Interpreting High Jarque–Bera Statistics in Return Diagnostics

Summary

The document raises a diagnostic question about Jarque–Bera test values reported alongside lag-one autocorrelation, a Ljung–Box statistic, skewness, and kurtosis. The author observes that the Jarque–Bera statistic rises sharply when kurtosis is high and asks whether a very large value is normal and suitable to report.

The exchange gives no explanation or resolution, so it does not establish whether the reported statistic is correctly calculated or how the sample should be interpreted. The underlying statistical point is that the Jarque–Bera test assesses departure from normality using skewness and kurtosis, and high kurtosis can contribute strongly to a large statistic. Interpretation still depends on the data, sample size, implementation, and test assumptions; the post provides no such details or evidence to diagnose a particular result.

Key ideas

• The question concerns Jarque–Bera values reported with other return-series diagnostics.
• The author notices that the statistic grows as reported kurtosis rises.
• The Jarque–Bera test uses skewness and kurtosis to assess normality.
• The document does not provide enough data or an answer to validate the reported statistic.

Tags

# Why Jarque - Bera values are so high? Is this normal?


# Why Jarque - Bera values are so high? Is this normal?












Please advise whether the following is a normal occurrence:

In the above table I have Autocorrelation at lag1, LB, Skew, Kurt and JB test.

I have noticed that whenever the value of Kurt increases, the JB increases exponentially. In some cases I even have JB=106000.35

Am using Eviews 9.5 to calculate it.

Please advise on the results of the above (as I have to report these values in the table).

Much obliged West

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.