| Bull Stat! |
25 August 2003
|
Fishy Statistics Make Economics a Dismal Science
Justin Lahart, CNN/Money Senior Writer, put together a beautiful piece last week on the significance of the 400,000 jobless claims that has everyone on Wall Street watching as an economic indicator. Mr. Lahart essentially performed the "Emperor's New Clothes" routine on the 400K figure, and backed it up beautifully. Briefly, he proved that there is nothing magic about the number. If only a few more such figures were attacked this way.
Pure and simple, Wall Street is a mix of greed and fear. The pros try to dress things up with GAAP figures, expert industry reports and "quant analysis," but not everyone in finance knows much about math. Thus, irrelevant numbers are often mistaken for significant facts, and business decisions go ahead based on factoids instead of knowledge.
Apart from the 400K jobless figure (above is recessionary, below is expansionary), core inflation springs to mind as a huge misleading non-fact. Core inflation refers to the inflation rate that excludes food and energy prices because of their volatility. Presumably, this may make statisticians happy because the curves are nice and regular, but the unpleasant truth is that food and energy are consumed everyday by everyone. Taking out other things may make sense, but not the essentials of life.
Per capita anything is another sort of figure that smells bad. If GDP is $20,000 per capita in a country of 1,000, the distribution of the income is very important. If evens spread out, the country is a pretty balanced society. On the other hand, if one man earns $20,000,000 and the rest have nothing, its a different story. Changing the per capita figure here might be meaningless, but if it stays the same, one can still have radical social change. Change "capita" to "store" and "GDP" to "company revenues" and the criticism holds at the microeconomic level.
This list is potentially endless. The problem lies in the fact that business and economics have numbers available to them -- prices. But price is a relative, not an absolute, concept. If apples are 4 for a dollar, and grapefruit is 1 per dollar, then apples are worth one fourth what grapefruit -- are these gold-backed 1900 dollars? Or 1970s inflating dollars? This sets limits to the usefulness of the numbers that sciences like physics and chemistry avoid (an erg is an erg is an erg). At that point, qualitative factors kick in. Foremost is the ability to think in critical ways about the data, as Mr. Lahart did.
Home