Unbiased Estimators vs. Minimizing a Quadratic Loss Function

Unbiased estimators are important in statistics. I guess because of Cramér Rao bound, for the variance. In the sense that if  , then  https://latex.codecogs.com/gif.latex?\text{Var}[\widehat{\theta}]\geq%20I_\theta^{-1}, where   denotes Fisher information (the proof was writen in an old post).

But what could we be the variance if   is not unbiased ?

Consider the following simple case, with a Gaussian i.id. sample   from a  . We know that the estimator of the Method of Moments is the same as the Maximum Likelihood estimator, i.e.  . And this estimator is efficient, in the sense that its variance is equal to Cramér-Rao lower bound,  .

But what if we consider another estimator? For instance  , with   not necessarily equal to 1. In that case, this estimator is (usually) biased since

 

And its variance is

 

We can visualise those two functions (the bias and the variance) using


n=10
alpha=seq(0,2,by=.01)
b=1-alpha
v=alpha^2/n
plot(alpha,b,xlab="alpha",col="red",type="l")
par(new=TRUE)
plot(alpha,v,col="blue",type="l",axes=FALSE
axis(4,)
mtext("bias",side=2,line=-1,col="red")
mtext("variance",side=4,line=-1,col="blue")

Observe that if is small (smaller than 1), the variance is smaller than the Cramér-Rao lower bound. And here, the mean squared error, defined as

which is, here,

The optimal value is obtained when the first order condition is satisfied

i.e.

So biased estimators can be more interesting than unbiased estimators, if the goal is the minimize the mean square error.


5 thoughts on “Unbiased Estimators vs. Minimizing a Quadratic Loss Function”

  1. That’s an interesting presentation of the bias vs. variance trade-off. By the way, I think there is a misplaced bracket in the mse formula (2nd line): bias should be written [(\alpha – 1)\theta]^2.

Leave a Reply

Your email address will not be published. Required fields are marked *