Direct answer: what evidence suggests rational expectations does not hold in forex?
Rational expectations is the idea that people (and, by extension, markets) use information available at the time to form forecasts that are, on average, correct. Evidence that it does not hold in forex typically shows up as systematic forecast error or predictable patterns in how exchange rates respond to information.
In the inflation-expectations context, the most relevant evidence is whether currency moves and inflation-expectation updates are consistent with a model where expectations are formed efficiently from the same information set.
How this shows up in practice (mechanics)
In a rational-expectations framework, one expects that forecast errors are not reliably forecastable using information that was already known when the forecast was made.
Concrete, verifiable signs include:
-
Predictability in forecast errors If you can sort historical periods using information available at the forecast date (for example, variables related to inflation expectations), and you find that later exchange-rate outcomes are systematically higher or lower, that indicates the forecasts were not fully efficient.
-
Systematic bias Bias means the average difference between realized outcomes and the forecasts is not centered around zero across samples.
-
Delayed or incomplete adjustment to new information If inflation-expectation information arrives and exchange-rate responses occur in a staggered way (rather than immediately on average), it suggests expectations may update more slowly than the rational-expectations assumption implies.
-
Model dependence and missing variables If adding reasonable information improves forecast performance in a way that is hard to justify under the original rational-expectations setup, that suggests the initial information set or the forecasting structure was insufficient.
Example checks (what to measure, and what would count as evidence)
You can frame tests around the inflation-expectations channel without assuming any specific forecasting model:
- Compare expected vs. realized inflation expectations: If measures of inflation expectations consistently miss later inflation outcomes in a way that could have been anticipated, then expectations formation is not behaving as “average-correct”.
- Link inflation-expectation updates to currency changes: If currency movements systematically track inflation-expectation changes in a way that can be predicted from pre-forecast variables, that is consistent with non-efficient expectation updating.
- Test residuals for patterns: A rational-expectations benchmark implies no exploitable structure in forecast residuals (within the limits of the chosen model). If residuals remain predictable, the rational-expectations assumption is strained.
Limitations and uncertainty
- No single test “proves” a violation: Different models, measurement choices, and sample periods can produce different results.
- Inflation expectations are measured imperfectly: Any comparison depends on how inflation expectations are proxied (for example, survey-based vs. market-implied measures), and measurement error can look like irrationality.
- Structural breaks can mimic irrationality: Regime changes, large shocks, or changes in risk/hedging demand can create persistent patterns that are not purely about expectations being “wrong.”
- Rational expectations is a conditional statement: It is about forecasts formed with a given information set and model assumptions; evidence against it is evidence against that specific conditional structure, not necessarily against expectations in general.
If you keep the claim bounded to what can be measured—forecast error behavior, bias, residual predictability, and adjustment timing—you can evaluate whether rational expectations is consistent with the inflation-expectations evidence, while acknowledging uncertainty and model dependence.