Customer Reviews

Efficient Asset Management: A Practical Guide to Stock Portfolio Optimization and Asset Allocation

5 star:		(1)
4 star:		(5)
3 star:		(3)
2 star:		(0)
1 star:		(2)

 

 

I would agree with the comments of the first 2 reviewers. That is, the book is honest, concise and thorough in addressing the pitfalls of using Mean-Variance optimization techniques for finding optimum asset allocations (i.e., minimum risk for given expected return). However, if you don't do your own asset allocation calculations (i.e., process historical trends to find the "efficient frontier") the only value of the book is to make you aware of the issues around using Markowitz mean-variance techniques and, therefore, be questioning of any asset allocation models you come across. In other words, for the user (vs. creator) of asset allocation models be aware that if the creator wasn't careful in his statistical techniques the models could be wrong. Also, what I also got out of the book was, in many cases, rebalancing of a portfolio may not be needed as frequently as many suppose as the efficient frontier is more of a cloud then a line.

This short, simple book offers a synthesis of research about the uses and practical problems associated with Markowitz optimization procedures. It will give you a good opportunity to see in a few interesting hours what can go wrong in implementing MV optimization and what to do to improve the process. Things that are relatively obscure, but have a direct practical relevance, such as considering the efficient frontier as having a variance, and offering some pointers on where to get arcane Stein-like estimators for the variances and covariances (Ledoit estimators).

There is no math entrance barrier (almost no equations), so this book will be of benefit to users of MV optimization who want to understand the issues deeper and not just press on a button and assume that the weights they get make sense. It is to be noticed that this is not the book for those interested in quadratic programming algorithms per se, as the focus is more from a user point of view. Also notice there are no new results in the book and that sometimes I wished some discussions were more detailed - but they may be too detailed for some other readers as well.

In brief an honest book, not too dumb and not too hard. An interesting and useful reading for all users of MV optimization. Also, a perfect book to complement an undergrad education in finance.

NOTE: Although the presentation, printing and binding is similar to the infamous NYSE "technical" books or Wiley trader's advantage series, this is actually a good vulgarization book written by somebody having an academic training. No chaos, technical analysis or other arbitrary opinions are to be found here. In case you'd be scared by the look of it...

Michaud's resampling methodology is quite rigorous, although the patentability of application of econometric concepts that are over 40 years old to a theory advanced by Markowitz in 1952 should be seriously questioned by any rational reader. The applicability of resampling and improvements to the inputs to estimation are clear, and should be strongly considered by anyone in the asset management industry. The book glosses over other approaches to optimization that are not based on Markowitz, all but ignoring the huge body of literature that has been built up around other optimization approaches. Except for this shortfall, this is an excellent book, and shoulds be a part of your library on quantitative asset management.

Michaud raises several important issues that one is sure to encounter in portfolio optimization. Michaud exposes the fallibility of mean-variance optimization and suggests several techniques to obtain more reliable results. His conclusions merit consideration. Props for increasing the breadth of statistical scope of efficient asset management. Michaud is also a fluid writer. My largest complaint is that the majority of his work utilizes sign-constrained (long-only) optimization. If you manage, advise or consult on portfolio management and you utilize optimization techniques or have considered them, you should become knowledgeable with the contents of this book...

This is the old edition and the latest version is out and available. I talked with the author and there are quite a few changes in their new methodology making this book obsolete

Michaud raises several important issues that one is sure to encounter in portfolio optimization. Michaud exposes the fallibility of mean-variance optimization and suggests several techniques to obtain more reliable results. His conclusions merit consideration. Props for increasing the breadth of statistical scope of efficient asset management. Michaud is also a fluid writer. My largest complaint is that the majority of his work utilizes sign-constrained (long-only) optimization. If you manage, advise or consult on portfolio management and you utilize optimization techniques or have considered them, you should become knowledgeable with the contents of this book.

This book has a couple of good chapters on the problems of implementing the Markowitz model. Interestingly enough, I understood better the resampling idea in Scherer's article (a clean and honest review in my opinion which Michaud understandably hates) than in the entire Michaud's book. If I'm not mistaken the basic resampling idea was originally due to Jorion. Michaud patented it a variation of it. Now if you want to implement the ideas presented in this book, you are by law forbidden to do it (because it is patented). In that sense I think this book should be free (since as another reviewer below put it, it is a sales pitch.) There is an interesting review on this troubling trend of patenting really basic algorithms in Wilmott Magazine. Still it makes interesting reading. 2 1/2 stars.

Actually the author does lead a consulting firm, as possibly conjectured by the last reviewer. No wonder his lack of clarity in the how-to phase.

For interested readers, Scherer has a good explanation of how resampling actually works. He also claims, in my view rightly, that Bayesian optimization is much better ... pity that commercial software is not as readily available as in standard portfolio optimization.

Anyway, overall I agree with the judgments of earlier reviews: the book is good as a reminder of the weaknesses in standard optimization, but the solutions it proposes are suboptimal. Read Scherer instead for theory, although implementation isn't any easier and price is even worse.

I am surprised that this process has gained the clout that it has. It almost seems like the finance industry got bored of existing processes and so came up with something new and possibly promising, evn though the author himself cannot explain why hes doing what he is doing.

The book is short, but could have so been shorter had it not been for Michauds constant repetition of ideas. Save your money and google the concept if u really want to know about it.

This is an excellent book for the readers with solid quant skills. This is not a course in investing for poets. So be honest with yourself about your capabilities and needs.