Noise-filtering techniques improve accuracy of key economic and financial tests.
When time series have noise, common tests like ADF and Johansen's test can give misleading results. This study explains why and suggests three filtering methods to improve the tests' accuracy without affecting their reliability. By applying these filters, the tests performed better on noisy data in simulations and real-world examples involving electricity prices and stock trading strategies.
