NEUTAUR/Data Glossary/Mann-Kendall Trend Test

Mann-Kendall Trend Test

The Mann-Kendall test is a standard statistical method for deciding whether a time series is rising, falling, or has no monotonic trend.

The Mann-Kendall test is a non-parametric statistical test that evaluates whether the values in a time series tend to increase or decrease over time. Because it is rank-based, it does not assume the data follow a particular distribution, which makes it robust for noisy economic and physical series.

How NEUTAUR uses it

NEUTAUR applies the Mann-Kendall test to physical and trade series to describe direction in plain language — rising, falling, or flat — and pairs the long-run trend with the most recent window so a series that is turning shows both readings side by side. The output is a description of what the data did, not a projection of what it will do.

Data source

Mann (1945), Econometrica 13:245-259 · Kendall, Rank Correlation Methods