Estimate Your Personal Rate of Return for Multiple Years

Since I wrote about a simple formula for estimating your personal rate of return, someone asked whether the same formula works for multiple years as well. The answer is yes and no. It works well, provided that

1. the net investments during the period are roughly even; and
2. the beginning balance is large relative to the net investments

The simple formula doesn’t work well if these two conditions are not met.

After applying the simple formula, you get a cumulative return. You will have to annualize the result using this formula:

average annual return = (1 + cumulative return) ^ (1 / number of years) – 1

Here’s an example:

Beginning Balance 1/1/1998: $20,000
net investment in 1998: $1,000
net investment in 1999: $1,100
net investment in 2000: $1,200
net investment in 2001: $1,300
net investment in 2002: $1,400
net investment in 2003: $1,500
net investment in 2004: $1,600
net investment in 2005: $1,700
net investment in 2006: $1,800
Ending Balance 12/31/2006: $50,000

Total “net in” = $1,000 + $1,100 + … + $1,800 = $12,600

cumulative return = ($50,000 – 0.5 * $12,600) / ($20,000 + 0.5 * $12,600) – 1 = $43,700 / 26,300 – 1 = 66.2%

This means for this example, the cumulative return in 9 years from 1/1/1998 to 12/31/2006 is 66.2%. To convert it into an annualized number, you do:

average annual return = (1 + 66.2% ) ^ (1/9) – 1 = 5.8%

If I use the more precise XIRR method for this example, I get 6.0%. So 5.8% is close enough. But, if we have a different set of numbers,

Beginning Balance 1/1/1998: $2,000
net investment in 1998: $1,000
net investment in 1999: $1,100
net investment in 2000: $1,200
net investment in 2001: $1,300
net investment in 2002: $1,400
net investment in 2003: $1,500
net investment in 2004: $1,600
net investment in 2005: $1,700
net investment in 2006: $1,800
Ending Balance 12/31/2006: $25,000

Note the beginning balance of $2,000 is small relative to the net investments of $12,600. The simple formula gives the result of 9.4%, while the more precise XIRR method gives 11.0%. In this second example you are better off with the XIRR method because I think the 1.6% difference is too big.