Follow the law.
Understand the effect.

Keyword-led discovery. Codex investigates one topic at a time, following mechanisms, related laws and references through a local legislation corpus. This is a selective research collection, not a complete census. No result does not mean no effect.

Direction and confidence are separate. “Worsens” and “improves” refer to the stated keyword criterion. They describe expected effects relative to a stated counterfactual, not necessarily measured outcomes or an overall verdict on a law. Mixed means material opposing effects; unclear means there is no defensible direction.

Check the source. Findings include quotations, causal reasoning, assumptions and countervailing effects. Historical and unverified legal status are labelled. Quotes are checked against the local snapshot; this does not establish that a provision is in force today. Counts represent findings, not economic magnitude.