7 - An index-number tournament

Summary

For the Colonel's Lady and Judy

O'Grady

Are sisters under their skins!

Kipling: The Ladies

Use every man after his desert, and

who should 'scape whipping?

Hamlet, Act II

This essay tries to answer three simple questions:

1. Is a Soviet-type index of industrial production as bad as it looks?

2. Is a Federal Reserve-type index as good as it is reputed to be?

3. Is the latter index clearly superior to the former?

Introduction

The indictment of the Soviet official index of industrial production, as drawn by our experts, usually runs as follows:

1. the use of the 1926–27 weight base long since obsolete;

2. the introduction of new products into the index at inflated prices;

3. the use of value-of-output (or of price) weights instead of the more respectable value-added weights;

4. miscellaneous offenses, such as the exclusion of small-scale industry in early years, unreliable reporting of original data, arbitrary prices, use of multiple prices since 1952, if not earlier, and many others.

Of all these indictments, I will deal here with only the third – the use of value-of-output, or of prices, as weights, all others having been thoroughly discussed elsewhere. My concern with this particular item stems not from its numerical importance – if anything, our experts tend to minimize its effects as compared with those of the other transgressions and are even a bit uncertain about its sign – but from plain curiosity about an aggregate index of industrial production with such obviously wrong weights, a curiosity which has remained unsatisfied because of the scarcity of analytical literature on the subject.