Tranche is a French word meaning “to slice or section.” When I see this term I immediately think of Fidelity Investments as this is exactly what they did with their mutual funds in the 1980s when they created their select funds. The idea was to divide the market into thin slices or sections so each month, in some financial magazine, they could tout that they had the top performing mutual fund. While it was a great marketing tool, it did little for investors as each month Fidelity also had the poorest performing mutual fund – some other select fund. How was the investor to know which end of the dog they were buying? The same is now true with ETFs where we can find an ETF that specializes in some niche of the stock market.
We are using the Tranche Model, not as an advertising gimmick, but as a way to reduce the luck-of-review-day. To implement this model, check out the Tranche Spreadsheet.
Referenced elsewhere on this blog is a link to an article titled, “Minimizing Timing Luck With Portfolio Tranching.” You can get the gist of the article without digging in or understanding all the mathematics. Ernie Stokely and John Dishman brought up this idea earlier. In a few days expect to read Ernie’s back-testing results.
Assume you have a portfolio of $100,000. To reduce rebalancing luck, experiment using the following guidelines.
- Break the portfolio into quartiles of $25,000 each.* I recommend tracking each using the TLH Spreadsheet.
- Set up your momentum model, not unlike that used with the Galileo or Rutherford portfolios.
- On your computer calendar, set up rebalancing periods.
- In my case, I have the portfolios set up to be reviewed every 33 days.
- If we use this coming August as an example, portfolio 1 is scheduled to be rebalanced August 4th, portfolio 2 on August 12th, portfolio 3 on August 20th, and portfolio 4 August 27th.
- When a portfolio rebalancing date falls on a Saturday, move that single evaluation up to Friday and when it falls on Sunday, move it to Monday. Do something similar for holidays.
Excluding the passively managed portfolios, the remaining eleven (11), while belonging to different individuals, are managed in a “tranching” fashion.
* We now have a spreadsheet that accomplishes what we are trying to do much better than the quartile approach.