A recent survey that reviewed scores of research studies on franchising
shows that the “real world” accords well with the theoretical models.^22 (The
studies covered all business sectors having a significant incidence of franchis-
ing: traditional and fast-food restaurants, gasoline service stations, hotels and
motels, retail chains, and so on.) Across all these sectors, the incidence of fran-
chising consistently increased when the agent’s effort (rather than other
uncontrollable factors) was the main determinant of output and profit.
Conversely, researchers found a greater incidence of workforce integration
when the firm could monitor employee effort levels and decisions at low cost.
(In other words, franchising’s high-powered incentives were unnecessary when
effort could be monitored and controlled directly.) Finally, researchers found
that franchising was more prevalent in sectors where management decisions
depended on information about local market conditions. As noted earlier in
Table 14.2, the dispersion of decision-relevant information favors decentral-
ized decisions, that is, external franchising.606 Chapter 14 Asymmetric Information and Organizational Design(^22) See F. Lafontaine and M. E. Slade, “Incentive Contracting and the Franchise Decision,” Chapter 5
in K. Chatterjee and W. Samuelson (Eds.), Game Theory and Business Applications(Boston: Kluwer
Academic Publishers, 2001).
CHECK
STATION 4
Using the facts in Table 14.3, confirm that the worker’s profit share must be just greater
than 60 percent to ensure that he adopts an efficient level of effort.
EVALUATING INDIVIDUAL PERFORMANCE An understanding of incentives
in the principal-agent problem leads to an additional result.
All information bearing on an individual’s effort and contribution to profit should
be included in the measure of performance. Any variables not bearing on effort or
profit contribution should be excluded.
This proposition is sometimes called the informativenessprinciple. In other
words, the more precise the measure of performance (combined with the
appropriate incentive structure), the more efficient will be the agent’s behav-
ior (and the smaller will be the resulting agency cost). The proposition makes
intuitive sense. Pertinent performance information should not be ignored, nor
should irrelevant information be included.
The difficulty, of course, is in accurately monitoring and measuring per-
formance. Frequently, it is difficult to disentangle and identify the contribu-
tion of a particular worker. Rather, the worker’s effort and performance
influence the firm’s profit in conjunction with the contributions of many
other workers and in combination with market forces. First, imperfect per-
formance measurement reduces the incentives for efficient behavior. Second,
using aggregate measures exposes the worker to significant risks in the com-
pensation scheme.
c14AsymmetricInformationandOrganizationalDesign.qxd 9/26/11 11:03 AM Page 606