The failure of financial econometrics: Confirmation and publication biases

Confirmation and publication biases are demonstrated by using an example in which the extent of exchange rate misalignment (as applied to the yuan/dollar exchange rate) is measured by using six different models. Testing these models initially reveals that they form cointegrating vectors according to one test but not the other two, thus creating the potential for confirmation bias. A wide range of estimates of misalignment is produced, allowing for both confirmation and publication bias. It is argued that financial econometrics is the tool that enables these biases in finance research.