This study proposes a novel econometric approach to estimating market power in homogenous product markets. We use a composed error model to estimate the stochastic part of firms' strategic pricing equation. This part is formed by two random variables: a traditional error term, which captures random shocks, and a random conduct term, which measures the degree of market power. This approach allows for the conduct parameter to vary flexibly across firms and within firms over time, and avoids ad hoc structural restrictions for identifying firms' conduct. The empirical application of our approach is based on a well-known California wholesale electricity market dataset, which has been rigorously used to study market power. Our results suggest that realization of market power varies over both time and firms, and reject the assumption of a common or time-invariant conduct parameter.