How are the retirement paycheck and guardrails calculated?

"How much can I spend?" vs “What are the chances I'll run out of money?”

One of the most amazing and valuable parts of Income Lab is the ability to quickly and directly answer the question “How much can I spend?”. In traditional legacy financial planning software, this question is not addressed. Instead, typical planning software answers the question “What are the chances I'll run out of money?”. This does not deliver a plan for life, but rather a score called “Probability of Success”. 

Moving your planning to Income Lab means discarding “Probability of Success” and embracing a more client-centered approach that directly answers the following questions in dollar terms.

1. How much can I spend? ("retirement paycheck" or “spending capacity”)
2. What would change that?
3. What could those changes look like?

These are the questions answered on the main Income dashboard in Income Lab.

But the obvious question is, “How did you get to those numbers in particular?!?” The answer involves a 3-step process:

1. Run a simulation analysis to determine a “cloud of possible futures” based on plan assumptions.
2. Determine the income levels that would be possible in each possible future in this cloud
3. Pick a reasonable, relatively conservative income level from this “cloud” that allows good enjoyment of life but does not take on too much risk

In the rest of this article, we go over these steps. We then address how the guardrails and adjustment levels are calculated.

Simulation Analysis and a “Cloud of Possible Futures”

To answer the core question of “How much can I spend?”, we use “Simulation Analysis”, which is just a fancy term for modeling a wide range of ways the world might work in the future. Everyone knows that we don't know the actual future. But we do have ways of making reasonable assumptions about the range of possible outcomes. For example, we know it's possible that stocks could turn in a -50% year or a +30% year, but we also know these are not a common occurrence. We know inflation is more likely to be 2% to 3% than 12% in a given year, but 12% remains a remote possibility. 

Using your assumptions, we simulate a thousand ways* that the world could work by varying:

• Investment returns: Returns could be high, low, medium, or any mixture of these
• Inflation: Inflation could be high, low, medium, or any mixture of these

By exploring these uncertain parts of the future, Income Lab builds a large “cloud of possible futures”. (In addition to this, for joint plans, Income Lab also applies mortality adjustments to the plan when doing the simulation analysis to recognize the fact that we do not know how long someone will live, and that how long someone lives can affect the resources available to the surviving spouse.)

In each of these scenarios, there is an amount that someone could spend based on the particular thing in a particular plan - the perfect spending level (“retirement paycheck” or “spending capacity”) for that specific scenario, if you happen to know you are in that scenario. If it were a good scenario (high returns, low inflation), it would support a high spending level. If it were a bad scenario (low returns, high inflation), it would only allow for lower spending. 

The spending level that is available in any given scenario depends on the plan itself - its resources, investments, asset allocation, fees, plan length, the timing and character of its non-portfolio income (Social Security, pensions, etc.), the timing and character of its special spending goals (other/variable expenses), and so on. Even small changes in a plan will change the spending that is available in each scenario, and therefore change the plan's retirement paycheck.

So, the next step is to determine the possible spending levels for each scenario.

Picking a “Retirement Paycheck”

To determine a plan's retirement paycheck, Income Lab delivers one of the spending levels from this “cloud of possibilities”. Which one do we pick? That depends on the plan's settings, but the default is to pick a very reasonable, conservative one from the cloud of possibilities. It won't be the lowest (though that is an option, if you prefer) or the highest, or even the median. By default, a plan will show the spending level that is lower than 80% of the possible retirement paychecks from the “cloud of possibilities”. (But you can adjust the Income Settings of the plan to pick a different one.)

For more, see What are the default values for the Income Settings slider?

Risk of Overspending and Risk of Underspending

Diving deeper into how this spending level is picked, imagine that instead of a big, unorganized cloud of possible retirement paychecks, we lined up all of the paychecks in that cloud by size and put them in a bell curve (called a “histogram”). That graph would look something like this.**

Some spending levels from the cloud of possibilities are very low (bad returns, high inflation) and some are very high (good returns, low inflation). But these extremes are uncommon and find themselves in the thin "tails” of the graph. As we get toward the middle of the bell curve, there are more scenarios that group together. The middle, or median spending level is, mathematically, the “most likely” or “best guess” spending level from the simulation analysis. Anything lower than this is, by definition, expected to be underspending the resources of this plan, or, to put it more positively, living within this plan's means. Anything above this median level is expected to be overspending the resources and running too hot.

Think of moving right on this graph as increasing the risk of overspending and decreasing the risk of underspending. Moving left on this graph  decreases the risk of overspending and increases the risk of underspending. For more information on framing risk in terms of overspending and underspending, see Reframing “Retirement Risk” as “Over- and Under- Spending” to better communicate to clients on Kitces.com.

By default, all of Income Lab's initial retirement paycheck options are left of the median, meaning they are all in the “underspending zone” or “living within our means”. The default is a spending level that is at the 20th percentile in this graph: 20% of the cloud is to its left and 80% is to its right.

The Income Setting slider in our software allows you to choose anywhere from the 1st percentile (all the way to the left on the slider) to the 40th percentile (the right-most setting).

Finding the Guardrails

To answer the question “When would this change?”, we need to find upper and lower guardrails. Conceptually and mathematically, these guardrails are:

• Upper Guardrail: Answers the question, “When is risk of underspending too high?”
  When are we spending so little that we could afford to spend more?
• Lower Guardrail: Answers the question, “When is risk of overspending too high?”
  When are we so sure that we're spending too much that it's worth making an adjustment to the plan?

Although guardrails can be customized, for all of our default settings in the Income Setting slider we use the same settings. For the upper guardrail, we define “risk of underspending is too high” as a 100% chance of underspending. To translate this into an actual balance dollar amount, we figure out the balance at which the current retirement paycheck would be the smallest spending level in the cloud of possible spending. In other words, we find the balance at which the current retirement paycheck would have a 100% of underspending.

We want to peg the lower guardrail to a level in the overspending zone. (To adjust while current spending is within the underspending zone would be irrational.) But adjusting at a 51% chance of overspending would in all likelihood be jumping the gun. (At that point it's still fairly likely we're NOT overspending.) So, our default is to wait until the risk of overspending is 75% (and risk of underspending is 25%). In other words, we wait to trigger a downward adjustment until we are in the middle of the overspending zone. To translate this into an actual balance dollar amount, we figure out the balance at which the current retirement paycheck would be at the 75th percentile in the cloud of possible spending.

Finding the Change in Retirement Paycheck

Once a guardrail is hit, question #3 arises: “What could those changes look like?”. In other words, now that I've hit a guardrail, what should I do? In practice, there are many adjustments someone could make after hitting a guardrail. Among these are:

1. Change monthly spending
2. Adjust future planned spending
3. Adjust portfolio legacy goal

However, when setting expectations for a plan, the Income Lab Income Dashboard shows the change in retirement paycheck that would be triggered if the guardrail were hit exactly. In other words, it asks, “If my portfolio went from where it is today to the guardrail level, how would I adjust my retirement paycheck?” (If the balance went far above the upper guardrail or far below the lower guardrail, the answer would of course be different.)

The answer to this question depends on the “speed” with which you adjust. Based on extensive testing and client preferences our defaults are:

• Hit Upper Guardrail: Increase retirement paycheck 100% of the way back to the original target risk of underspending/overspending. In other words, embrace all of the newly available extra spending.
• Hit Lower Guardrail: Decrease retirement paycheck 10% of the way back to the income we would have if we were at the original target risk of underspending/overspending. In other words, ease into the reduction in spending in order to avoid overreacting and unnecessarily restricting spending.

If someone hits the upper guardrail, their risk of overspending is now very low. Importantly, it is below their risk at the beginning of the plan. So, when faced with a decision about how much more to spend, it's normal for someone to say, “Hey, why not just go back to a spending level that has the same risks I had been accepting when I started this plan?” So, the adjustment is 100%. This also leans into our normal preference to embrace good news and a more abundant life.

If someone hits the lower guardrail, in contrast, the discussion is about how and where to trim. The two ends of the spectrum would be (a) make a huge adjustment and drastically reduce risk now, or (b) make a small adjustment and see if that is enough (if it's not, continue to make small adjustments until things look stable or turn around). Typically, people prefer option (b). That's because people don't like reducing their standard of living or restricting spending. But it's also because a lower guardrail hit is likely due to market turmoil. If reversion to the mean is likely in markets, then when a lower guardrail is hit it may be that we are nearing a recovery. In that case, clients would rather not make huge spending cuts now, only to be told months later that they can now spend more. Instead, clients prefer to make small adjustments and see if they're enough. This approach prefers the possibility of several small adjustments to just one big one.

As an example of how this 10% adjustment works: If the current retirement paycheck is $10,000 and the paycheck at the original 80% risk of underspending is $6,000, a 10% adjustment would take the retirement paycheck to $9,600 ($10,000 - $6,000 = $4,000 * 10% = $400).

This approach also means that, after hitting the lower guardrail and adjusting the retirement paycheck, the lower guardrail will still be relatively nearby. You will see this clearly in the Retirement Stress Test.

FOOTNOTES

* For Monte Carlo analysis types, Income Lab uses 1000 scenarios in the simulation analysis. For the historical analysis type, the number of scenarios depends on the plan length and is usually much more than 1000. For more on how the number of scenarios affects simulation analysis, see How many Monte Carlo simulations are enough? on Kitces.com.

** For those who are more mathematically inclined: The shape of this histogram for retirement paychecks is actually “lognormal” rather than “normal”, meaning the shape is not symmetrical and the right tail is longer than the left tail.