Shrinkage Anyone?

Which of the following concerning Shrinkage Estimators is FALSE? A) the term ‘shrinkage’ refers to the approach’s ability to reduce the impact of extreme values in historical estimates B) any choice for the target covariance matrix will lead to an increase (or at least not a decrease) in the efficiency of the covariance estimates versus the historical estimate C) shrinkage estimator is an inferior approach for estimating the population covariance matrix for the medium- and smaller-size datasets that are typical in finance

I was in the pool! B or C… due to the word “any” i will go with B.

i’ll take C.

I said C. Though I have no idea.

and just for giggles since i’m now looking at my notecards and it’s my next one after shrinkage, remember with volatility clustering, high volatility periods are often followed by periods of more high vol or low with low. the higher the rate of decay b/t volatility b/t one period and the next, the greater the tendency for volatility clustering. i like these quizes- reminding me of the little stuff i completely forgot even existed in this cirriculum.

C final answer and score 1 for Skipp’s seinfeld reference

word. i am banking on the graders having a sense of humor and crediting me for quips in the exam

these eggs are delicious… c cuz shrinkage is a superior way for esitmationg not inferior

C is the correct answer…superior is the right word, not inferior as pimp stated “it shrinks?”…“like a frightened turtle!”

c… is the right answer…it is not an inferior way doing things

Bouncing between B and C. “A surprising fact concerning the shrinkage estimator approach is that any choice for the target covariance matrix will lead to an increase…” from book. Thanks, mcap. Good Q, …copied to my googdocs.

Has to be C. B is literally a direct quote from the book. I read that chapter again on Thursday. It also talks about using a model that will produce a result of between 0 and 1 between the historical covariance matrix and the shrinkage covariance matrix, so even if the target matrix was terrible the model would produce a weight of 1 to the historical. Although it went on to say how the model worked was outside the scope of the reading.


C shrinkage fits medium and small data like financial market