After having had a few requests, here are some references for the rigorous modeling
of market shares. For SDers this means application to marginal (new) market share
only, even through econometricians use this stuff for the average, straight away.
Rationality, "irrationality," and information-uncertainty come naturally with this
method. Read the texts like a SDer and not like an optimal-world-type economist
and you will see that there is a causal interpretation that works. Daniel McFadden
is the "Jay Forrester" of all this. His work is not for the mathematically
challenged. Note that OLS do not work for parameter estimation; MLE or Kalman
Filtering is needed.
The book by Keeney and Raiffa makes sure that the equation/model of the
decision-makers "utility" function is validly specified. Again, love math or leave
it.
Our experience over twenty years and hundreds(?) of case studies has been that
these methods have been shown retrospectively to be very, very accurate and the
causal interpretation of the results legitimate (and possibly even valid). Our
feedback model forecasts (that include this market share approach) have often had
tenths of percent error over 10 years! We have had no blind-sided forecasts.
The nearest comparison with a "strong" economtric forecast showed a 30% maximum
error to our 0.56% maximum error. (Still our model is this later instance was
eventualy discarded in favor of the "bad" model because everyone was familiar with
the bad one! Yes, retrospectively, our market share model did indicate that we
would lose this market share. We just never thought about the feedback on
ourselves. Its the pits to be part of the feedback system you model.)
Here are the refenences:
Keeney, R. L. and Raiffa, H., Decisions with Multiple Objectives, John Wiley &
Sons, New York, NY, 1976.
McFadden, Daniel L., "Conditional Logit Analysis of Qualitative Choice Behavior,"
in Frontiers of Econometrics, P. Zarembka (ed.), Academic Press, New York, 1973.
Qualitative Choice Analysis
Kenneth Train
MIT Press 1986
Discrete Choice Analysis
Moshe Ben-Akiva and Steven Lerman
MIT Press 1985
Structural Analysis of Discrete Data With Econometrics
Charles Manski and Daniel McFadden
MIT Press 1986
Good reading and mind-opening.
G.
George Backus Email: gbackus@boulder.earthnet.net
Policy Assessment Corporation phone: (303) 467-3566; fax: (303) 467-3576
14604 West 62nd Place Denver, Colorado 80004, USA