[ show plain text ]
Strongly second Julian's comments re (1).
If people aren't aware of this critical literature re "social welfare
functions", then they should check it out before promoting the concept of a
socialist SWP. Briefly, the literature concludes that a SWF can be
constructed only under the following conditions:
(a) that all people have the same preferences
(b) that preferences do not change with income.
Cute: in other words--though neoclassicals don't put it this way--you can
have a meaningful SWF (which can rank any two combinations of commodities
and unambiguously say which is preferable) if society consists of one
individual, and there's only one commodity to consume.
The one out which may apply in this debate is the third, independent way
out: (c) that incomes are fixed.
This may sound appealing from a socialist point of view, whereas it's not
from a (neoclassical) capitalist one, since in the latter changing prices
will change incomes (hence neoclassical rationalisations of this failure of
their Benthamite project choose the former two absurd escape routes over
this one!). However, I doubt that the analysis holds water once you move
out of the static framework in which SWF analysis occurs into dynamic
analysis--let alone if you make preferences and commodities co-evolutionary.
In a nutshell, I think that one of the last things Marxist economists
should be doing is legitimising a line of economic analysis which even
neoclassical economists have shown to be seriously shonky!
If you want to chase this up, some of the relevant references are:
Gorman 1953, Community Preference Fields (lost journal reference!)
Sonnenschein, H, 1973, Do Walras' Identity and Continuity Characterise the
class of community excess demand functions, Journal of Economic Theory
1972, Market excess demand functions, Econometrica
1973, The utility hypothesis and demand theory, Western Economic Journal
1982, Market demand and excess demand functions in Arrow, K.J, Handbook of
Debreu, G., 1974, Excess Demand Functions, Journal of mathematical economics
Note the names, by the way: these are the high priests of neoclassicism. If
they've shown that something is shonky, then it's shonky--despite their
failure to interpret it properly.
At 11:55 AM 3/2/00 -0000, you wrote:
>I think Duncan has posed three issues here:
>1) a socialist society could develop a social welfare function (SWF)
>2) such a function could/should be maximisable (or minimisable; e.g.
>minimising necessary labour-time)
>3) it would generate numbers (distributed along some single scale) which
>administrators could use for allocating resources without further reference
>to those concerned.
>>From Paul (and Allin's) past comments and their published writings, I gather
>that they would subscribe to all three of these (in the case of (2), to the
>version in brackets).
>I'm very sympathetic to the view that the role of socialist economic
>planners should be to minimise necessary labour-time, which I think commits
>me to something like (3): minimising general (but, contra Paul and Allin,
>not abstract) labour, which is evidently just measured in hours.
>My problem, however, is with (1), and hence with (2), or at least with
>versions other than the one in parentheses.
>There's a standard literature, starting with Arrow, which suggests one
>should at least be sceptical about the possibility of an SWF -- and, apart
>from this, the whole conception of an algorithm to tell planners what they
>should tell people to do seems anti-participatory and anti-democratic.
>Why can't people just get together from time to time and decide what seems
>to best in the circumstances (no doubt *advised* by specialists)?
Dr. Steve Keen
Economics & Finance
University of Western Sydney Macarthur
Building 11 Room 30,
Goldsmith Avenue, Campbelltown
PO Box 555 Campbelltown NSW 2560
email@example.com 61 2 4620-3016 Fax 61 2 4626-6683
Home 02 9558-8018 Mobile 0409 716 088
Home Page: http://bus.macarthur.uws.edu.au/steve-keen/
Workshop on Economic Dynamcs: http://bus.macarthur.uws.edu.au/WED
This archive was generated by hypermail 2b29 : Fri Apr 21 2000 - 09:47:56 EDT