What do you think?
Rate this book
Active Portfolio Management: A Quantitative Approach for Producing Superior Returns and Controlling Risk
"This new edition of Active Portfolio Management continues the standard of excellence established in the first edition, with new and clear insights to help investment professionals." -William E. Jacques, Partner and Chief Investment Officer, Martingale Asset Management. " Active Portfolio Management offers investors an opportunity to better understand the balance between manager skill and portfolio risk. Both fundamental and quantitative investment managers will benefit from studying this updated edition by Grinold and Kahn." -Scott Stewart, Portfolio Manager, Fidelity Select Equity ® Discipline Co-Manager, Fidelity Freedom ® Funds. "This Second edition will not remain on the shelf, but will be continually referenced by both novice and expert. There is a substantial expansion in both depth and breadth on the original. It clearly and concisely explains all aspects of the foundations and the latest thinking in active portfolio management." -Eric N. Remole, Managing Director, Head of Global Structured Equity, Credit Suisse Asset Management. Mathematically rigorous and meticulously organized, Active Portfolio Management broke new ground when it first became available to investment managers in 1994. By outlining an innovative process to uncover raw signals of asset returns, develop them into refined forecasts, then use those forecasts to construct portfolios of exceptional return and minimal risk, i.e., portfolios that consistently beat the market, this hallmark book helped thousands of investment managers. Active Portfolio Management, Second Edition , now sets the bar even higher. Like its predecessor, this volume details how to apply economics, econometrics, and operations research to solving practical investment problems, and uncovering superior profit opportunities. It outlines an active management framework that begins with a benchmark portfolio, then defines exceptional returns as they relate to that benchmark. Beyond the comprehensive treatment of the active management process covered previously, this new edition expands to cover asset allocation, long/short investing, information horizons, and other topics relevant today. It revisits a number of discussions from the first edition, shedding new light on some of today's most pressing issues, including risk, dispersion, market impact, and performance analysis, while providing empirical evidence where appropriate. The result is an updated, comprehensive set of strategic concepts and rules of thumb for guiding the process of-and increasing the profits from-active investment management.
624 pages, Hardcover
First published November 1, 1994
72 people are currently reading
809 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 10 of 10 reviews
September 6, 2020
No student of finance would consider this book to be serious. It is shockingly a standard and unfortunately one of the only books that attempts to do what it does.
They are not academic. They quote themselves (only), they only publish in non-academic journals (J Port Mgt, Fin Analysts’ J), and just about the only thing they do is come up with entirely new labels for everything. Omega, IR, IC etc. They are also inconsistent with notation. It’s a scientist’s nightmare. What’s with that? Everything they do is to relabel something that is well known which either is named differently or nobody bothered naming. IC is just correlation. Most people call it rho. Would they? No it’s their secret sauce. IR? Why not call it a Sharpe ratio?
An example of truly retarded notation is on page 265. It is meant to be a regression of one dependent variable (returns) on one independent with an intercept. But formula 10.9 would never let you guess that.
So recall if you have (trying to use their nonstandard notation and struggling to not use linear algebra since they refuse to do so)
r = a+ phi g + eps then
phi-hat = cov(g,r)/vat(g) = std(r) corr(r,g)/std(g)
This is what they do in formula 10.9 with no explanation. Is there any insight in this? As a mathematician and a finance professional for over 20 years I would say none whatsoever. Completely vacuous.
They don’t bother telling you what they’re really doing for instance.
Just as an example I have in my book on page 269ff they list a formula for phi. (10.20). You won’t recognise this but anyone who has studied statistics should know that if you are estimating the regression Y= X beta + epsilon, an estimated beta (beta-hat) is (X’X)^{-1}X’Y. Formula 10.20 is basically that. Formula 10.21 etc is basically describing a QR decomposition or Gram-Schmidt orthogonalization. Can they tell us that? Again it would ruin their magic.
Page 311 describes a Cholesky decomposition in some truly mind numbingly dumb way. Again no mention of the actual method so you can look it up elsewhere and realise it’s no big deal.
Even the fundamental law is just relabelling a regression. Big deal. Ok they give little rules of thumb. If you have orthogonal bets. (Who does? Ever?). It is ok as a toy model sort of framework. But again to call it the Fundamental Law of Asset Mgt clearly shows the authors disproportionate self-regard.
Can they publish with serious academics or are they disregarded entirely? I would say for the most part they cannot except Kahn did write an article together with the eminent Stephen Boyd together with some 5-6 other authors. Boyd is truly amazing so i will allow him one mistake. I have a feeling they didn’t actually have to talk shop much because I’m sure hearing someone who has new names for everything is worse than a bullet in the head. It’d be like taking to a Martian. If you read the paper (which came out with cvxportfolio, a new package for applying convex optimisation specifically to multi period portfolio Mgt problems with transaction costs) it is clear that there is not a single IR, IC, etc in sight.
There are serious academics who study this area. Whether it is Stephen Boyd and his coauthors and students or Lasse H Pedersen amongst others. They unfortunately have not written textbooks which completely supplant the need to go to this horrible book.
This is basically a Neanderthal approach to portfolio Mgt. The hope is that since they are old we will all soon forget this horrible relabelling of basic regression and mean variance optimisation.
I own a copy. I sometimes look at it but need a strong stomach. It is unfortunately a standard.
They are not academic. They quote themselves (only), they only publish in non-academic journals (J Port Mgt, Fin Analysts’ J), and just about the only thing they do is come up with entirely new labels for everything. Omega, IR, IC etc. They are also inconsistent with notation. It’s a scientist’s nightmare. What’s with that? Everything they do is to relabel something that is well known which either is named differently or nobody bothered naming. IC is just correlation. Most people call it rho. Would they? No it’s their secret sauce. IR? Why not call it a Sharpe ratio?
An example of truly retarded notation is on page 265. It is meant to be a regression of one dependent variable (returns) on one independent with an intercept. But formula 10.9 would never let you guess that.
So recall if you have (trying to use their nonstandard notation and struggling to not use linear algebra since they refuse to do so)
r = a+ phi g + eps then
phi-hat = cov(g,r)/vat(g) = std(r) corr(r,g)/std(g)
This is what they do in formula 10.9 with no explanation. Is there any insight in this? As a mathematician and a finance professional for over 20 years I would say none whatsoever. Completely vacuous.
They don’t bother telling you what they’re really doing for instance.
Just as an example I have in my book on page 269ff they list a formula for phi. (10.20). You won’t recognise this but anyone who has studied statistics should know that if you are estimating the regression Y= X beta + epsilon, an estimated beta (beta-hat) is (X’X)^{-1}X’Y. Formula 10.20 is basically that. Formula 10.21 etc is basically describing a QR decomposition or Gram-Schmidt orthogonalization. Can they tell us that? Again it would ruin their magic.
Page 311 describes a Cholesky decomposition in some truly mind numbingly dumb way. Again no mention of the actual method so you can look it up elsewhere and realise it’s no big deal.
Even the fundamental law is just relabelling a regression. Big deal. Ok they give little rules of thumb. If you have orthogonal bets. (Who does? Ever?). It is ok as a toy model sort of framework. But again to call it the Fundamental Law of Asset Mgt clearly shows the authors disproportionate self-regard.
Can they publish with serious academics or are they disregarded entirely? I would say for the most part they cannot except Kahn did write an article together with the eminent Stephen Boyd together with some 5-6 other authors. Boyd is truly amazing so i will allow him one mistake. I have a feeling they didn’t actually have to talk shop much because I’m sure hearing someone who has new names for everything is worse than a bullet in the head. It’d be like taking to a Martian. If you read the paper (which came out with cvxportfolio, a new package for applying convex optimisation specifically to multi period portfolio Mgt problems with transaction costs) it is clear that there is not a single IR, IC, etc in sight.
There are serious academics who study this area. Whether it is Stephen Boyd and his coauthors and students or Lasse H Pedersen amongst others. They unfortunately have not written textbooks which completely supplant the need to go to this horrible book.
This is basically a Neanderthal approach to portfolio Mgt. The hope is that since they are old we will all soon forget this horrible relabelling of basic regression and mean variance optimisation.
I own a copy. I sometimes look at it but need a strong stomach. It is unfortunately a standard.
January 25, 2016
I read this book because it was recommended for Coursera course: Computational Investment I.
It was my first book on Portfolio Management, although it has very good ratings on goodreads and amazon, I surprisingly found this book rather obscure and not-easy-to-follow. The book tries to do a mathematical approach to portfolio management, but mathematical formulas come out of the blue, with no previous explanation or justification. The level of math required is not a big deal, it is just that formulas are completely unexplained. Since the book is so well rated by many other readers, I guess this is not an introductory course for first-timers.
It was my first book on Portfolio Management, although it has very good ratings on goodreads and amazon, I surprisingly found this book rather obscure and not-easy-to-follow. The book tries to do a mathematical approach to portfolio management, but mathematical formulas come out of the blue, with no previous explanation or justification. The level of math required is not a big deal, it is just that formulas are completely unexplained. Since the book is so well rated by many other readers, I guess this is not an introductory course for first-timers.
December 7, 2016
If you're an investing professional, you should already know about this book, whether you use it or not. It's a highly quantitative read that will make your undergraduate math courses valuable, literally. It may not make you rich, and it may not make the people you invest for rich, but you will at least understand why or why not after understanding the math. If you're interested in how indexes (benchmarks) are constructed for specific purposes, this is the book. If you're interested in serious measures of investment skill and performance, ditto. Do be prepared for a lot of linear algebra and calculus and probability theory, though.
March 30, 2016
A reference classic, and surprisingly well written. I refer to it weekly at work.
August 22, 2018
Academic financial text books have, to a large extent, focused on beta and the so called efficient market. Active Portfolio Management was groundbreaking when it was first published in 1994 as instead it was devoted to the practical process of generating alpha from a quantative angle. Richard Grinold and Ronald Kahn, today retired and at BlackRock respectively, share a history in academia, at BARRA and above all at the quant behemoth Barclays Global Investors where they both held leading positions while writing this book.
Even though the book is full of financial theory the approach is practical. The topic at hand is the generation of risk adjusted relative returns. The market returns are always the baseline and success is measured by the IR (the ratio of residual return to residual variance) rather than an academic Sharpe ratio. When I first read this book 10 – 12 years ago I didn’t by any means find it enjoyable. It’s thick, theoretical, filled with formulas and I was frankly not ready for it. Yet, over the years I find myself returning to the key concepts of the book over and over again. Out of the four parts the first lays out the authors’ theories and then the latter three cover the practical work of a quant PM.
The claim to fame of the book is a concept called The Fundamental Law of Active Management that reads IR=IC*√BR. It states that there are two sources of oportunities to increase the information ratio. The first is the ability to forecast asset’s residual return, measured by the information coefficient. The IC is the connection between forecasts and eventual returns, IC=2*(N1/N)-1 where N1 denotes the number of correct bets and N the total number of bets made. It’s a measure of skill. Active asset management is all about forecasting. Success in forecasting doesn’t only hinge on doing things right but also on doing the right thing, i.e. not only on the PMs skill but on wheather the pond he is fishing in is promissing. This is where BGIs quant signals come into play but also Warren Buffett’s concept of Graham and Doddsville. The second source of IR is breadth – the number of independent active oportunities per year the PM have to use his skill on. IC is about the quality of investment opportunities while BR is about the quantity of investment opportunities through coverage of more securities or a higher frequency of opportunities. It is for example more valuable to be able to forecast the returns of 1600 stocks than 1100 stocks. To increase the IR from 0,5 to 1,0 one would need to double skill, increase breadth by a factor of four or some combination of the two. This additive value of further breadth requires investment opportunities to be totally uncorrelated (this favours an eclectic investment style). If a new opportunity is fully correlated to a previous one it adds no IR. Most opportunities fall somewhere in between.
Often the asset management process focuses excessively on the quality of bets versus the quantity. The concept of breadth emphasizes the negative secondary effects that come with placing limits on an investment process with an edge. It’s not only that limiting the investment universe deducts IR, the implications are broader. Limiting yourself to being long only lowers IR. Placing restrictions on the amount of cash in the portfolio lowers IR, so does demanding sector neutrality in an equity portfolio etc. In reality you only need to have a very small edge, say a 0,52 hit ratio to create a great IR if the breadth is large – so don’t constrain yourself without good cause.
The opening line of the book reads “The art of investing is evolving into the science of investing.” In many ways the book was a precursor to today’s world of multifactor risk factor models, exotic betas etc., i.e. the systematization of all the alpha signals that can be systemized - leaving precious little pure alpha for traditional active investors. In an interview Grinold stated “The goal is to replace heroic personalities contending in an atmosphere of greed and fear with compelling hypotheses subjected to hard data.” This is an important book, but I kind of like my heroes.
Even though the book is full of financial theory the approach is practical. The topic at hand is the generation of risk adjusted relative returns. The market returns are always the baseline and success is measured by the IR (the ratio of residual return to residual variance) rather than an academic Sharpe ratio. When I first read this book 10 – 12 years ago I didn’t by any means find it enjoyable. It’s thick, theoretical, filled with formulas and I was frankly not ready for it. Yet, over the years I find myself returning to the key concepts of the book over and over again. Out of the four parts the first lays out the authors’ theories and then the latter three cover the practical work of a quant PM.
The claim to fame of the book is a concept called The Fundamental Law of Active Management that reads IR=IC*√BR. It states that there are two sources of oportunities to increase the information ratio. The first is the ability to forecast asset’s residual return, measured by the information coefficient. The IC is the connection between forecasts and eventual returns, IC=2*(N1/N)-1 where N1 denotes the number of correct bets and N the total number of bets made. It’s a measure of skill. Active asset management is all about forecasting. Success in forecasting doesn’t only hinge on doing things right but also on doing the right thing, i.e. not only on the PMs skill but on wheather the pond he is fishing in is promissing. This is where BGIs quant signals come into play but also Warren Buffett’s concept of Graham and Doddsville. The second source of IR is breadth – the number of independent active oportunities per year the PM have to use his skill on. IC is about the quality of investment opportunities while BR is about the quantity of investment opportunities through coverage of more securities or a higher frequency of opportunities. It is for example more valuable to be able to forecast the returns of 1600 stocks than 1100 stocks. To increase the IR from 0,5 to 1,0 one would need to double skill, increase breadth by a factor of four or some combination of the two. This additive value of further breadth requires investment opportunities to be totally uncorrelated (this favours an eclectic investment style). If a new opportunity is fully correlated to a previous one it adds no IR. Most opportunities fall somewhere in between.
Often the asset management process focuses excessively on the quality of bets versus the quantity. The concept of breadth emphasizes the negative secondary effects that come with placing limits on an investment process with an edge. It’s not only that limiting the investment universe deducts IR, the implications are broader. Limiting yourself to being long only lowers IR. Placing restrictions on the amount of cash in the portfolio lowers IR, so does demanding sector neutrality in an equity portfolio etc. In reality you only need to have a very small edge, say a 0,52 hit ratio to create a great IR if the breadth is large – so don’t constrain yourself without good cause.
The opening line of the book reads “The art of investing is evolving into the science of investing.” In many ways the book was a precursor to today’s world of multifactor risk factor models, exotic betas etc., i.e. the systematization of all the alpha signals that can be systemized - leaving precious little pure alpha for traditional active investors. In an interview Grinold stated “The goal is to replace heroic personalities contending in an atmosphere of greed and fear with compelling hypotheses subjected to hard data.” This is an important book, but I kind of like my heroes.
August 13, 2023
Was highly interesting and gratifying to trace and confirm the links and 'lessons via experience' of discretionary investing, from the mathematical proofs of quant and systematic management. CAPM posits that the expected residual return is zero, hence the optimal portfolio will only differ from the market portfolio if forecast excess returns differ from CAPM consensus returns. Benchmark (market) timing is sus, and breadth and skill compose the information ratio. Nice crumb trail that strengthens the story from the fundamentalist camp, that active managers NEED a differentiated view. I'd just be highly interested to know some in the fundamental camp's (Howard Marks namely) pushback to the authors' doctrine of 'active management is forecasting'. And superior returns boil down to the search for superior information. WHAT that superior information actually entails is beyond me - is it a shrewder interpretation of public info than active managers? Or is it digging deeper to uncover more info and building a bigger mosaic?
This book is also starting to show its age. APT is just a typical, beaten-to-death multi-factor model. Elementary discussions of behavioural finance were amusing. Can the IR even be seen as a budget constraint? - (information is not static.) Plus, was Fama and French's objective of the 3-factor model even to hard disprove the CAPM? And does the 3-factor model even overthrow the CAPM? In my view, it was just a natural extension of the too-good-to-be-true simple linear regression of the CAPM. Not different from how Sharpe extended Markowitz's work.
Most of this shit just flew over my head tbh, but thankfully for math noobs like me, many of the core ideas are contained in the chapters. More intensive derivations are reserved for the technical appendices. The core texts are fairly accessible, and I think economics-level math, which isn't a high hurdle for the intended audience (people who are even considering trying their hand at institutional active management), can get one through this book.
This book is also starting to show its age. APT is just a typical, beaten-to-death multi-factor model. Elementary discussions of behavioural finance were amusing. Can the IR even be seen as a budget constraint? - (information is not static.) Plus, was Fama and French's objective of the 3-factor model even to hard disprove the CAPM? And does the 3-factor model even overthrow the CAPM? In my view, it was just a natural extension of the too-good-to-be-true simple linear regression of the CAPM. Not different from how Sharpe extended Markowitz's work.
Most of this shit just flew over my head tbh, but thankfully for math noobs like me, many of the core ideas are contained in the chapters. More intensive derivations are reserved for the technical appendices. The core texts are fairly accessible, and I think economics-level math, which isn't a high hurdle for the intended audience (people who are even considering trying their hand at institutional active management), can get one through this book.
February 6, 2020
A classic book that defined the quant portfolio management & factor investing approach. I had to read more than 10 times when I was working at Barra.
One negative point: While the second edition was better organized, sometimes the book seems more as collection of papers.
=> A must have on a any quant's shelf.
One negative point: While the second edition was better organized, sometimes the book seems more as collection of papers.
=> A must have on a any quant's shelf.
July 29, 2020
A good and near-comprehensive resource on quantitative theory and its applications. It approaches active investing from a “humble” point of view (their words) and ensures a deeper understanding of the risk involved and how it may be mitigated.
April 15, 2016
This is the textbook for the active portfolio management course at Haas School of Business taught by Dr. Ronald Kahn, the author of the book. It is more like an encyclopedia, and not an easy reading for business people: it is definitely not rhetoric. However it does provide everything you need to know to construct, backtest, and evaluate your portfolio. I would keep it on my shelf for future reference. Not recommended for fun read, but a complete must-have for active portfolio managers' knowledge base.
December 1, 2014
This book communicates the background of investment extremely well. Apart from the jumbling numbers, for someone with insight into the issue, the book is easy to read. I learned mostly about the underside of the investment process, a issue that I think not communicated enough today. I would recommend this book to anyone who is curious about scientific and investment theories.
Displaying 1 - 10 of 10 reviews