Clicky

The “Perfect” Retirement Calculator

New Reader? Get free regular updates from Can I Retire Yet? on saving, investing, retiring, and retirement income. New articles about twice/month. Join more than 15,000 subscribers. Unsubscribe at any time:

The perfect retirement calculator doesn’t exist yet. But having just finished reviewing dozens of calculators in recent posts — The 3 Best Free Retirement Calculators and The Best Free Retirement Calculators: Round 2 — I can’t stop thinking about the possibilities….

For me, the perfect calculator isn’t one that is somehow more accurate at analyzing your retirement. I just don’t see how it’s possible to precisely predict the trajectory of the economy, the markets, government policy, or your life for decades into the future.

No, a better calculator to me is simply one that is easier to use, more flexible, and better at showing you the range of possible outcomes for your money.

With that in mind, let’s look at my “wish list” from days spent crunching retirement scenarios using the current best crop of retirement calculators….

Input Data

Let’s start with data entry. Every calculator must collect a set of personal data from you before it can attempt to project anything about your future retirement. It takes a surprisingly small amount of key data to run a basic retirement analysis, and that’s one reason I favor simpler calculators to start. I see little sense in collecting lots of detailed data, such as when you’ll sell your house or when you’ll buy your next car, because the “noise” created by all the other major future unknowns will simply drown out any apparent accuracy to be gained.

However, there are several areas where I’ve felt the need for more refinement in input data, in order to accurately model real-world retirements:

Calculators need to allow for flexibility with the inflation rate and Social Security payments. These are both critical parameters for retirement success: inflation directly impacts your real investment returns, and Social Security is a critical income stream making up a substantial portion of many retiree’s cash flow. Neither of these variables is predictable going forward. Both are highly subject to political events. And Social Security can begin at a range of ages, depending on your situation. For that reason, a calculator that assumes a fixed rate of inflation, or assumes Social Security that starts at some set “retirement” age and remains unchanged, is just not being realistic.

Also, surprisingly, in order to accurately and safely model lengthy modern retirements, I think you need to allow for additional post-retirement income. Because, even if you’ve left your career long behind, you may want to work. Or you may need to work, to deal with unexpected personal or global economic developments. A calculator that ignores this possibility for a lengthy retirement, especially a retirement that starts early, isn’t being realistic.

On the flip side, in order to accurately model a real-world retirement, you also need more flexibility in your spending. Many studies report that overall spending tends to go down as retirees age and become less active. Aside from that natural progression, you might adjust your spending based on economic developments: If your portfolio does well, you might take an extra vacation. If it does poorly, you might reduce your spending by a certain percent. These possibilities can be modeled as spending “policies” — rules for how to adjust spending in the face of changes in other retirement variables. And a calculator that explicitly allows for spending policies will help you gain confidence in your ability to handle the unexpected.

When it comes to actually entering your numbers, software should be both flexible and robust. It should provide clear prompts, and accept data in whatever format you choose, while converting it as necessary. Only after exhausting all possible interpretations, should a program generate an error message. For entering numeric values, I prefer modern user interface widgets such as sliders and dials, instead of simple entry fields. Those interactive widgets give clear feedback on the ranges of legal values, and they’re just more fun!

Modeling

There is no exact equation to compute whether you have enough money to retire, or how long that money will last. Far from it. All that retirement calculators can do is try to simulate some key economic variables going forward and calculate where you’d wind up, if your initial assumptions are right and nothing changes. So by “modeling” I mean the capabilities a calculator provides for simulating the future.

Obviously, the more flexible a calculator’s modeling capabilities, the more likely it will churn out at least one set of results that could match your actual future. The heart of the matter for most calculators is how they simulate stock market investment returns going forward. I’ve seen three broad types of models:

  1. average return
  2. historical
  3. Monte Carlo

Average return is the simplest and least sophisticated approach, on the surface. You simply pick a long term average for stock returns, and assume the same average will hold for every year in the future. The big problem with this approach is that it doesn’t account for volatility: Market fluctuations act to rob you of returns, especially when you are withdrawing from a portfolio. You could simply reduce the average return to account for this, though by how much is still a judgment call. Some experts argue that using average returns with lowered values, based on today’s relatively high market valuations, is actually the most sophisticated approach to modeling retirement.

Historical modeling usually samples or replays historical data in some manner to simulate future market returns. The argument for this technique is that we’ve seen all possible variations in the past, so future data will probably behave in the same way. This approach has the appeal of being based on historical “reality,” even if that is a past reality. If you’re of the opinion that things can’t get much worse, or better, than they’ve been in the past century, then you may be comfortable with this approach. Indeed, when you ponder the world wars, depressions, and political upheavals of the last hundred years, you realize the world has already survived a great deal of change. You might conclude there is no reason to expect anything worse to show up. But nobody can be certain.

Monte Carlo is a more academic approach that uses probabilistic methods to model volatility. Typically the technique couples historical market return averages with a standard deviation to generate random volatility, in the hope that this will simulate the possibilities for actual market volatility going forward. Hundreds or thousands of “runs” are performed to calculate the ending values of your retirement portfolio for each random variation of market returns. Those ending values are then organized into “confidence” bands, giving you results such as an “85%” chance you won’t run out of money. This is a very popular approach and if you are comfortable thinking in terms of probabilities, it may be for you. But there are serious issues: It turns out that the “artificial” randomness generated by retirement calculators may not accurately reflect the real-world randomness of the stock market. And some experts argue that probabilistic results simply can’t be applied effectively to individual cases, like your own retirement.

Each of these models has its proponents and detractors. There are strong arguments from some very bright and experienced people that one approach or another is better. But not all the most trustworthy voices have settled in the same camp. And that’s where I take issue with the current state of affairs….

I’m just an early-retired engineer who wants to see all the data before making my own personal decisions. So I want a retirement calculator without an agenda — one that runs each of these possible models on the same input data and then compares the results for me to evaluate.

Perhaps, when using my own data, the results from all of the models are quite similar. Or maybe one model is an outlier. In any case, I’d like to see them all on the same page. I don’t want to have to choose just one modeling philosophy when I use a retirement calculator, or to have to use multiple calculators, with attendant input variations, just to get the complete picture.

Output Results

If a calculator provides extra flexibility in how it models variables like inflation, Social Security, and investment returns, then it’s probably not going to produce a single set of results for your retirement. Rather, it will produce a range of possibilities, depending on the values of the different input variables.

What happens if we have historical investment returns, but inflation runs high and Social Security is reduced? I’d like a calculator that makes generating and comparing multiple scenarios very intuitive. At a minimum it should be possible to easily create and label those scenarios, review their input data, then clearly compare the results on a single screen.

Related to comparing scenarios, the way a calculator presents its results is critical to usability. You use a retirement calculator for critical guidance on major financial decisions, so the answers should be as clear as possible. Many calculators do a decent job in this area, though few get everything right, in my opinion.

There are two primary means of presenting output: tabular and graphical. I think a good calculator should offer both. Tabular output means creating a table or spreadsheet, with a line for each year, and columns for key parameters such as beginning balance, contributions, investment earnings, spending, taxes, and ending balance.

Graphical output is simply a chart, usually with years on the horizontal axis and portfolio value on the vertical axis, showing what happens to your savings over time. A picture is worth a thousand words. If your savings craters into zero at age 71, this is not good. If it grows exponentially into your 90’s, then you might be able to live it up a bit. The reality of course is a band of possibilities — not just one outcome. So a compact method for displaying median values, and 10% and 90% success bands, on the same chart, can be very instructive.

Also, not to be forgotten, a restatement of the input data and key modeling assumptions — What monthly expenses did I input? What inflation rate was used for the calculation? When did Social Security start? — is extremely useful when reviewing an analysis at a later time. Only a few calculators get this feature entirely right.

Overall Metaphor

My final wish list item for the “perfect” retirement calculator is both the simplest and the most radical.

For starters, understand that I’m uncomfortable with the entire thesis of most modern retirement calculators — that they are going to tell me precisely whether I can retire or not — a simple “yes” or “no” answer.

As I’ve discussed before, the retirement question is ultimately unanswerable in precise terms, because that would require predicting the future of the stock market, the economy, government policy, and your lifespan. Predicting future investment returns, even over long periods, is an inexact science at best. Guessing what inflation will be or what governments will be in power decades from now is silly. And for most of those in average health, being required to predict your life expectancy is inherently flawed.

So I object to the entire premise of most retirement calculators. Trying to resolve such a complex question with a simple “yes” or “no” answer borders on lunacy. There simply is no such simple answer to the retirement question. To me that means the underlying pass/fail metaphor that most retirement calculators use is totally wrong.

All a retirement calculator can really provide is a prediction for how long your assets would last under current conditions. So, for that reason, I want a simple fuel gauge metaphor that tells me: Given these assumptions, this is how long your money will last.

Then I will look at that fuel gauge, look outside the window at the condition of the world, look in the mirror at my own personal condition, and I will decide how I should be driving my retired life forward at the moment….

I just want a retirement calculator to tell me how many years my money will last, based on what it knows now. I just want a fuel gauge.

That’s much more realistic than the pass/fail approach. And I suspect it’s how most people, myself included, will actually manage their retired life.

* * *

As I’ve shown in my previous retirement calculator articles, the best of the current generation of calculators is quite good. But, as I’ve detailed above, there is always room for improvement. Whether it’s input data options, modeling flexibility, output data visualization, or overall design metaphor, there is still much to be desired from retirement calculators in general.

Will one of the current favorites rise to the challenge? Or will a newcomer take the prize? Either way, I’ll be keeping my eye on the field of contenders to see which one raises the bar next….

Comments

  1. Did you look at Optimal Retirement Planner at http://www.i-orp.com/ ?
    This is the calculator I like.

    • Thanks Dave. I did look at Optimal Retirement Planner briefly. The optimization approach to minimize tax on withdrawals was interesting. I remember seeing that the recommendation for withdrawal rate was on the high end. And I believe it uses historically computed averages for return on investment and inflation which may not hold up going forward. I didn’t know enough about linear programming/operations research to validate the approach being taken, but if you or anybody else has insights please add them.