Updated: 5/7/2004; 1:21:57 PM.
Tahsin Alam's Radio Weblog
        

Thursday, April 29, 2004

[Finance]

Kritzman, M. and S. Page, “Asset Allocation versus Security Selection: Evidence from Global Markets,” The Journal of Asset Management, Spring 2003.

  • The "received wisdom" that asset allocation is more important than security selection is not conclusive
  • Importance of a decision = the difference that it makes = dispersion caused by the decision
  • Importance of making an investment decision = the dispersion of possible investment outcomes that can result from the different options that one can decide between
  • NOTE - Which decision we should focus on depends not only on the importance of the decision but also on our skill in making the correct decision. We are separating the two issues here. But realize that answering the "importance" question does NOT answer the question about where we should spend our time.
  • Relative Volatility 
    • Take two stocks A1 and A2. If they have high "relative volatility", then they are very volatile w.r.t. each other. If their respective individual volatilities are given, then they have highest relative volatility at -1 correlation, and lowest at +1 correlation. Makes intuitive sense.
    • Higher correlation -> higher rel vol
      Higher std dev -> higher rel vol
    • The higher the relative volatility between A1 and A2, the more "important" is the decision to choose between them.
    • Given four stocks A1, A2, B1, B2 belonging to two asset classes A, B,
      importance of asset allocation decision = rel vol of A vs. B and
      importance of security selection decision = rel vol of X1 vs. X2 (X=A,B)
    • Higher stock correlation -> lower stock rel vol
      BUT higher stock correlation -> lower asset class std dev -> lower asset class rel vol
    • Breakeven is at 50% intra-asset class correlation, and 33% inter-asset class correlation.
  • Monte-carlo simulation of multiple asset classes
    • Take a 60-30-10 stock-bond-cash as the baseline allocation
    • Simulate security selection within the stock portion - pick 100 stocks at a time every year, and use bootstrapping to plot out a return distribution for 14 years
    • Take a 60-30-10 stock-bond-cash as the baseline allocation
    • Randomly vary allocation "normally" around  this given baseline every year using "bootstrapping" and plot out a 14-year return distribution
    • The security selection decision shows a bigger dispersion than the asset allocation decision -> security selection is more "important" than asset allocation
  • Conclusion:
    • #1: Security selection is important -> if you are not good at it, DON'T DO IT! To quote "unskilful but nonetheless perceptive investors choose to avoid it"
    • #2: Mutual funds out there don't show nearly as much dispersion as is achievable through security selection -> "the authors suspect that managers compress the natural opportunity set to reduce their exposure to business risk"
  • My thoughts:
    • The asset allocation decision simulation is "tactical asset allocation around a baseline" - it's not strategic long-term asset allocation, and it does not have wide-swings. That's ok since most of the time what people mean by tactical asset allocation is exactly what the simulation does.

      However, just wanted to point out that saying "X-Y-Z" is a good asset allocation strategy for the next 30 years for this investor is an extremely important decision. I bet that choosing 100-0-0 vs. 0-0-100 will show a much higher dispersion than is possible via security selection. So, strategic asset allocation (setting the baseline) is an extremely important decision.

      Now the other part is that the simulation makes incremental difference from the baseline decision, not wide swings like "go 100% cash now". If we allowed the asset allocation decision to be defined as deciding between wide swings like that, then I bet it will again bcome more "important". But people never do this, right? People do "tactical" asset allocation? But isn't that the same observation as made at the end of the paper that managers doing security selection don't take advantage of the full dispersion possible in the activity?
    • Don't forget that there are two different factors involved in deciding where to allocate one's decision-making resources - "importance" of the decision, and "skill" at making the decision.

      It's also possible that different decisions allow different levels of skill. I mean that there are factors inherent about a decision that allow certain ranges of skills to exist out there. This is unformed thinking, but I am trying to capture the idea that it may be "easier" to have skill at making certain decisions than others. You don't need to be a Buffett to know that there is a high probability of interest rates going up in the U.S. over the next 12 months, but maybe you need to be a Buffet to decide which security to buy. And nature allows fewer Buffetts to exist.



11:43:14 PM    comment []

© Copyright 2004 Tahsin Alam.
 
April 2004
Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30  
Mar   May


Click here to visit the Radio UserLand website.

Subscribe to "Tahsin Alam's Radio Weblog" in Radio UserLand.

Click to see the XML version of this web page.

Click here to send an email to the editor of this weblog.