A new technique for estimating market power in several markets simulta
neously is developed and applied to the Australian retail beef, lamb,
and pork markets. The hypotheses that market power is zero and that ma
rket power is the same for each meat cannot be rejected. Nor is there
evidence that market power increased over the period of analysis. Litt
le bias is created by examining markets is isolation, rather than with
in a system, when markets are competitive, but that bias can be large
when market power exists in some markets in the system.