A while back I proposed a possible retirement income strategy where you set aside a fixed number of years of spending somewhere safe (like a high-interest savings account (HISA)) and invest the rest of your savings with the same portfolio allocations you had before retirement. The strategy calls for using your current portfolio balance to choose a spending level.

To determine the amount you can spend, you would assume a fixed investment return and calculate the yearly spending level that would deplete the portfolio by some fixed age. Then if your portfolio either gets higher or lower returns than expected, your spending level would increase or decrease. The HISA savings serve to smooth spending levels somewhat. Further smoothing of spending could be done with some filtering.

If we just focus on yearly spending amounts rather than trying to adapt monthly, what we end up with is a table of spending percentages similar to the RRIF percentages. I created a spreadsheet to calculate the yearly spending percentages as well as the percentage of savings that would be held in a HISA. To edit this spreadsheet, you need to go to the “File” menu and “Make a copy”.

**Spreadsheet Inputs**

The inputs in the spreadsheet are highlighted in yellow. You need to choose an asset allocation (excluding the HISA) and the expected real return for each asset class. “Real return” means the return after subtracting out inflation. Beware of choosing high returns; the higher they are, the greater the risk of your portfolio returns coming up short and leaving you spending too much early on in retirement.

Another very important input is the fees you pay for owning each asset class. I’ve listed 0.12% for stocks because I pay 0.11% in MERs and another 0.01% in commissions and spreads each year. However, few investors have a portfolio with such low costs.

Other inputs are the real return of the HISA, number of years of spending saved in the HISA, your target life expectancy, and your minimum expected remaining life. This last input only starts to affect spending levels late in life. If your target life expectancy is 100, but set the minimum remaining life to 10 years, then spending percentages will stop rising at age 90 in an effort to preserve capital.

**Spreadsheet Results**

The results appear down the left side of the spreadsheet. It shows, for each age, what percentage of the total savings should be in the HISA, and what percentage of the total savings (including the HISA) you can spend each year. Although, the table starts at age 18, the most useful part begins further down the left side at more realistic retirement ages.

The idea is that at the start of each year you would shift portfolio assets into the HISA to match the target percentage. Then you’d withdraw the target spending amount from the HISA during the rest of the year. You might want to add some filtering to the spending levels to avoid sudden changes from one year to the next, but such filtering is not included in the spreadsheet.

It might seem that this spending strategy has you spending more money later in retirement because the percentages rise with age. The idea is that if your portfolio’s return exactly matches expectations, your total savings will drop (due to your spending) by exactly the right amount to keep your yearly spending the same from one year to the next.

If you try to use such a retirement spending strategy, it’s important not to double-dip. If you are spending dividends or interest from your portfolio, that counts as part of the spending allocation for the year. If you are spending forced RRIF withdrawals, the RRIF assets count as part of your portfolio and the withdrawals count as spending. Don’t forget that you may not actually be able to spend all your withdrawals because of income taxes.

This retirement spending strategy can be conservative or aggressive depending on the inputs you choose. The default inputs have a 65-year old spending 4.43% of savings during the year. However, if we reduce the target life expectancy from 100 to 85, drop the years of HISA savings to 3, and increase real stock return to 6%, that percentage jumps to 7.37%. If, instead, we extend life expectancy to 110, increase HISA years of savings to 10, and drop real stock returns to 3%, the spending percentage for a 65-year old drops to 3.14%.

I can’t tell you what spreadsheet inputs give a good balance between safety and living a decent life in retirement. This will depend on your ability to adjust your lifestyle, your other income streams from CPP, pensions, or annuities, and other factors. As always, you can’t follow anything here blindly; think for yourself.

Wow, no comments. I love the idea, but I've been too swamped with work to go through the spreadsheet in detail and say anything intelligent. Just wanted to let you know that there are still people out here who appreciate this sort of thing!

ReplyDelete@Potato: Thanks. I've been motivated to figure this out lately to help some of my older family members. I also want to be able to know when I think I'm truly financially independent. I hope to say more about this next week.

DeleteThank you for your withdrawal worksheet

ReplyDeleteI have a question.

If c is the cash rate and s is the portfolio rate, m is the number of years to keep in cash, and n is the remaining life then you calculate

HISA PV as PV(c,m,-1,0,1) which makes sense. I am not able to understand the Portfolio PV calculation. In my mind it should be PV(s,n-m,-1/(1+s)^m,0,0) but what you have is PV(s,n-m,-1+HISAPV*c,0,0). I believe that the growth of portfolio during the cash years is not considered in your calculation. Can you please explain?

Thank you

Hi ispeuq,

DeleteEach year, the HISA level is steady because we refill the HISA from the portfolio after taking out the year of spending. The amount we pull from the portfolio is one year of spending (this is the -1 part in the PV calculation) less the interest produced by the HISA (this is the HISAPV*c part). The portfolio has to support these payments for n-m years, at which point the portfolio is depleted. The m years of living on cash at the end of life doesn't need to be considered in the portfolio PV calculation because the portfolio will be gone after the n-m years. The HISA PV calculation handled the HISA amount necessary to live for the remaining m years.

If this doesn't clear things up for you, let me know your reasoning behind your version of the portfolio PV. I'm just guessing, but it seems like we have different ideas for how the HISA and portfolio are being drawn down.

Thank you for your explanation. I was looking at it differently.

DeleteI was interpreting PV(c,m,-1,0,1) as how much do I need now to fund $1 (present value) for the next m years at c% interest.

And was looking to see the 2nd formula in the same light, i.e. How much do I need to fund $1 (present value) starting to be paid after m years for n-m years.

I think I part I missed was that some parts of the portfolio will not grow at 's' rate because they get moved into cash and can grow only at 'c'. So my calculation will result in a higher number than actually what is possible.

Does that make sense?

Hi ispeuq,

DeleteI think I understand what you mean. Your interpretation of PV(c,m,-1,0,1) is correct, but this happens at the end of retirement rather than the beginning. For the first m-n years, the cash interest feeds the annual withdrawals and the rest of the annual withdrawals comes from the stock portfolio.