Many of us have run through retirement calculators. We punch in our current investment account balances, how much we’ll save each year, what rate of return we expect, and it spits out a projected future value. That projection will be wrong.

Even if we could predict the future rate of return accurately (which we can’t), it would still be inaccurate. Why? The sequence of returns cannot be predicted. The sequence the returns come matters, A LOT.

Note: If you are not adding to your portfolio or taking distributions from your portfolio, the sequence of returns will not have much of an impact. Your returns will fall more closely in line with the time weighted rate of return, assuming good investor behavior…..

How Returns Actually Happen

Let’s imagine the year is 1999 and you are mapping out the next 20 years of your projected savings and future balances. With a proprietary market forecasting formula you’ve developed, you determine that the market will earn 6.01% over the next 20 years.  You already have $1,000 saved and plan to save exactly $15,000 annually. Based off your savings rates and your return projections you will have $555,590.59 at the end of that 20 year period. You decide you don’t want to be part of the emotional roller coaster of markets though, so you won’t look at your balance again for 20 years.

Fast forward 20 years, you do the math and determine you were exactly right on your investment returns and savings projections! You saved exactly $15,000 each and every year and the market return over that period was 6.01% annually. You are very proud of yourself and excited, but you are quite pleasantly surprised to look at your account balance and notice there is over $700,000, as opposed to the $555,590.59 you planned. How can this happen? You saved exactly as planned, markets performed exactly as planned, but you have nearly 30% more than planned. Did your investment firm send you someone else’s statement?

No. The problem is (or in this case the benefit), the market didn’t earn 6.01% each and every single year, it actually had volatility. Since you were making contributions the volatility could have helped or hurt you, let’s look:

The money weighted return can differ greatly at sometimes from the time weighted rate of return, as the money weighted return (also referred to as internal rate of return) adjusts for cash flows in and out of your investments.

This isn’t the norm, there is no norm, other than nothing will go as projected. Let’s run another scenario, but this time you are planning your retirement and you have accumulated $1 million at the start of the year 2000. You are trying to figure out how much you can withdrawal, so you once again need to forecast future market returns. Given your high level of investment experience and not being bothered by market volatility, you stick with a strategy that is 100% invested in the US stock market.

You determine the market should earn about 5.79% the next 20 years, and since you don’t mind spending a little bit of principal you factor your annual withdrawal to be 6% of your initial portfolio value ($60,000 annually, $5,000/monthly). Also, since you are taking a high initial withdrawal rate you decide never to adjust your withdrawal for inflation.

Let’s see where you are by the end of September 2019.

Uh Oh! You didn’t make it to September of 2019. You actually ran out of money in March of 2018. You can see the first 3 years were catastrophic to your plan as by the end of 2002 your balance was more than cut in half. Even though the market recovered, you never recovered.

Now maybe it is a bit reckless for a retiree to be so concentrated in one segment of the investable universe, like the US stock market. Had they just take their $1 million portfolio and split it into thirds between US stocks, non-US stocks, and bonds then they’d still have about $398k at the end of September 2019.

Monte Carlo Simulation

A better way to test your chances of success is to use a monte carlo simulation tool. You will set a projected return over the future time horizon and the software will run 1,000 random trials with different return sequences with different average returns than projected, then tell you a success rate based off those trials. You can see an example of this below, of the 1,000 random trials ran for this client 9 of them had an average annual return between 5.60% and 5.70% compared to the planned return of 7.30%. Of these 9 trials, 4 were successful and 5 were not.

You can see there is a wide range of outcomes even when using the same assumed returns.