If you hope to retire early, the big question is: Do I have enough money to retire safely? Tons of websites, retirement calculators, and rules-of-thumb (like the 4% rule) try to answer this question simply. But most of them don’t paint a complete picture because:
- The details behind the calculators or rules-of-thumb are often murky.
- You need to know the most reasonable or best inputs for your situation.
- You can’t entirely tailor them to your specific situation.
I think one of the best of these resources is FIRECalc, which allows for fairly detailed variations and estimates the probability that you might run out of money in retirement by calculating all possible sequences of historical investment returns. But even in this case, some details of the calculations aren’t completely explained, it assumes a certain level of understanding about the appropriate calculator inputs, and many potential specifics are omitted. I’ll cover a bit more about the features and limitations of FIRECalc in Part 2 of this series coming soon.
For my own early retirement decision, I conducted independent calculations tailored specifically to my situation. And I suspect I’m not the only would-be early retiree out there who would prefer not to leave this momentous life decision solely in the hands of standard rules or generic calculators. So, in this two-part series, I’m going to give you an overview of my own calculations (Part 1) and how I evaluated the results (Part 2). If after reading this, you’d like to use my spreadsheets as a starting point for your own calculations, just contact me, and I’ll send you a copy.
Because I’m not a big fan of blogs with lots of “personal finance updates”, I’ll try my best to make these posts less “show and tell” and more an “instruction manual” to help you with your own tailored retirement calculations.
Gathering Information
If you think you might retire soon, you need to compare the size of your nest egg (or estimate your nest egg size a few years from now) to your long-term spending needs. Many retirement calculators are unhelpful at this stage of life because they focus on the accumulation phase and ask for inputs like current salary, savings rate, and years to retirement. For the would-be early retiree, that stuff is mostly just water under the bridge.
Accordingly, to decide whether it’s time to retire early, you’ll need to gather the following more relevant information:
- Total retirement savings, broken down by tax-advantaged and taxable accounts.
- An at least rudimentary plan for how you will invest your retirement savings (e.g., all stocks, 50% stock/50%bonds, etc).
- An annual spending budget (either before tax or after tax, depending on how you like to think about it).
- A spending plan for the period before your tax-advantaged accounts are accessible without penalties.
- Social security or pension income and when you might choose to receive that income.
- Any other income (such as rental properties or ongoing “side gigs”)
- Other specifics that may substantially impact any of the above, such as plans to,
- Sell a home and invest the proceeds,
- Move to a cheaper area
- Pay off a mortgage
- Use tax minimization tactics
- Achieve more or less rental income over time
- Have one time or periodic extra spending needs (kids college education, living abroad, big purchases, etc.)
If developing all this information seems daunting, then I’d say your generally unprepared for early retirement, not for financial reasons, but for planning reasons. Retiring early without being clear on these sorts of details is pretty much like playing Russian Roulette. It might all work out in the end, but only because you got lucky.
If instead you’re worried about exactly how much flesh to put on these bones, I’d suggest to you start simple and work towards more detailed information where necessary. Starting with broad estimates allows you to discover which parameters drive your chances of success and where you need more detailed inputs.
For example, I didn’t build an elaborate spending budget, because my wife and I practiced for several years by spending what seemed reasonable for our future retirement. For social security income, I used the projections of future income mailed out periodically by the Social Security Administration. I also had a pretty good ball-park estimate of how much money I would have accumulated for various potential retirement dates and used my past tax records to estimate future rental income. My investing plan is mostly consistent with the articles presented here at Mindfully Investing. In a nutshell, this means I have a portfolio of 80% stocks and 20% cash at the start of retirement for reasons detailed in the articles. The other details were all pretty obvious for us, including that we had already paid off all mortgages, had no plans to sell any
Calculation Set Up
I actual spreadsheet I set up was pretty simple, at least in the first iteration. Here’s a blank example of the set up.
The rows represent each successive year of retirement. I projected out for 34 years*, when I will be 86 years old, if I make it that far. It seems like a reasonable time span based on my family history and the fact that my wife is a bit older than me.
The right-most three columns start at the top with the amounts of money we accumulated in each of these accounts as of retirement, with the fourth column being the total. “Retirement accounts” refers to tax-advantaged accounts like IRAs and Roth IRAs.
The rows in the “annual return” column contain the nominal (not inflation adjusted) annual return assumed for each year. Of course, no one can accurately predict future asset returns in the next year, much less for the next 34 years. To represent this uncertain future, I conducted multiple iterations using various historical sequences of nominal returns to see how the growth or decline of the overall portfolio might vary. Looking at many potential future return scenarios is critical, because sequence of returns is one of the key determinants for retirement portfolio success. Because the non-cash accounts are 100% stocks, I used the historical sequences of stock returns from Robert Shiller going back to 1871. If you have a portfolio that includes bonds, you can also get historical estimates of bond returns from this same article. (For historical returns of other assets, you’ll need to do your own research, but Portfolio Visualizer is a good place to start.)
You often see probabilistic analyses that use every possible sequence of historical returns. In my case, I simplified this considerably by using 12 historical return sequences with the start of each sequence shifted forward by 10 years. So, the first sequence was the 34 years from 1872 to 1905, the next sequence was from 1882 to 1915, and so on, with the last sequence ending in 2015. I also added two more “worst case” sequences, one of which represents the lowest annualized return for any 34-year period and another that starts with the largest single stock crash around 1929.
The “spend budget” is the amount we expect to spend each year. This value is inflation adjusted at the rate assumed in the next column. For my first iterations, I assumed a 2% (1.02) annual inflation rate. Also, in this column, I added specific big-ticket expenditures in the years that I expect those to occur.
The next columns are social security and rental income we expect to receive. For rental income, I assumed some moderate rent increases to keep pace with inflation.
The “withdraw need” column calculates the amount that we would need to withdraw from one of the first three columns to fulfill our spending needs after including rental and social security income. Consistent with the Mindfully Investing bucket plan, I made some simple assumptions about the source of the spending money in any given year. The cash accounts are drawn upon first. The tax-advantaged accounts are drawn upon last, mainly because the government penalizes withdraws from tax-advantaged accounts before you reach retirement age. Also, as you get older, there are Required Minimum Distributions (RMDs) from tax-advantaged accounts. Our actual withdrawal plan is considerably more nuanced than this for tax minimization reasons. But the main goal of this exercise is to check whether we could run out of money in retirement. The exact details of money shuffling between accounts won’t impact that determination much.
The last two columns are mainly informational and track the cumulative effect of the selected sequence of annual returns and how much the portfolio is drawn down during sequences involving market declines.
The key result is whether the “total money” column goes to zero before the end of the retirement period (34 years in the future). You might also be interested in how much money is left over at the end of the period, if you are hoping to pass on some inheritance to relatives or charity after your death, but that’s not a big concern for us.
Calculation Results
The set up of the various sequence of returns and some select results are shown in this table.
| Scenario | First Year | Last Year | Nominal Annualized Return (CAGR) | Annualized Inflation Rate | Worst Cumulative Draw Down | End Value/ Start Value | Lowest Value |
| 1 | 1872 | 1905 | 7.2% | -1.2% | 0.0% | 5.59 | 0.79 |
| 2 | 1882 | 1915 | 6.3% | -0.2% | 14.1% | 3.52 | 0.79 |
| 3 | 1892 | 1925 | 7.9% | 2.6% | 13.8% | 6.95 | 0.80 |
| 4 | 1902 | 1935 | 7.2% | 1.7% | 10.2% | 5.39 | 0.86 |
| 5 | 1912 | 1945 | 8.0% | 2.0% | 3.4% | 7.15 | 0.87 |
| 6 | 1922 | 1955 | 11.0% | 1.4% | 0.0% | 25.74 | 1.20 |
| 7 | 1932 | 1965 | 13.5% | 2.4% | 5.8% | 38.70 | 0.92 |
| 8 | 1942 | 1975 | 11.5% | 3.7% | 0.0% | 30.95 | 1.14 |
| 9 | 1952 | 1985 | 10.9% | 4.3% | 0.0% | 23.91 | 1.08 |
| 10 | 1962 | 1995 | 11.3% | 5.0% | 9.2% | 17.10 | 0.89 |
| 11 | 1972 | 2005 | 11.0% | 4.8% | 26.0% | 22.28 | 0.70 |
| 12 | 1982 | 2015 | 11.3% | 2.8% | 0.0% | 28.75 | 1.13 |
| Worst Start | 1931 | 1964 | 10.4% | 2.0% | 63.2% | 3.90 | 0.37 |
| Worst CAGR | 1881 | 1914 | 5.5% | 0.2% | 13.9% | 2.60 | 0.77 |
The second to last column displays the key result in terms of the ratio of the ending portfolio value divided by the starting portfolio value.** The key finding is that none of these ratios is zero, which means that none of these sequences of returns resulted in our portfolio running out of money within 34 years. This graph traces the ups and downs of each scenario.
It turns out that our starting annual spending budget is about 2.8% of our starting portfolio. Given that a spending rate starting at 4% of the initial nest egg is considered pretty safe for 30 years, it’s understandable that my retirement portfolio can successfully weather just about any stock market storm.
However, before launching into retirement, there’s one detail about this type of analysis that needs some further consideration. I noted that these calculations are all on a nominal-return basis, with the withdraw rate boosted each year by 2% to account for inflation’s impact on the purchasing power of the money. This assumption seems reasonable given that the table above shows it’s pretty rare for inflation to stay above 3% for more than 30 years. But some long periods of 4 to 5% inflation have occurred in the past. So, what if I instead assume that inflation is 4%? Here’s what that looks like.
| Scenario | First Year | Last Year | Nominal Annualized Return (CAGR) | Annualized Inflation Rate | Worst Cumulative Draw Down | End Value/ Start Value | Lowest Value |
| 1 | 1872 | 1905 | 7.2% | -1.2% | 0.0% | 3.81 | 0.78 |
| 2 | 1882 | 1915 | 6.3% | -0.2% | 14.1% | 2.08 | 0.78 |
| 3 | 1892 | 1925 | 7.9% | 2.6% | 13.8% | 5.12 | 0.80 |
| 4 | 1902 | 1935 | 7.2% | 1.7% | 10.2% | 3.73 | 0.84 |
| 5 | 1912 | 1945 | 8.0% | 2.0% | 3.4% | 5.28 | 0.86 |
| 6 | 1922 | 1955 | 11.0% | 1.4% | 0.0% | 22.09 | 1.20 |
| 7 | 1932 | 1965 | 13.5% | 2.4% | 5.8% | 34.96 | 0.92 |
| 8 | 1942 | 1975 | 11.5% | 3.7% | 0.0% | 29.42 | 1.14 |
| 9 | 1952 | 1985 | 10.9% | 4.3% | 0.0% | 21.72 | 1.08 |
| 10 | 1962 | 1995 | 11.3% | 5.0% | 9.2% | 13.62 | 0.89 |
| 11 | 1972 | 2005 | 11.0% | 4.8% | 26.0% | 19.73 | 0.69 |
| 12 | 1982 | 2015 | 11.3% | 2.8% | 0.0% | 27.09 | 1.13 |
| Worst Start | 1931 | 1964 | 10.4% | 2.0% | 63.2% | 0.67 | 0.31 |
| Worst CAGR | 1881 | 1914 | 5.5% | 0.2% | 13.9% | 1.43 | 0.73 |
Even with the 4% inflation assumption, none of scenarios run out of money before the end of the projection, although the “worst start” scenario comes close. The worst start scenario illustrates how the annual withdraws start to sky-rocket pretty quickly as the inflation adjustment is increased. For example, with 4% annual inflation, an initial withdraw of $50,000 per year becomes almost $110,000 per year after 20 years. This shows the value of doing your own projections. You can change a specific assumption and see exactly how it impacts your results. It also shows that the amount you plan to spend strongly impacts your chances of retirement success.
Conclusion
These results raise some new questions. What’s the best way to include inflation in these sequential return estimates and evaluate the results? Is the 2% or 4% inflation assumption more realistic? Do my projections appear conservative or aggressive in comparison to more generic online calculators? How could I adapt my spending to decrease the chances of failure in a “worst start” type of scenario, particularly if inflation starts to increase substantially?
In other words, completing the calculations is really just the end of the beginning of deciding whether you can safely retire. In my next post I’ll get into the more intricate matter of evaluating the results, which should finally answer the question, “Can I Retire Now?”
* This specific number comes from the fact that last year I conducted these projections for a 35 year period and now one of those years is in the rear-view mirror. I didn’t want to revise all the calculations just to add one year back.
** I know some readers were probably hoping to discover my net worth. When it comes to broadcasting my financial details all over the internet, I guess I’m pretty old-fashioned as compared to a lot of personal finance bloggers. So, I’m presenting all the results as ratios of the starting nest egg value.
