From: Stephen Brodeur <sbrodeur@primeon.com>
Subject: Re: Asset Allocation
Date: Thu, 08 Oct 1998 19:01:39 -0400
Previous Article (by Author): Re: Refinancing Scott Barnett
Articles sorted by: [Date]
[Author]
[Subject]
I understand that asset allocation techniques can dampen the swings in portfilio values, much like diversification in general can reduce wild swings vs. concentrated holdings. But I also believe that I heard that asset allocation can actually *beat* the performance of the best asset class over the long run. I wish I could remember where I (think I) heard this claim. The rationale for such a claim is something like this: You pick an allocation scheme that you want to follow for some long period, based on risk tolerance and time frame and all that stuff. You stick with that allocation mix, and "regularly" (yearly?) adjust your holdings to keep in line with the mix. This has the effect of capturing gains in strong segments and putting them into weak (underperforming, and possibly undervalued?) segments. Superficially, this sounds like you are selling your winners to buy losers (very unappealing). But it also sounds like selling high and buying low (brilliant). I think the result is supposed to be that the overall effect of capturing some of your high performance gains and buying low into out-of-favor asset classes will give you a higher net result than taking the "highest performing class" and just riding the storm. In order for this to work, the gains that you miss by holding the sub-par performers during the boom times must be more than offset by the loss you would suffer when the high flyers crash. So, my question is, does asset allocation actually outperform individual asset classes performance over the long run, or does it just impact the risk and the fluctuations? Is it a performance vs. risk/comfort issue, or is it a performance enhancing technique? - -Steve Brodeur x222 - -
Previous Article (by Author): Re: Refinancing Scott Barnett
Articles sorted by: [Date]
[Author]
[Subject]