Ventana software support forum

Dr. Duggan,

You should get a fair number of responses to your question, because much
work has been done modeling commodity cycles with the System Dynamics
stock-and-flow approach. John Sterman devotes an entire chapter of his book
Business Dynamics (2000, Irwin-McGraw Hill, see Chapter 20) on these
formulations, built upon the work of Dennis Meadows (he modeled livestock).
The references in BD are as follows:

Meadows, DL. (1970) Dynamics of Commodity Production Cycles. Waltham, MA:
Pegasus Communications.
Weymar, H. (1968) Dynamics of the World Cocoa Market. Cambridge, MA: MIT
Press.
Guvenen, O., W. Labys, and J-B Lesourd. (1991) International Commodity
Market Models: Advances in Methodology and Applicatios. London: Chapman and
Hall.

I and many other System Dynamicists find the classic econometrics models of
simultaneous equations problematic, beyond the usual problems of
identification of model parameters. John Sterman provides the following
reasons on p. 798 in BD:

(1) Econometric models do not represent the stock and flow structure of
these markets (inventories, WIP, production capacity).
(2) They are formulated in discrete time, which may lead to spurious
dynamics. The time period is usually dictated by the intervals over which
data are reported.
(3) It is assumed that the commodity can be produced within the chosen time
period (e.g. indicated production in period t-1 is realized in period t).
(4) They do not distinguish between production capacity and capacity
utilization.

To summarize, John states, ""Proper commodity models, like all dynamic
models, should represent the stock and flow structure, time delays, and
behavioral decision processes of the market. The delays should be set to
their actual values, not multiples of some arbitrary time period, and the
models should be formulated in continuous time.""

I hope this helps,
Tim Quinn

System Dynamics Group
Massachusetts Institute of Technology
Sloan School of Management

30 Wadsworth Street
Bldg E53, Rm 358A
Cambridge, MA 02142

Telephone: 617-258-5585
Email: tdquinn@mit.edu