Should law respond readily to society’s evolving views, or should it remain fixed? This is the question of entrenchment, meaning the insulation of law from change through supermajority rules and other mechanisms. Entrenchment stabilizes law, which promotes reliance and predictability, but it also frustrates democratic majorities. This paper uses economic theory to study this tension. It argues that nearly all laws should be minimally entrenched. This is because bare majority rule can systematically harm society - even when voters are rational, and even when no intense minority is present. Then it argues that minimum entrenchment is conceptually straightforward but optimal entrenchment is not. It depends on factors like whether the costs of legal instability are variable or fixed and who has power to set the agenda. Existing scholarship ignores these factors, casting doubt on its prescriptions. The paper provides guidance for legal designers, and it has implications for constitutional law, including Article V, amendments, conventions, and judicial updating.