Previous research on price determination for non‐ferrous metals at the London Metal Exchange (LME) suffers from three limitations: first it has employed single equation methods only, which cannot explain the simultaneous determination of spot and futures prices; second, by focusing on current and lagged prices, previous research does not analyse the effect on price determination of critical variables such as expectations, consumption and inventories; third, the outcome of prior research regarding market efficiency is ambiguous. This paper, which addresses these issues, develops a simultaneous model of the copper market at the LME, with representation of the activities of hedgers, speculators and consumers. This model produces post‐sample forecasts of the spot price which outperform conventional benchmarks, thus providing evidence against the efficient market hypothesis. Model‐derived forecasts are employed as the foundation of a trading program which produces risk‐adjusted profits (net of commission costs) for holding periods of one week and one month, thus fulfilling the ‘sufficient condition’ for market inefficiency. This study, therefore, provides new insights into price determination on the LME copper market, and resolves the ambiguity of previous research regarding the efficiency of that market. This is the first application of the model forecasting approach to the question of performance of the market for copper.