12
Mandatory vs. Voluntary Approaches to Food Safety
consumer chooses a consumption level of y to maximize the perceived
expected net benefits from consumption, i.e., to solve
(7) Max B(y)-Py-(l-s)q Dy,
where the last term in (7) is the perceived expected damages that will not
be compensated by the firm and hence will be borne by the consumer. For
an interior solution, the choice of y solves:
(8) B'(y) -P-(l-s)q D = 0.
Similarly, the firm chooses the levels of y and q to maximize perceived
expected profit,35 i.e., to solve36
(9) Max P(q)y - C(q,y) - sqBDy,
yielding the following first-order conditions:37
In this case, to achieve an efficient solution, one generally needs two policy
instruments, one to create the correct incentive for risk reduction and the other to
ensure efficient risk allocation. If transfer mechanisms are available to ensure
efficient risk allocation, then risk reduction incentives can be modeled as though the
parties were risk neutral (see Miceli and Segerson 1995). If only one instrument
(e.g., a single liability rule) is available, then the instrument must be designed to
address both distortions. Unless it can simultaneously satisfy both objectives, a
first-best outcome will not be possible.
35 This is clearly a stylized model of producer behavior. A more detailed model
would distinguish explicitly between alternative strategies that firms could use in
response to consumer demand for food safety (see Caswell and Johnson 1991).
36 This specification assumes that the only losses the firm might incur as a result
of a contamination episode are those related directly to victim damages (D). We do
not explicitly incorporate potential losses due, for example, to consumer boycotts or
other manifestations of public outcry. In practice, these costs could far outweigh
any direct Hability payments the firm would have to make (Caswell and Henson
(1997)), and inclusion of them would yield greater incentives for investment in food
safety. We also model a single producer. In reahty, the production and processing
of a product might involve a chain of producers. For example, pesticide residues on
food stem from the production activities of pesticide manufacturers and the use
activities of farmers. See Segerson (1990) for an explicit consideration of safety
incentives in a model involving a chain of producers.
37 We assume here that the firm is perfectly competitive in its output market.
Imperfect competition would add still another source of distortion into the analysis,
which would genera∏y require an additional policy instrument for efficiency or
analysis of second-best outcomes. Note, however, that to the extent that the firm
overproduces the good due to imperfect internalization of damages (see below), the
existence of market power and the associated underproduction incentives would be
efficiency improving. See Bamett (1980) for a related discussion in the context of
environmental extemahties.