What Gets Measured

The question of what gets measured is prior to the question of what gets rewarded, because institutions cannot reward what they cannot measure and will routinely fail to notice what they do not measure. Volume III's analysis of the failure of existing metrics — its examination of how GDP, engagement metrics, and other dominant measures systematically discount regenerative value — establishes the diagnostic case for different measurement. This chapter addresses the constructive case: what Providence measures and how.

Providence's measurement architecture begins from the principle that what is measured must be genuinely connected to the constitutional values rather than serving as a proxy that, once optimized for, becomes disconnected from what it was originally tracking. This is Goodhart's Law applied to social institutions: when a measure becomes a target, it ceases to be a good measure. The appropriate response to Goodhart's Law is not to avoid measurement — unmeasured outcomes are unmanaged and institutional drift toward what is measured is structurally guaranteed — but to design measurement systems that are resistant to Goodharting.

The specific measures Providence requires include: relational health indicators (the quality and depth of the relationships within and between communities, measured through participant experience rather than through behavioral proxies); developmental progress indicators (the actual development of the human capacities — relational maturity, stewardship capacity, governance wisdom — that the architecture is designed to cultivate); ecological health indicators (the actual state of the ecological systems in the bioregions where Providence communities operate); and constitutional fidelity indicators (the degree to which the governance mechanisms are functioning as designed, measured through governance process audits rather than through outcome metrics that can be gamed).

None of these measures are easy to produce. All of them require ongoing human judgment rather than automated data collection. This is a feature rather than a bug. The measurement of what Providence actually values requires the kind of engaged human attention that automated systems cannot provide, and the practice of that engaged attention is itself a development of the capacities Providence is trying to cultivate.