William Sharpe: Institutional Asset Allocation Studies “Inferior”

20 December 2006

Alpha Male has just finished reading William Sharpe’s new book, “Investors and Markets“.  While it’s not an easy read (think second year university microeconomics), Sharpe is always relatively accessible when compared to others of his pedigree.  The book raises a number of issues that are relevant to this blog.  Over the next few weeks, we will attempt to do the book justice by touching on these various intersections. 

For those of you who don’t have a Nobel Prize or who have a Nobel Prize in an discipline other than Economics, don’t be turned off by the title of this book.  Despite trying its best to scare people away with one of the driest titles in the history of publishing, this book is worth reading.  

Essentially, Sharpe’s “CAPM 2.0″ accounts for the heterogeneity of investor preferences, positions, and risk tolerances by building a CAPM-world from the ground up using a discrete number of investors.  By doing this, Sharpe avoids having to assume everyone operates the same as the “representative investor” in a CAPM framework and he offers financial advisors and consultants an academic foundation upon which to ply their trade (matching investments to the panoply of unique investor profiles). 

Sharpe has developed a simulator in which hypothetical investors can be created and instructed to trade with each other.  Sharpe’s virtual trading world (available for free at Sharpe’s homepage) lacks the pizzazz of the much ballyhooed and somewhat freaky “Second Life” virtual world, but it operates in much the same way.  For example, Sharpe creates two characters Mario (a fisherman from Monterey, California) and Hue (a fisherman from Half Moon Bay, California).  He assigns them various risk preferences and tells them to trade with each other until both are as happy as possible.

Then, like a master storyteller, Sharpe introduces a variety of other characters into the plot in order to illustrate the effect of changing investor preferences.  Marie, Mario’s ne’er-do-well richer sister and Hugo, Hue’s bon vivant brother enter the plot during Act 3 to illustrate how wealth impacts investing decisions.  Other characters include: David & Danielle, who both have “decreasing relative risk aversion” (as a result of a tragic car accident on the way to a skiing holiday in Tahoe we assume), Kevin & Warren, who have “kinked marginal utility curves” (Sharpe does not address their orientations, leaving readers guessing), and Quade & Dagmar, pool boys by day and shrewd index investors by night (unbeknownst to both of their Republican parents).  Yes, Sharpe spins quite a yarn as he builds a CAPM of discrete, rather than continuous distributions.  (Okay, we embellish just a little).   

But he departs from this pedagogical framework in the final chapters to discuss its significance for advisors and institutional consultants.  In Chapter 8, he covers the “Asset/Liability” studies regularly undertaken by institutional investors around the world.  His conclusion: these studies lead to inferior portfolio choices because they involve two separate steps: selecting a target policy mix and filing the resulting asset class buckets with various active funds.  Says Sharpe:

“Institutional investors engage in asset allocation studies or, if liabilities are taken into account, asset/liability studies.  Such studies normally involve members of an investment board, which selects a target policy mix (asset allocation) and establishes allowable ranges around target values.  Between studies, staff members are directed to ensure that the fund’s actual asset allocation stays within the pre-specified ranges.  Since such studies are time-consuming and expensive, they are performed infrequently, typically at intervals of one to three years….

“While simple to execute, such stepwise procedures are likely to be inferior to a more integrated approach to the problem.  In a typical two-stage approach, the preferred asset allocation is selected on the assumption that all funds will be invested in passive and costless index funds.  But actual investment vehicles typically have costs, added risks, and more complex relationships with the underlying asset classes.  The result is likely to be inferior portfolio choices and overly optimistic forecasts of future performance.

“A far more rational approach uses only one stage, dealing directly with the actual investment vehicles, with all their attractive and unattractive features.  This does not imply that asset classes should not play a role.  Quite the contrary, As we have seen, asset classes can serve well as risk factors.  Whether or not their returns are assumed to differ from those implied by their beta values depends on one’s view of market equilibrium.  But approaching the problem from two separate stages can only lead to inferior results.”

Not only can active funds can have “complex relationships with the underlying asset classes”, but we would add that asset classes themselves tend to have a high correlation, as this timely piece of Morningstar junk mail I received today illustrates.  

We believe asset allocation is a blunt instrument that brands active managers as being part of an asset class bucket mainly for ease of understanding and communication to investment boards.  Instead of doing the “two-step”, why not actually follow-through with the assumption that the chosen funds in a portfolio be “passive costless index fund”?  Then, depending on the mandate, buy active management separately in the form of (high alpha content) hedge funds.  For starters, the benefits of active management (if any) could be better measured and rewarded.

Investors & Markets By William F. Sharpe     

 Email this post to a friend   Print This Post 

3 Responses to “William Sharpe: Institutional Asset Allocation Studies “Inferior””

1. All About Alpha » Blog Archives » Net Inflows and Time-Varying Alphas: The Case of Hedge Funds Says:
   January 9th, 2007 at 1:00 pm

   […] But how can this be?   Intuition suggests that more assets chasing the same inefficiency will eventually arbitrage- (”iron”) out that inefficiency.  As Alexander Ineichen says in his new book, the market “learns” or “becomes immune” to the arbitrageur (although he does say that markets cannot be perfectly efficient).  But according to Beltratti and Morana, market participants have heterogeneous utility functions (a notion also argued by Max Darnell, Joanne Hill, and William Sharpe): “The wide variety of real world investors, including noise traders and investors with heterogeneous time horizons and objectives, seems to provide plenty of opportunities for hedge funds managers to exploit: the limits of arbitrage do not seem to have been met yet.” […]

2. All About Alpha » Blog Archives » The Whole Enchilada Says:
   January 25th, 2007 at 8:26 pm

   […] And finally, he suggests (as does William Sharpe in his new book) that strategic asset allocation is a false religion… “Rather than “equity managers” competing with other “equity managers” in the investor’s equity piece of the pie, all alpha generators will compete with each other for the whole enchilada.” […]

3. allaboutalpha.com: Welcome to AllAboutAlpha.com Says:
   May 9th, 2007 at 6:27 pm

   […] The report suggests that Modern Portfolio Theory’s need to divide the world into asset classes is behind much of this “semantic confusion”.  In an apparent affirmation of William Sharpe’s contention that asset allocation studies are “inferior” (see related posting), over half of participants said that the very concept of a “policy portfolio” needed to be either revised or reinvented… […]