Dissection of an Investment Presentation (Part 4)

Without doubt, the “Wealth Indexes 1927 – 2014” graph is impressive. The comparator (S&P 500) gets dwarfed by both high value and high momentum curves.

As with the Part 3 line graph before it, though, this graph raises lots of questions. Again, no drawdown (DD) analysis is shown. Graphing “growth of $1” on the y-axis makes it difficult to tell whether any DDs exceeded initial equity value. This is a no-no because merely changing the starting date could mean the difference between survival and Ruin. I think the proper way to backtest a strategy is to use constant position sizing and an initial equity value that exceeds maximum DD [perhaps with a margin of safety, too]. If they did this then supporting evidence should have been given.

Posting index returns raises questions about dividend yield inclusion and taking steps to avoid survivorship bias.

Index benchmarks imply buy and hold (B&H), which raises the tax issue. I asked Longs Peak* how GIPS handles taxes:

     > Reporting performance net of tax is not very common so
     > taxes themselves will not affect performance. If the strategy
     > is different because the accounts are managed in a tax
     > sensitive way then… [compare] accounts with similar tax
     > situations… many firms create… taxable and non-taxable…
     > composites because they implement their strategy in…
     > different way[s]… You want to ensure you are making an
     > apples-to-apples [comparison]… [via e-mail].

If I knew trades in the test and benchmark accounts involved different levels of capital gains then I might include an adjustment to err on the side of conservatism. For example, I could subtract the long-term capital gains rate from the highest tax rate and decrease equity curve growth by that amount.

Along with DD analysis, another phenomenon often excluded from comparison studies is flat time. Flat time is the [longest] time interval between equity highs (-water marks). Flat times are inversely proportional to new account highs.

While it’s hard to measure precisely given the wide axis scale, the high-momentum case shows at least a couple instances of 5-10-year flat times. It might be difficult for some people to stick with the strategy through these periods. I certainly would not recommend this to anyone looking to trade full-time as a business or for those seeking an optimal investing approach. I do believe people can benefit from a B&H strategy, however. I think this is characteristic of most garden-variety offerings that present significant improvement in the face of subpar returns.

Finally, the exponential y-axis suggests the graph reflects remaining 100% invested. This is either not realistic or carries a greater risk of Ruin (e.g. here or here). I believe optimal investing should allow for the possibility of locking in profits over time, which remaining fully invested does not.

* Sean Gilligan has been very helpful with GIPS compliance questions over the phone and through e-mail.