comparing oranges with apples. The comparison, while conceptually flawed, also exaggerates the role of corporations because sales figures across the entire economy will add up to numbers that will vastly exceed the GDPs of the countries where these sales occur. [p.166]

This idea of corporations ruling the world was manufactured by left-wing critics of globalization in order to instill fear. Corporations, despite all their incentives to please populaces and customers, are not democratic, so people prefer to see democratic governments remain powerful enough to control corporations when necessary. “Wal-Mart and GM are stronger than most governments!” is a great catch-phrase to scare people undecided about the desirability of globalization.

Martin Wolf took this Institute for Policy Studies “paranoid delusion” to task over two years ago; it’s too bad that Professor Cowen blogged without catching this criticism of the data. Here’s how Wolf reads the data in his February 6, 2002 Financial Times column:

In fact, only two of the top 50 economies, measured by value added, and 37 of the top 100 were corporations. For the critics, GM is bigger than Denmark and Wal-Mart is bigger than Poland. Properly measured, Denmark’s economy is more than three times bigger than GM. Even impoverished Bangladesh has a bigger economy than that of GM.

But the flaw in such claims is not just factual but also conceptual, since countries and companies are radically different. A country has coercive control over its people and its territory. Even the weakest state can force millions of people to do things most of them would far rather not do: pay taxes, for example, or do military service. Companies are quite another matter. They are civilian organisations that must win the resources they need in free markets. They rely not on coercion but on competitiveness. [GlobalPolicy.org]

In short, leftists trying to critique corporate power have bastardized the data until it produced an interesting statistic. Hopefully the Marginal Revolution will revise its take on the matter to include a stronger disclaimer than its mere “these comparisons are problematic.” Something along the lines of “these comparisons are ridiculous and misleading” would be more appropriate.