When making regular purchases of an investment (i.e., dollar cost averaging into it), volatility tends to reduce the average price of shares purchased. When dollar cost averaging out of an investment, volatility does the same thing–it reduces average share price.
Of course, when you’re on the selling side, reducing the average price per share is not a good thing. As a result, any asset allocation method for a portfolio in its distribution phase must either:
- Accept the fact that volatility will probably be harmful to returns, or
- Make an effort to temper the volatility.
Which Volatility to Reduce
If your portfolio consists entirely of an S&P 500 index fund, and you’re selling shares each day to fund living expenses, the volatility you’d be concerned with is the daily volatility of the S&P 500. If you’re selling once per month, you’d be concerned with month-to-month volatility. If you place one sell order each year, it would be annual volatility that concerns you.
In other words, the volatility that matters is the volatility of the investment(s) you’re selling over periods of time equal to the frequency with which you’re selling. And, therefore, it’s likely beneficial to make efforts to set your selling frequency equal to periods over which the investment(s) you’re selling has historically had the lowest volatility.
When DCA’ing out of stocks, this suggests that reducing the frequency of your sell orders as much as possible is likely to be beneficial (because volatility of stock returns is inversely related to the length of the period of time considered).
The “Buckets Method” of Asset Allocation
I’ve written before about the “bucket method” of asset allocation in retirement. It typically consists of setting up the following three “buckets” (or something similar):
- 2 years of living expenses kept in a money market account,
- 3 years of living expenses kept in short-term bond funds,
- The remainder of the portfolio uses a static allocation (often near 40/60 stock/bond).
Then the portfolio is rebalanced annually, making sure to fill the first and second buckets back up each time.
The problem is that this method still leaves an investor with the return-damaging effects of DCA’ing out of volatile investments, because the annual rebalancing will amount to annual selling of stocks and/or long-term bonds in order to refill the first two buckets. In other words, even though the investor is spending out of the money market, he’s still funding the money market with regular, frequent sales of stock.
Jim Otar, however, offers a slightly different method in his Unveiling the Retirement Myth–a method which makes an effort to minimize the return-damaging effects of volatility while selling.
Otar suggests the following rules are followed:
- Living expenses are paid for out of the money market bucket, then (if that is depleted) out of short-term bonds.
- When the portfolio receives a cash inflow (dividends, interest, tax refund, etc.), first top off the money market bucket (up to 2 years again), then top off the short-term bonds (up to 3 years again), then invest any remainder in equities or long-term bonds.
- If rebalancing from equities to fixed income, first use the cash to top off the money market bucket, then top off the short-term bonds bucket, then purchase long-term bonds with any remaining money.
- If rebalancing from fixed income to equities, first use the cash to top off the money market bucket, then top off the short-term bonds bucket, then purchase equities with any remaining money.
- Keep rebalancing of the 3rd bucket (the one made up of equities and long-term bonds) to a minimum. Otar suggests once every 4 years.
The idea is to do everything possible to reduce the frequency with which stocks are sold–by keeping enough cash to fund living expenses and by keeping rebalancing to a minimum.
Sounds like a reasonable approach to me.