ABSTRACT
There are two standard arguments for disaggregated modelling of the company sector. The first is that the modeller — perhaps shareholder, takeover bidder, or union negotiator — is primarily concerned with the performance and prospects of an individual company (see Goudie and Meeks, 1981, 1982). The second applies to the modeller who is primarily concerned with the sector aggregate: the diversity of company behaviour may be so great that disaggregated modelling is the most efficient way to capture sector behaviour. 1 In this paper a third argument is presented: even if the modeller is interested only in the sector and not in the individual agent, and even if all companies are identical in their response to given market stimuli, nevertheless threshold effects in companies' responses may still make disaggregation the most efficient route to aggregate conclusions.
