Wallyver raises a question about the seeming parallel between my remarks about index numbers and my discussion of accounting problems in Part Three of my paper, THE FUTURE OF SOCIALISM. First of all, let me say that until he asked this question, the parallel had never occurred to me. It is very interesting.
My point about index numbers was this: because there is no solution to the index number problem -- i.e., no objective, ideology-neutral way to form such measures of social reality -- every effort to grasp conceptually the nature of social reality is unavoidably ideological. This reveals a deeper truth, which is that social reality itself is ungetoverably normatively structured, embodying [in a way that physical reality does not] a perspective or point of view or normative dimension, or indeed more probably several that are in conflict with one another. From this insight, I draw the conclusion that a value-free social science is impossible [and that includes economists as well as sociologists and political scientists, of course].
The point about the accounting discussion is different, but in many ways strikingly similar. Because there is no objective "scientific" solution to the accounting problem once capitalism develops beyond its most primitive stages, supposedly objective decisions concerning profitability in a firm inevitably turn into quasi-political decisions or negotiations among the interested parties. This, I suggest, is part of the process anticipated by Marx by which the new social relations -- of socialism -- grow in the womb of the old capitalist order. If I am correct, then profit-maximizing capitalists have no choice but to engage more and more in decision making whose structure is political rather than, in the old sense, economic, and this fact robs the defenders of capitalism of their strongest argument, which is [pace Hayek] that market-driven decisions are necessarily the most efficient. The problem is not that they are less efficient, but that they are impossible. [By the way, for those who are really into this sort of thing, this is exactly the criticism that Oakeshott has of rationalism in politics, namely that it is impossible, not bad.]
Does that clarify things a bit?
I think there is a useful distinction between your argument for the inevitable ideological background to social analysis and my point about index numbers. Your unemployment blog some months ago about the Waltons, the Galt(s), and the Marxes highlighted the social meaning attaching to the concept of unemployment, and the consequences of this for trying to measure unemployment.
ReplyDeleteI was referring to a narrower issue of indeterminacy, where it is in principle impossible to identify a single number as correct. In standard examples of index number theory, Paasche and Laspeyres indices constitute bounds, typically lower and upper respectively. If the bounds are narrow, then this indeterminacy may not matter much. For example, with adjacent years of a time series, the Paasche and Laspeyres typically differ by less than 1%. For cross-sections, such as firms in the same industry but different regions, the two can differ by 50%. (This comes from one of Erwin Diewert's survey articles.) Economists worry less than accountants about small numerical discrepancies; I could live with 1%, but a 50% divergence would be troublesome.
Not having read Thomas on accounting, I'm not sure which of these two interpretations would more appropriate for him. But for index numbers, let me give another example. There has been quite a bit of research work done on "democratic" and "plutocratic" consumer price indices. Standard officially-published CPIs are typically plutocratic, i.e. richer households have higher weights because they spend more. "Democratic" indices weight sample households equally. The choice between these two indices isn't indeterminacy in my sense, but rather a political one in your sense. Though my indeterminacy point would also apply to each of the democratic or plutocratic indices.