Tuesday, January 20, 2004

Complexity: Devious Management and Investment Risk

A common feature of much recent malfeasance is the construction of highly complex corporate structures, apparently beyond the legitimate demands of Requisite Variety.

Complexity – whether requisite or otherwise – adds to investment risk. Dodgy management practices may cause a continuous erosion of corporate value, or they may trigger a sudden collapse of value (Enron, Parmalat). Some companies have a sufficiently robust business model that they remain viable even with a certain level of malfeasance. Other companies turn out to be merely pseudo-viable – only remaining solvent thanks to dodgy accounting. Some investors may be willing to tolerate erosion, but do not wish to be confronted with sudden collapse.

Complexity is (or should be) a warning sign. The purpose of complexity is what it does. If it doesn’t serve a legitimate purpose, then it is surely reasonable to assume it is there to serve some other agenda.

Using the theory of complexity, we should be able to construct geological maps of the corporate world, showing (probabilistically) where it might be worth drilling for the next accounting black hole.

For example, it now seems that Parmalat was non-viable, only sustained in pseudo-viability by paper cashflows from a non-existent bank account. But such transactions can only be concealed by having lots of apparently genuine intra-company transactions. There is therefore a control mechanism that forces complexity onto the company at the operational level, and a higher level mechanism that manages the smoke and mirrors. It might not be easy to detect the fraud by looking at the operational company alone; it may be the existence of the control mechanisms that gives the game away.

Conversely, if a management is obliged to construct and present evidence of its bona fides, this evidence needs to include a properly grounded account of the control mechanisms, including a justification of the degrees of complexity and intracompany coupling.

Update

See new Squidoo lenses on Complexity and POSIWID