Skip to content

Impact Assessment Table

Governance register — instantiates Deadweight Loss Reduction

A single comparable grid that lays each affected party's estimated gain or loss beside the protected constraints and the post-change signals to watch, so governance can see the incidence at a glance.

Deadweight-loss language is unusually good at hiding people. "Net welfare rises" can be true while a specific group is made distinctly worse off, and the aggregate number gives a reviewer nowhere to look. Impact Assessment Table is the artifact that fixes this by rendering rather than computing: it lays out one row per affected party — who bears the distortion today, who benefits from it, who gains or loses from the repair, who is only indirectly touched — and puts beside each row the estimated gain or loss, the protected constraints in play, and the specific signals to watch after the change lands. Its defining contribution is legibility for human governance: it does not decide whether a repair is worth it (that is the Cost–Benefit Assessment Protocol) and it does not map the wedge (that is the Distortion-Reduction Review); it makes incidence and the monitoring plan visible on one page so a board, council, or regulator can review the distribution of harm instead of taking a summary on trust.

Example

A state is weighing an expansion of pharmacists' scope of practice — letting them prescribe a defined set of routine medications directly. The analysis exists; the question is whether the elected board can see what it does to whom. The Impact Assessment Table gives each stakeholder a row. Patients: faster, cheaper access, especially where clinics are scarce — a large gain, some of it in rural rows that a state-wide average would erase. Primary-care physicians: fewer routine visits — a real loss, concentrated on a specific group. Pharmacists: added revenue and added workload — a mixed row. Payers: ambiguous, flagged as uncertain rather than guessed.

Across the top run the columns that keep the repair honest: a protected-constraint column (diagnostic safety, the boundary that must not be crossed) and a monitoring-triggers column naming exactly what would signal the repair is going wrong — an adverse-event rate above baseline, a referral-rate collapse suggesting missed diagnoses. The table does not tell the board what to do. It tells them who wins, who loses, what must be protected, and what to watch — and it makes the rural access gain and the physician loss impossible to read past.

How it works

  • One row per affected party. Include the losers, the indirectly affected, and the intangibles, not just the intended beneficiaries — the rows a headline number would collapse are the ones the table exists to preserve.
  • Comparable cells, honest blanks. Put every party's effect in the same terms where possible and mark genuine uncertainty as uncertainty rather than filling it with a confident guess.
  • Carry the protected constraints alongside. A column for what must not be sacrificed keeps the efficiency story next to the safeguard it could quietly erode.
  • Name the post-change triggers now. For each row, state the observable signal that would show the repair is producing rebound, exclusion, or quality loss — so monitoring is designed before the change, not improvised after.

Tuning parameters

  • Row granularity — how finely affected parties are split (one "consumers" row versus rural, low-income, and elderly sub-rows). Finer rows surface concentrated harms an average hides, but a table too fine to read protects no one.
  • Effect vocabulary — whether cells are dollars, qualitative ratings, or mixed. A common unit aids comparison; forcing intangibles into dollars can bury the very effects the table is meant to expose.
  • Uncertainty display — how visibly disputed or unknown effects are flagged versus point-estimated. Louder uncertainty resists false confidence; too much noise and reviewers stop reading.
  • Trigger sensitivity — how large a post-change movement counts as a monitoring trigger worth acting on. Set tight and every wobble alarms; set loose and real harm accrues before anyone notices.

When it helps, and when it misleads

Its strength is that it makes distribution and aftercare reviewable by humans who are not analysts. A governance body can scan rows and ask "why is this group's loss acceptable?" and "what happens if this trigger fires?" — questions an aggregate net-welfare figure silently forecloses. It is also where the monitoring plan gets committed to paper before the change, which is what distinguishes a genuine repair from efficiency theater.

Its failure mode is that a table looks authoritative — clean cells imply settled facts. A row can be filled with a fabricated-feeling precision that launders a guess into a number, uncertain rows can be quietly point-estimated, and inconvenient parties can simply be left off the grid, which is the table's version of running the analysis backwards. The tell for that last misuse is a table with no meaningfully harmed row — real repairs almost always have one. The discipline is to require that every plausible loser appears, that uncertain cells stay visibly uncertain, and that the person who drew the rows names who is not on the table and why.

How it implements the components

Impact Assessment Table realizes the archetype's transparency-and-aftercare components — the ones that make effects legible and trackable:

  • affected_party_incidence_map — its core structure: a party-by-party layout of who bears the distortion, who benefits, and who gains or loses from repair, with distributional effects kept explicit.
  • monitoring_and_rebound_check — the trigger columns define, up front, the observable signals that would reveal rebound, exclusion, or quality loss after the change, so post-change tracking is designed rather than improvised.

It does NOT weigh the ledger or run the sensitivity case (cost_benefit_assessment_frame, sensitivity_analysis, distributional_review) — that is the Cost–Benefit Assessment Protocol; it does not map the wedge or size the surplus (distortion_map, surplus_estimate) — that is the Distortion-Reduction Review; and it changes no rule (redesign_lever).

Notes

The table is a display, not an analysis: its cells are only as good as the estimates poured into them from the review and the cost–benefit protocol. Its distinct value is entirely in arrangement — the same numbers a spreadsheet already holds become governable once each affected party has a row and each row a monitoring trigger. Treat a polished-looking table with no harmed row as a warning, not a result.