The Correction Freeze Problem Survives Independent Severe Tests: Route 4 on Inflation Expectations and Climate Belief
Description
The Correction Freeze Problem (CFP) asks why beliefs often fail to correct when the world changes slowly. Its fourth route, the information limit, says that a slow drift can be invisible because it carries too little information to be told apart from noise. This paper puts Route 4 through two independent tests, one on inflation expectations and one on climate belief. The predictions came from CFP itself and were fixed before any model was estimated.
The inflation test uses official monthly data for 2000 to 2026. The sources are the University of Michigan Surveys of Consumers, the Bureau of Labor Statistics, and the Energy Information Administration. Route 4 predicts that people notice large, salient price changes but miss the slow monthly trend. The data show this pattern. Gasoline moved by 10% or more in 42 months. In 33 of those months when expectations changed at all, they moved in the same direction as gasoline 94% of the time, against 62% in other months. Relative to the monthly inflation trend, expectations moved in the matching direction only 51% of the time, which is chance.
The climate test links 37,417 respondents of the Climate Change in the American Mind survey (2008 to 2025) to NOAA temperatures for their region. It separates long-term regional warming from the weather of the most recent year. Belief that global warming is happening did not rise with long-term warming, as Route 4 predicts, and the pre-registered falsifier was not triggered.
The noise that people face was also measured directly, from everyday item prices and regional temperature records. In both domains, observed accuracy stays below the ceiling that the available information allows. Two results are weaker than predicted. The monthly inflation trend has a small but real effect on the level of expectations, and recent weather moves climate belief in the predicted direction but not significantly.
Route 4 was exposed to failure on independent data and did not fail. This is survival, not final confirmation, and the decisive next tests are set out. The other three routes of CFP remain untested, because the data needed to test them do not yet exist.