Applying the law of averages to expected returns

First time posting something on Reddit, so let me know if i fucked it up.

A while back i was watching a TV show (i think it was brain games) where they had a group of people estimate the number of marbles in a jar. Turns out that, while individual estimates were all over the place, the group's average estimate ended up being shockingly close to the true value, which got me thinking:

In Markowitz Optimization Theory, the optimal construction of the risky portfolio depends on 3 things:

1. The expected return on a stock;
2. The standard deviation of that stock; and
3. The correlations (more specifically, the covariances) of the returns between every possible pair of stocks in the risky portfolio.

Since the 3rd component is a matter of using historical returns to calculate the covariance of each stock to another, it should be uniform across investors. The other two components, however, depend heavily on the individual investor's opinion on future expectations (regardless of whether you're using CAPM, APT, or another method to estimate them), which is the reason why the composition of the risky portfolio varies among different investors.

So, my questions are:

Could you apply the same concept we saw with estimating the number of marbles in a jar to estimating the expected return and standard deviation by averaging as many analyst expectations on these two variables as possible?

Would this average value reflect the most likely outcome?

Does anyone know of any empirical studies relating to this?

(Disclaimer: Although I'm currently working on my Masters in Finance, I'm really only in the process of fully understanding these concepts, so please correct me if my thinking is flawed. I'm here to learn above anything else)

Submitted June 21, 2018 at 08:00AM by holistic_fkp https://ift.tt/2I7azeN