I'm finding some other comments that are also filling in various gaps.
For example:
That second message gets to the heart of the matter: It's loads of work to figure out the Active Share and Tracking Error of each funds. The question is whether there is a way to do or buy that work without it "costing" more than it offers.
For example:
toulamapS@fatwallet said:I read the paper cover-to-cover, with some parts more carefuly, and others less carefully.
Some thoughts and questions:
1. I think the results make sense, and do not contradict existing literature that active management generally underperforms passive indexing,
2. So among the 4 types of active funds: (a) diversified stock pickers, (b) concentrated stock pickers, (c) factor bettors / timers, and (d) closet indexers, only (a) is shown to have alpha relative to benchmarks. However, there are many caveats, such as the consistency of the diversified stock pickers, and the temporal opportunity of stock picking,
3. I believe there is no reason to expect that someone has refuted this paper. Several reasons: (i) author is building on similar research, (ii) analysis seems thorough, (iii) it is too new to be refuted any way.
4. Do I believe the results? I am tempted to. Market efficiency states that whenever there is alpha, it will be arbitraged away. Even picking better mutual funds is a source of alpha, so this paper is identifying how to consistently pick funds expected to perform better. However, consider the following: (i) the results are relative new, (ii) the alpha is relatively weak, (iii) the "Active Share" is not easily accessible for funds (while the tracking error is VERY EASILY accessible).
However, I believe that "Active Share" can be approximately computed for many / most funds without much effort.
At any rate, good paper, and thanks for sharing.
----
a good exercise for you, IMHO, would be to figure out the Active Share and Tracking Error of each of those funds listed in the OP. Similar to that spreadsheet you have posted for the non-retirement account funds, you can try to add the Active Share of each fund.
Because Acive Share is not readily available, this would require some work on your part.
Also, obtaining the Tracking Error, although easy, might not be trivial. I suggest you obtain the tracking error relative to the closest benchmark. E.g, for a fund that claims to be "Mid-cap value", you might want to determine what Mid-cap index they are following, and then determine the Tracking Error like that. Obtaining Tracking Error vs the SP500 might not be that precise for the purpose of Petajisto paper.
That second message gets to the heart of the matter: It's loads of work to figure out the Active Share and Tracking Error of each funds. The question is whether there is a way to do or buy that work without it "costing" more than it offers.
Last edited: