Humans are kin-selective pack omnivores who use economic exchange and display as a big part of mate selection for themselves and close kin, mostly descendants. As pack omnivores humans are also quite good at creating group fictions such as nations and businesses to increase, on average, the net returns of the participants. Economic policy ignores two things as irrelevant: mating display as an economic goal and kin selection. Up until 1968-82 two simplifying assumptions, group selection and generally non-economic mating decisions, worked well enough to create generally competent economic policy. A third simplifying assumption of ignoring how new industries operate is where the problems started. Mature industries have known economies of scale, scope, network and various other minor economies. New industries have to discover economies in production, supply and distribution. This process of discovery is a step function of Econ 101 results and sudenly a cost drop in supply, distribution or production results as a new economy is found. In rare cases like Moore's law of decreasing computer memory costs such increasing marginal returns become predictable. Increasing marginal returns are not merely ignored but outright forbidden in economic models. It is forbidden because it makes both mating display and group selection assumptions much less useful. Since 1982 ignoring increasing marginal returns has not worked and that's what's wrong with economic policy.