On Wilde's theory of risk homeostasis