This paper starts with the basic premise that the conventional measures of productivity growth, which ignore joint production of good and bad outputs, are biased. We then construct an alternative productivity growth measure using activity analysis. An application to U.S. agriculture demonstrates its usefulness. More specifically, we show that the Tornqvist index of productivity is biased upward when production of undesirable outputs or "bads" is increasing. Conversely, this same measure of productivity is biased downward when externalities in production are decreasing.