Client. A national legislative body, where every bill is an intervention on a living economic and social system —and where it is almost never measured who wins and who loses before voting on it.
Approach
This is perhaps the case where the method shows up earliest and in its purest form: the two left-hand links of the bridge —the sociology of organizations and data engineering— working together on the most political material possible. Here the “organization” being read is the economy itself, understood not as an aggregate of numbers but as a system of actors and networks with cross-cutting incentives, asymmetries of information and power, winners and losers.
It is the root of the bridge thesis: a technical intervention —a law, a system, an automation— can only be understood if you first read the human fabric it lands on. An aggregate number averages that fabric until it disappears; distributive analysis brings it back.
The problem identified
Legislative debate used to rest on the aggregate impact of a rule: how much a macro variable rises or falls. But a law almost never affects everyone equally: it redistributes. The aggregate hides exactly what matters politically —who it benefits, who it harms, what incentives it creates and what behaviors it induces. The problem identified was that gap: deciding on averages that conceal the distribution.
Functional assessment
The assessment consisted of modeling, for each bill, the map of actors of the economic process the rule touched: who participates, with what incentives, what market relationships connect them, where there are asymmetries and concentrated power. It is sociological work before it is econometric: defining the system well before measuring it, so as not to measure the wrong thing precisely.
Building the solution
- Quantitative and econometric analysis of distributive impact: estimating, beyond the aggregate effect, how it is shared among the different actors and segments.
- Reading market structure and social structure: identifying concentration, asymmetries and the mechanisms by which a rule shifts advantages from some to others.
- Automating the processing in Python, to repeat the analysis on each new bill without rebuilding it from scratch.
Information and data layer
The value was not only in the model, but in making it legible to those who decide and do not read econometrics. I built dashboards (Power BI, Tableau) that translated the distributive analysis into a reading that a legislator or their team could use to deliberate. In government, an analysis the decision-maker does not understand is an analysis that does not exist: the translation is part of the work, not an ornament.
How the work was conducted
The work predates agentic tools: the method was econometric and hand-crafted, automated in Python to gain reproducibility. I record it as an honest reconstruction of an early career stage —I do not disclose the body’s specific data or analysis—; what I do assert is the method, because it is the seed of everything that came after: reading a human system as a system of actors before intervening on it.
What this case proves
- The origin of the bridge: the sociotechnical perspective applied to public policy, years before naming it that way.
- Quantitative rigor with social purpose: econometrics put to answering “who wins and who loses”, not to impress.
- Translation to the decision-maker: from the model to the dashboard a non-technical person can use to decide better.