When we publish an error, we correct it here. This page is the permanent public record of material corrections and revisions to Convex analysis, going back to the beginning. Entries are never removed once published. A correction on an article also appears at the top of the article itself.
How We Handle Corrections
•Clarifications. Small rewordings that do not change the substance of the analysis. Noted inline on the page.
•Revisions. Updates prompted by new data or subsequent events. The prior version and the change are documented.
•Material corrections. Factual errors that change the meaning of the analysis. Logged here with the original and corrected statements.
Correction Log
Material Correction· Original 2026-04-29
Corrected the 30Y Treasury yield from 4.55% to 4.92% (FRED DGS30 April 2026 monthly average) and recalculated the mortgage-Treasury spread from 191bp to ~154bp throughout the page.
The April 2026 snapshot section originally cited the 30Y Treasury at 4.55%, which was a stale or incorrectly transcribed value. FRED DGS30 monthly readings for April 2026 averaged 4.92% (April 24 close 4.91%; April 29 close 4.97%). The dependent mortgage-to-30Y spread calculation was also off (the 6.46% Bankrate APR minus the corrected 4.92% gives ~154bp, not 191bp). All cascading numerical claims in the entry were updated, including affordability math and forward scenarios. The narrative direction (spread structurally elevated post-QT, MBS convexity at TCW-reported -3.3 most-negative) was unchanged.
Corrected the Bank of Japan policy rate from 0.5% to 0.75% across 7 occurrences. Verified against the BoJ April 28, 2026 monetary policy statement.
Several research-driven pair narratives stated the BoJ policy rate as 0.5% as of April 2026. The actual rate held at the April 28, 2026 BoJ Policy Board meeting is 0.75% — the highest level since 1995, reached via the documented hike sequence: 0.0% (March 2024 negative-rate exit) → 0.25% (July 2024) → 0.50% (January 2025) → 0.75% (October 2025). All pages now reference 0.75% with the underlying hike timeline. The dependent Fed-BoJ differential math (3.50-3.75% vs 0.75% = ~275bp) was also updated.
Corrected the Michigan 5-year inflation expectation from 4.4% (with "highest since 1981" framing) to 3.5% (highest since October 2025), per UMich Surveys of Consumers and Trading Economics.
The pair narrative stated the University of Michigan 5-year inflation expectation as 4.4% in April 2026, attributed as "highest since 1981". Both the value and the framing were wrong. The actual April 2026 final reading is 3.5% (up from 3.2% prior month), described by UMich as the highest reading since October 2025. The 1981 peak for the 5-year series was approximately 7%; recent recurring peaks have been near 3.5%. All occurrences in the entry were updated; the year-ahead reading (4.7%) was verified accurate and remains.