From: paul cockshott (clyder@GN.APC.ORG)
Date: Tue Nov 11 2003 - 05:39:07 EST
> > To measure it we need to do a least squares fit to price data: > > minimize with respect to x the square of (y - Mx) > > where x is a vector of hourly natural real wage rates, y is a vector > of recognized labor activity, measured by money, expressed in hours. > > In other words y is equal to a vector of firms' absolute value added > (nominal value added in dollars times the absolute value of money). > M is a matrix. The element M(f,t) gives the hours worked in firm f > by workers of skill type t. We need to have the number of firms much > greater than the number of skill types to get a good fit. > > The dimensions of y and M are hours. x is dimensionless -- wages > measured in money hours divided by clock hours. > > ------------------------------------------------- > Whatdo you expect these xs to be, >1 or <1? Phil: I would expect them to be dispersed around a mean of 1, but I have not tried prove that yet. ----------------------------------------- That would surprise me since wages on make up part of the value added. If you did get a mean of 1, you would now have a measure of exploitation by comparing your 'natural real wages' with the wages actually paid.
This archive was generated by hypermail 2.1.5 : Thu Nov 13 2003 - 00:00:00 EST