Kupiec test
FRM Part 2 glossary · Market Risk
Definition
Unconditional-coverage backtest comparing observed to expected VaR exceptions via a proportion-of-failures likelihood ratio.
Kupiec test in plain English
A risk model called VaR promises something like "you shouldn't lose more than $1 million on more than 1 day out of 100." The Kupiec test is a scorecard that checks whether that promise held up. You look back over, say, 250 days, count how many days losses actually blew past the limit ("exceptions"), and compare that count to how many you'd expect if the model were honest — roughly 250 × 0.01 = 2.5 days. If reality shows way more or way fewer exceptions than expected, the test statistically flags the model as untrustworthy. ("Unconditional" just means it only cares about the total number of misses, not whether they clumped together on bad days.)
Learn these first
Practice Kupiec test
In the Kupiec Likelihood Ratio test, what does the null hypothesis (H_0) state?
Related FRM Part 2 terms
FRM Part 2 practice questions · FRM Part 2 flashcards · FRM Part 2 study guide