Thomas Idzorek is the Global CIO of the Institute of Morningstar. His presentation is The Impact of Skewness and Fat Tails on the Asset Allocation Decision (see here for the paper that formed the basis for the presentation).
My session notes follow. As always, these are at-the-time notes. I make no guaranty as to their accuracy or completeness.
- The standard return model (bell curve) underestimates fat tails (following Taleb).
- The “flaw of averages” — average returns over time mean less due to potential extreme outcomes and any category can underperform, even by a lot, over long periods.
- Average returns focus can ignore risks.
- Markowitz mean-variance optimization has issues (e.g., the model ignore liabilities — the reasons why one is saving/investing).
- Paul Kaplan: Deja Vu All Over Again.
- “Bad crap happens about ten times more often than normal distributions suggest.”
- James Xiong: Nailing Downside Risk (more here).
- Standard deviation only helpful as a stand-in for risk when distributions are normal.
- U.S. REITs and global high yield are dangerous when analyzed with better metrics (fund-of-fund optimization) due to non-normal attributes.
- Optimizing Manager Structure and Budgeting Manager Risk (nearly as important as Markowitz).
- “There’s no substitute for common sense.”
- Re models — garbage in, garbage out.
- Asset returns are not normally distributed.
- Investor preferences often go beyond mean and variance.
- M-CVaR favors assets with higher positive skewness, lower kurtosis, and lower variance.
- Certain products (e.g., hedge funds) can depart from normal distribution a lot.
- We need a good fund-of-funds optimizer (recognizes that asset returns are not normally distributed).
- Idzorek and colleagues are working on it.