Pre-registration: How much can we learn from monetary policy surprises in Latin America? Contamination of swap-based surprises, predictability of survey surprises and statistical power

Pre-registration (in Spanish), written on 20 October 2026 before constructing the survey-based monetary policy surprise for Chile, before relating the Choi-Willems-Yoo (2024) swap-based surprise for Chile to the expectation in the Banco Central de Chile's financial operators survey, and before computing the number of meetings or countries required for adequate power. It fixes the definitions (meeting-by-meeting survey surprises for Colombia, Brazil and Chile; orthogonalisation with respect to pre-announcement information; minimum detectable effects by circular shifts) and six predictions with their refutation criteria: that swap-based surprises load on the expected component of the decision, that survey surprises are barely predictable, that the minimum detectable effects exceed the effects reported in the literature by a wide margin, the number of years or countries needed for 80% power, and how a central-bank study reporting 68% bands reads at 90%. No response of activity or prices to any surprise is estimated.