These results can be seen immediately by simply comparing the equilibrium investment
levels under the different ownership structures. There are some subtleties in the com-
parison between the employee cooperative and the partnership. When both employees
are sufficiently similar ex-ante . i.e.. α ≈ ⅜. then the average employee's investment level
is similar under both ownership structures. However, the variance of each employee's
investment level is higher under the partnership. Since the production function is strictly
concave the joint-employee ownership structure which has a lower variance is better. How-
ever. when employees are sufficiently different ex-ante. i.e.. ∣α — ⅜∣ » 0. the partnership
may dominate the employee coop. Since allocating the ownership to the better employee
will provide more incentives for the better employee than the employee ∞op.
An important difference with the case of no competition is that now outside ownership is
always the worst ownership allocation. It either gives rise to the worst underinvestment or
to the worst overinvestment. Thus, under internal competition between employees, third
party ownership is always dominated by other forms of ownership. The striking result
that the customer cooperative achieves the first best depends critically on a condition
that the customers are homogeneous that they all like the product of the firm. We will
discuss the case where this condition is relaxed in next subsection.
Moreover, another important implication of our theory is that the optimality of own-
ership structure may depend also on the job requirement, i.e. the qualification of the
employees. If the qualification requirement is low such that if people are about the same
in their talents for fulfilling the tasks, then allocating ownership to all the employees
is the second best. This result may shed some lights on man}* of the observations in
the literature. For example, in plywood cooperatives, worker-owners are semi-skilled and
commonly rotate over time through the various jobs: managers are generally* not members
of the cooperative, but rather hired as salaried employee (Greenberg, 1984). Similarly,
many car rental firms are owned by drivers jointly (Hansmann. 1996).
If the qualifications required in a firm are more related to peopled talent which is no easy
to identify* ex-ante, that is employees are heterogeneous in qualifications: moreover, an
employee's performance in the past may signal that he∕she is more likely to perform well
in the future, then allocating ownership to those employees who performed well in the past
is the second best. This may explain why partnership is the dominant ownership in the
service professions, such as law. accounting, investment banking, consulting, advertising,
architecture, engineering, and medicine (Hansmann. 1996. pp.66-69). Here, partners are
the employees who have performed well in the past. Take example of law firms, the
partners in a law firm are more qualified lawyers with higher skill and productivity, and
17