I am receptive to suggestions for future modifications to our calculators. I have mentioned one update in response to Mike's remarks on the From Chapter 3 thread.

Generally speaking, it is easy to introduce new data, allocation percentage changes, new allocation algorithms (extending beyond our current limitations) that are tied to specific years as opposed to years after beginning retirement. I can do that by introducing calculations in the region around row 2550. [If you don't have row 2550 in a particular version that you are interested in, I have given you detailed instructions for adding it. See the post in the From Chapter 3 thread.]

Do not feel constrained in making your requests. I know how to make changes based on years after beginning retirement and even for using absolute dollar amounts. My previous experience along these lines was with the JanSz feature, which removed a fraction of the increase (but not any decrease) in the total balance as compare with six years earlier. It was a tedious, painful process. I know how to do it better and with much less pain and suffering.

Consider this example of what might fall into both the easy category and the difficult category: holding bonds to maturity. The calculator currently treats all bonds as one year trading vehicles without any capital gains, losses or expenses. That is, if you buy a five-year bond at a 4% interest rate in one year and if interest rates jump to 8% in the following year, you draw 4% interest in the first year and 8% in the following year. You cannot lock in the 8% interest rate. If interest rates were to fall back to 4% in the second year, you would be back to getting 4% interest.

There is an easy way for me to do something similar locking in interest rates that is not exactly correct but which might be useful. I could use a trigger (such as a nominal interest rate threshold or a real interest rate threshold) to lock in a rate for 5 years. [We have interest rate information for 5-year treasury notes already in the calculator. If you want to add rates for a different type of bond, there is no difficulty at all. I am hesitant to locking in anything for 30 years.] In fact, we could use something similar to ibonds or TIPS, locking in a real interest rate, absolute minimums (e.g., ibonds can never fall below a 0% real interest rate) and a real dollar amount for par value at maturity.

Here is the catch. I would be locking in a rate for 5 years, not an amount. As long as the portfolio balance stayed reasonably steady, there would be no problem. But as the portfolio balance rose or fell, my bond amount and the interest amount would rise and fall as well. If the balance changed substantially, we would be making trades instead of holding on to previously purchase bonds. The trades would be at the artificially locked in rates.

The better approach, which is more difficult approach, is to define a bond amount that stays fixed for 5 years and to allow the percentage bond allocation to vary according to a well-defined algorithm. I would have to know something about the portfolio's balance relative to the beginning of a particular historical sequence, not a previously specified year. The total dollars of bonds purchased would have to be determined when the initial threshold was crossed. The percentage allocation of everything would vary during the 5-year period.

This is the key step for doing this: use the same procedure as I use for summarizing portfolio balances and annualized returns (both nominal and real). All of my Special and Deluxe versions of my calculators have these tables.

Locate the cell that for the very first 12/31 balance for each year. For 1871, it is cell B215. For 1872, it is cell C233. And so forth. Make a table with the years in column A and the number of years into retirement in columns B, C and so on. (The final column is BJ and it is at 60 years.) In the table corresponding to 1871, write =B215 in column B and use the fill handle to drag the formula all the way to BJ. For the year 1872, write =C233 in column B and use the fill handle to drag the formula to column BJ. And so forth.

Having this table and others like it make it very easy to write algorithms that are based on the balances at a specified number of years into one's retirement. Or so I think.

It is much better for you to ask for five or six very specific calculators than for you to ask for one with a more general capability. If the five or six can be put together, I probably will. But if we start by looking only at a combined capability, there may be a hidden difficulty that causes us to end up with none.

Have fun.

John R.

## Calculator Wish List

**Moderator:** hocus2004