Saturday, January 25, 2014

Chris Dillow — (Ab)uses of statistics


Chris Dillow links economics and power.
Macroeconomic data do not, and should not, over-ride individuals' local, dispersed knowledge of their personal situation.

You might think this is an Austrian point. But it is also a Keynesian one. Keynes described macro aggregates such as the price level or aggregate output as "vague and non-quantitative concepts" which are "unsatisfactory for the purposes of a causal analysis." To use a good if over-used metaphor, macro stats are only an imperfect map, not the terrain....

In this sense, Alasdair MacIntyre identified the purpose of macro stats - the claim to possess them is a claim to power: "What purport to be objectively-grounded claims function in fact as expressions of arbitrary, but disguised, will and preference" (After Virtue, p107). This sounds like a Foucauldian point, but it has been shared by people who wouldn't know Foucault from off stump....

If we focus upon macro data to the detriment of individuals' actual experience, we risk mistaking the map for the terrain. Which is one reason (of several) why Kenneth Boulding was right to say (pdf) that "almost all organizational structures tend to produce false images in the decision-maker."
(Use the embedded link above to download Boulding's "The Economics of Knowledge and the Knowledge of Economics.")

In a footnote, Dillow makes another good point.
Strictly speaking, "data" is the wrong word. It comes from the latin for "something given", but in fact macro data aren't given at all but rather constructed.
So-called data is the input into the black box of information processing and emerges as information. The data (given) is shaped by the data-gathering process, both limitations upon it — it's not possible to collect every detail — and selection criteria — choosing what to include and what to exclude. Then the input is processed into information using methods that involve further assumptions. As a result the output of information is only an approximation of the given from a POV, and the given itself is only an approximation of the sum of what may be relevant. For example, power is excluded from macroeconomic consideration.

Stumbling and Mumbling
(Ab)uses of statistics
Chris Dillow | Investors Chronicle

No comments: