Tuesday, March 6, 2012

Designing Portfolios-How to Allocate Among Asset Classes or Models

In the old days deciding how to allocate a portfolio came down to a decision on stocks vs. bonds.  With all the new ETFs and other investments out there it is becoming more and more difficult.  Now investors have access to commodities, currencies, alternative strategies, etc.  There a number of ways to decide how to allocate among these different areas, including but not limited to:

1.      Traditional Modern Portfolio Theory (MPT)—Here an investor uses some sort of optimizer that either has past returns, correlations, and volatility of a number of asset classes or some prediction of future ranges for these variables.  The optimizer then creates the “optimal” portfolio mix.  The drawbacks of this approach are apparent to anyone who suffered through 2002 and 2008 as asset classes that weren’t that correlated or volatile, became correlated and volatile.  MPT works in an up market (along with everything else) but suffers horribly in a down market.

2.      Core and Satellite- This approach typically has a fixed core of around 80% that is in an index fund or a mix of index funds to get market returns.  It will then have around 20% that will be used to be opportunistic or defensive.  If the investor has a solid strategy for the satellite portion then this can be better than MPT but the large index exposure will still suffer horribly in a down market.

3.      Balanced portfolio—This is the simplistic 60% stocks/40% bonds approach.  The idea is that stocks and bonds are uncorrelated so the bonds can cushion the portfolio in a down market.  While bonds can provide some cushion the inherent problem with this design is that stocks are much more volatile than bonds.  So bonds may go up in a down market but stocks will go down much more than bonds will go up.

4.      Risk Parity---This is an approach that is gaining a lot of favor in the institutional marketplace because it typically needs to use leverage.  In a risk parity approach, asset classes, or in our case tactical models, are weighted by risk so that each asset class or model contributes equally to portfolio risk.  When done using uncorrelated assets or tactical models this can significantly smooth out returns.

5.      Equity Curve Feedback—In this approach you monitor the returns of each asset class, or in our case each tactical model, money is then taken away from the asset classes or models doing the worst and moved towards the ones doing the best.  This is in contrast to the rebalancing done with MPT portfolios where you sell your winners to buy more of your losers.

6.      Leveraged Space---This is something new we are working on so expect to see more in the future.  Basically this is the Optimal F approach that specifies how much should be allocated to one asset class or strategy taken to the portfolio level.  So far this looks like a very promising way to reduce risk while still generating attractive returns.

No comments:

Post a Comment