Assume managers have two sources of knowledge, self (S) and organization (O) (as in March's model), and two subjects they act to benefit, also S and O. A manager may be controlled at any given moment by S knowledge and the S subject, or O knowledge and the O subject, but not by both S and O. Whether S or O controls the manager at any given moment is determined by a stochastic process in which the probability that S or O will have control is affected by votes for S or O of an internal and external electorate E that reacts to S and O actions.
Some positive outcomes result from the assumed "two party" system with the manager divided into separate S and O control centers subject to an electoral check, as opposed to a "one party" system with a single control center acting in its own interest:
1) Benefit to organization of being able to learn from S knowledge (as in March's model); S knowledge will be more readily available to the organization if it is bundled with S subjectivity that benefits the actor than if S knowledge has no benefit to the actor and is hence repressed in favor of adopting the organization's knowledge;
2) Benefit to individual and organization of having actions aligned with S or O rather than with the interests of a single control center, which may accord with neither S nor O;
3) Benefits to individual and organization of having both S and O checked from taking value-reducing actions by the part of E that represents a "swing vote" not committed to an S or an O position.
On the other side of the ledger, there are some negatives associated with a divided managerial self with S and O control centers and an electoral process. The obvious drawback consists of the greater costs of an individual having two competing control centers and an electoral process, compared to having only one permanent control center and no elections. The other major problem with S and O rivalry inheres in the possibility that the electoral process may be manipulated by S or O so that it does not align S and O with value-enhancing actions that take the other side or the electorate as a whole into account. If either side can hijack the electoral system so that some or all votes for the other side are not registered, suboptimal outcomes will result that take advantage of O for the benefit of S or the other way around. Though these outcomes may be better than the outcome from a single control center within the individual that acts to benefit itself rather than either S or O, they are inferior to the value-enhancing outcomes that will result from a well-functioning electoral system that accurately registers all relevant internal and external reactions to S and O actions.
Comments