According to https://en.wikipedia.org/wiki/Fisher_information#Sufficient_statistic we have if and only if, but according to https://projecteuclid.org/download/pdfview_1/euclid.imsc/1362751193 we don't. I have not yet read measure theory so I don't really understand the second link.

Could someone give me some clarity on the if part of the theorem? Is it dependent on regularity conditions?

  • 3
    $\begingroup$ I also knew it as an iff condition. See e.g. problem 6.4.15 here where the author mentions a consultation with C.R. Rao regarding the result. $\endgroup$ – StubbornAtom Nov 30 '19 at 20:55

The cited paper by David Pollard indeed analyses an example due to Kagan and Shepp [The American Statistician 59 (2005) 54–56]. That example gives a statistic which is not sufficient, but still the Fisher information based on the insufficient statistic is equal to the Fisher information based on the complete data.

I will not state that example here, but note that it is a mixture distribution where the support of the distribution varies with the unknown parameter $\theta$. That is known to cause many hickups.

So the conclusion, this is not an iff condition. But the cited paper continues to state and prove a theorem 7, which gives a necessary regularity condition for iff to hold.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.