2 
advocated by Roncaglia. As an example I mention the input-output 
models. They assume prices as given and deal with some 
problems of quantities, say a multiplier analysis, on this 
also 
bssis. Or they start from a given final output, as Sraffa 
does, and deal with prices and distribution. Again, all 
Kaleckian economics always work with given and constant prices 
and operate on the deflated values of macro-economic 
variables. In this sense, Kalecki certainly was a classicist! 
The simultaneous treatment of quantities and prices involves 
non-linearities: The flex price relations are typically 
non-linear ( agricultural and other primary commodities) 
bottle-necks generally involve non-linearities. For an 
abstract equilibrium theory this involves no trouble. 
But for an applied economist such simultaneous equation systems 
are not very helpful. He has to start from given initial 
conditions and work out the process as it evolves in time, 
for example by means of difference equations, j This is 
what Paul Davidson fimpjifes when he says that economic processes 
are not ergodic, that is, they do not lead to a steady state 
independent of initial conditions ( I should add that even 
if they are ergodic they do not converge quickly so that 
they never get old enough, in practice, to reach the steady 
state, being again and again interrupted by disturbances from 
outside). Now if you try to work with difference or other 
functional equations non-linearity will not be easy to deal 
with. I have no recipe to offer, unfortunately. 
In the field of micro-economics ( information approach ) 
Streissler has haitiLy given us much encouragement to pursue