IT columnist Bob Lewis often makes insightful comments about technology in his "Keep the Joint Running" newsletter. His most recent installment is no exception.
Commenting on a recent editorial in the Washington Post, Lewis pointed out that the editors had made an error in logic that stems from a misunderstanding of mathematical principles explained in an academic paper on investments in the U.S. vs. those overseas:
According to the editorial, "A 2008 study by economists Mihir Desai, C. Fritz Foley, and James Hines of Harvard Business School found that domestic investment by U.S. firms grows by 2.6 percent for each 10 percent increment in the companies' investment overseas." It cited this as evidence of the economic benefits of sending work offshore. Had the editorial team been more adept at mathematical reasoning, it might have compared this 2.6 percent benefit to the growth of domestic investment by U.S. firms when they invest the same 10 percent increment domestically instead of in foreign subsidiaries. Which would be ... let's see ... carry the one ... that would be ... hmmm ... 10 percent -- nearly four times bigger. And why percentages? If a company has $1 billion in U.S. assets and a hundred bucks worth in Malaysia, then investing another ten bucks in Kuala Lumpur should yield $26 million here. Sign me up! The most important reason to learn more math (and science; the two go hand in hand) is, clearly, self-defense.
Lewis also correctly points out that correlation is not the same as causality:
The editorial's phrasing also strongly implied a causal relationship. Scientists are more careful. In the cited paper, the researchers were clear that correlation was the best they could manage.
Lewis' entire column is worth reading:
