In a previous article, I looked at the question, “How Much Do I Need To Retire?” In that article, we estimated an annual budget, determined how much would be covered by income from things like social security, pensions and immediate annuities. Then we used the 4% rule to estimate the required size of your nest egg. I mentioned that the 4% rule was for rough planning only. When retirement is imminent, more detail is necessary.
Meet Engineer Wally
Let’s suppose Wally is 52 years old and works at IBM as senior engineer designing computer chips. Wally earns $150,000 per year. He has diligently saved and invested. Wally’s nest egg is currently $2.0 million, according to his latest mutual fund statement. Wally has a balanced 50/50 portfolio and intends to maintain that mix throughout his retirement.
One day Wally comes into work and finds out that his job function has been shipped to India where a PhD EE will do his job at 1/4 the salary.
Divide Budget Into Essential and Non-Essential
First of all, in the budget estimate, Wally should divide his expenses into essential and non-essential or basic and luxury. An important element of a successful retirement is flexibility. When hard times come, he must be able to reduce spending by foregoing luxury items. Suppose Wally determines his desired annual spending in retirement is $100,000 of which $60,000 is essential. Let’s say that Wally’s social security will be $1000 per month and he has no pension. Wally needs $48,000 per year from his portfolio to cover essential spending.
1. Determine Your Longevity
Based on age and expected longevity, determine the number of years Wally will need to fund. To be conservative estimate, assume either Wally or his spouse will live until age 100. Wally should plan for his nest egg to last for 48 years.
2. Determine the Average Balance in your Portfolio
Look at the end-of-month balances for at least the past 12 months. Add them up and divide by 12 to get the average. Use the average of recent samples, rather than the latest balance because it is a better estimate of the size of your portfolio than using a single sample. Wally’s average balance over the past 12 months is $1.8 million, which is less than the current balance.
3. Estimate the Worst-Case Real Internal Rate of Return
Based on the assets in the portfolio, estimate the worst-case internal rate of return. This is not so easy to do, but I can give some guidelines. But first, some explanation is in order.
IRR vs. Market Return
Suppose you put a lump sum into a mix of 50% stocks and 50% bonds for a period of 30 years. Every year there will be some return realized, which will vary year to year. Some years there will be gains, and some years there will be losses.
At the end of the period, you can calculate the annualized rate of return. This is a fictitious number. It is the market’s annualized return, which is the the annual return you would need to have gotten if the return was the same every year, which it was not. For the lump sum investor, the order of returns is irrelevant because 0.5 x 2 = 2.0 x 0.5. (A fancy way of saying this is that multiplication is commutative.)
But for an investor, either adding to or taking out from a portfolio, the IRR will not be the same as the market return. It could be higher or lower and depends on the exact sequence of returns and the timing of the inflows and outflows. Large gains or losses when the portfolio is small will have a small effect, and likewise when the portfolio is large.
This chart shows the market return (in blue) and IRR for someone depleting a 50/50 balanced portfolio down to zero over 30 years (in red).
Notice that the IRR with outflows (red line) will be more than or less than the market’s real return (blue line). The IRR with outflows also has wider swings than the market return. A lucky sequence of returns gives the investor with outflows a higher IRR than the market return.
Use Worst-Case IRR
The second consideration is that we are using the worst-case internal rate of return, not the average or expected rate of return. Think of building a bridge. Engineers don’t design for the average expected load. When the circus comes to town and the elephants march across, the bridge will fail. So you design for worst case plus a margin-of-safety to account for uncertainty.
In the chart above, observe that while the worst case real return for the 50/50 mix was just under %3, the worst-case IRR for the depleting portfolio was about 1.3%. To keep things simple, and provide a margin of safety, we’ll round that down to 1%.
Estimating Worst-Case IRR for Different Periods
We used the chart above for a 50/50 balanced portfolio to estimate the worst-case IRR for a 30-year withdrawal period. But Wally’s retirement needs to last for 48 years. Here is a chart that plots the worst-case IRR for withdrawal periods from 10 to 40 years.
Observe the lower blue line. For a 50/50 balanced portfolio, the real IRR is zero at 20 years, and negative for periods less than 20 years. At 40 years, IRR rises to just above 2%. Wally’s 48-years is beyond the right-side of this chart, but it appears that the IRR asymptotically approaches something over 2%. We’ll use 2% as the worst-case IRR.
Use the PMT Function
Now we have three numbers:
- The number of periods, 48 years
- The average portfolio balance, 1.8 million dollars
- An estimate for the worst-case real IRR, 2%
With these three numbers, using either Microsoft Excel, a financial calculator, or an amortization table, we can calculate the annual payment that will deplete the portfolio down to zero. From Excel, the annual payment number is $58,680. [note: This works out to 3.26% of the average portfolio balance, which is a bit less than the 4% rule of thumb would imply. This is expected because the period is 48 years while the longest period considered by the Trinity Study was 30 years.]
We figured out above that wally needed $48,000 per year for essential spending. It looks like this is covered. Plus Wally should get some luxury spending as well.
Recalculate Every Year
Wally should re-do this calculation annually. This is important, because it provides a course correction every year in case our estmate for the worst-case IRR was way off.
Each year, the term is one less year and his average balance will be different. If Wally’s lucky, his average balance will grow and Wally can give himself a raise. If unlucky, Wally might have to give up some luxuries. If things get really bad, he may have to slash spending to the bone, or even take a part-time job or move to a cheaper location. A great deal of what will happen is purely a matter of luck.
Estimating Worst-Case IRR for Different Asset Allocation
The key to using this methodology is being able to estimate the worst-case IRR for your portfolio. This is not a simple task because the future sequence of returns is unknown. Plus it depends on which assets are in the portfolio.
The chart above was for a 50/50 balanced portfolio and was based on history. The chart also shows the IRR for 100% bonds. Similar charts can be made for other mixes of stocks and bonds using historical data. I will present that in another article.
There are other methods for estimating the worst-case IRR that don’t rely on historical data. I will discuss those ideas in another article as well.
The views expressed in this article are those of the author and do not necessarily represent the views of Laguna Beach Bikini, its editors, staff or any other organization.
