This is a problem with much of mainstream economics. Start by defining everything worth measuring as the things you can measure, then build elaborate models around those measurements, then declare success when your models yield an output. Finally, influence policy to try get more of one measurable and less of another.
