Unblock America

Policy Research Prompt

This is the reusable prompt for researching one plank of the platform. Run it once per policy. Each run produces one file in this folder, named policy-NN-slug.md (e.g. policy-01-housing.md), numbered to match the plank order in the README and on the homepage.

These docs serve three audiences, in priority order:

  1. Us, in conversation. Talking points we can deliver from memory and defend when pushed.
  2. The ebook. Each doc becomes the first draft of a chapter, so citations must be real and checkable.
  3. A skeptical expert. An economist or journalist reading this should find it honest — including about the weak spots.

The prompt

Research the following policy from the Unblock America platform: [paste the plank title and full text from README.md].

Produce a markdown document following the template below exactly. Rules:

  • Primary sources first. Government statistics, academic papers, and legislative text over think-tank summaries; think-tank summaries over journalism. Every citation must be a real, checkable source with author/institution, title, and year. Never fabricate or approximate a citation — if you can't verify it, say "unverified" next to it.
  • Quantify or flag. Every empirical claim gets a number and a source, or an explicit "(estimate, weak evidence)" flag. A talking point we can't back up is a liability, not an asset.
  • Name the beneficiaries of the status quo specifically — companies, industries, lobby groups, with dollar figures on what they gain and what they spend on lobbying where available. Do not substitute vague villains ("elites", "Washington") for specific ones.
  • Be honest about losers. Every one of these policies creates transition pain for someone. Naming who loses, how much, and for how long is what separates this platform from marketing. If a compensation mechanism exists (like shifting ethanol subsidies to fruits and vegetables), evaluate whether it actually covers the loss.
  • Steelman the opposition. Find the strongest published arguments against the policy — from credentialed critics, not strawmen — and respond to each. Where the critics are right, concede it in writing.
  • Date everything. Statistics drift. Every number gets the year it was measured, and the doc gets the date it was researched.

Template for each policy doc

# Policy NN: [Title as it appears in the README]

Researched: YYYY-MM-DD. Status: draft | reviewed.

## The plank

[Verbatim text from the platform. If README and index.html disagree, note it — that's a bug to fix.]

## The problem

What is broken today, with numbers. How much does the status quo cost, who pays,
and how does that cost show up in an ordinary person's life?

## Who profits from the status quo

The specific companies, industries, and lobby groups that benefit; what they gain
per year; what they spend defending it. This is the villain section — it must be
factual enough to survive a defamation review.

## The proposal, concretely

What actually changes: which statute, regulation, or agency rule. What the
mechanism is (the Japan yōto chiiki model, the shortest-split-line method, etc.)
and how it works where it's already in use.

## Evidence

Case studies, natural experiments, and academic literature. For each: what
happened, where, when, measured how, and how closely it maps to the US case.
Rate the overall strength of evidence: strong / moderate / thin.

## Who wins, who loses

Distributional table. Winners and losers, sized in dollars and people where
possible. Transition costs and duration. Whether proposed compensation actually
compensates.

## Fiscal impact

Effect on the federal budget: revenue, outlays, net, over what window. Use CBO,
JCT, or academic scoring where it exists; label back-of-envelope estimates as such.
The platform's core claim is fiscal — this section is not optional.

## The opposition's best case

The 3–5 strongest published arguments against, each attributed to its strongest
proponent, each answered honestly. Concessions in writing where they're right.

## Talking points

- **One line:** deliverable in a sentence.
- **Thirty seconds:** the problem, the fix, the proof point.
- **Two minutes:** adds the villain, the evidence, and the honest cost.
Every factual claim in the talking points must trace to a sourced claim above.

## Open questions

What we still don't know, what evidence would change our mind, and what an
economist reviewing this plank should be asked to check first.

## Sources

Full citations with year and link. Mark each: primary / secondary / unverified.

Working agreements

  • One policy per doc, one doc per policy. Cross-cutting research (movement strategy, case studies of reform movements) stays in docs/ root, not here.
  • A doc is draft until a second pass has verified its citations resolve and say what we claim they say; then mark it reviewed.
  • When a doc changes what we believe, update the plank text in README.md and app/templates/index.html (both — they drift) in the same change.
  • The suggested order of research is by argumentative load-bearing weight, not plank number: start with the planks we cite most in conversation (housing, tax filing, healthcare–employment), and use what we learn to decide Policy 12.