On risk aversion, classical demand theory, and KM preferences