The Last Conversation You'll Ever Need To Have About Planning For Retirement Withdrawals

A couple months ago, I read an amazing article titled "The Last Conversation You’ll Ever Need to Have About Eating Right". In it, the author had a simple conversation with an expert on nutrition, wherein they asked the questions that all weigh on our minds and were given straightforward and research-backed answers. Because I think every subject deserves this kind of article, here's a "Last Conversation" piece on planning for retirement withdrawals!

Just tell me: how much money do I need to retire? What's my "number"?

I wish it were that simple. The problem is that there are just too many variables to calculate a simple number that works for everyone. What are your living expenses? What's your anticipated retirement income? Is that income going to start right away, or later (for example, waiting until 70 to take Social Security)? How aggressively are you invested? How efficient is your portfolio in terms of risk-adjusted returns? How much of it is in Roth, Traditional, and non-tax-advantaged accounts? How long-lived is your family? The answers to all those questions can drastically change your "number".

But can't I just assume a 4% withdrawal rate of my portfolio? Multiply my annual expenses by 25, and there's my answer. Isn't that good enough?

You're referring to the 1998 "Trinity Study" which indicated that, for the portfolio they simulated, a 4% withdrawal rate led to a 95% success rate of retirement. Again, there are variables at play here that can drastically change that outcome. On page 2 of Wade Pfau's recent Forbes article on the subject, you can see that the results vary widely depending on your portfolio and how long you need the money to last. Also, it makes assumptions like constant withdrawals and no taxes or fees, which are big assumptions to make.

But why not assume constant withdrawals? I retire, I start taking Social Security, I pull 4% from my portfolio every year, adjusted for tax and fees. I can make a spreadsheet to run those numbers and turn out just fine.

That's not a terrible idea, but there are a couple of problems. For one, your portfolio is going to be going up and down with the market. Do you want your spending to go up and down like that?

OK, fine. I start with 4%, then increase with inflation in up years and don't increase in down years.

That's smarter. But are you really comfortable leaving your lifestyle in retirement -- likely several decades -- up to a rule of thumb? You might run out of money if the market has an ill-timed downturn...or, arguably worse, you might be leaving money on the table if it has a well-timed upturn!

That's true, but while I don't like the idea of running out of money, I don't really care about leaving money on the table, as long as I can live in retirement the way I am now.

In that case, an immediate fixed annuity might not be a bad idea.

I've heard of those. Exactly how do they work again?

There are several kinds of annuities, and they're all similar: an insurance company sells you income for life. With an immediate fixed annuity, it's pretty straightforward: you pay a lump sum for the annuity, and it pays you for the rest of your life.

Sure, that sounds great...so what's the catch?

You're leaving money on the table. Most annuities are expensive, if not in fees, then in terms of opportunity cost -- if you invested the money instead, you could have a higher standard of living, even taking into account the potential for downturns.

OK, I see the tradeoff there. But you said there were other problems with the consistent withdrawal rate, right?

Yes. If you want to optimize your retirement, it's nearly always better to wait until 70 to take Social Security -- the opportunity cost of withdrawing more of your investments earlier is outweighed by the increased Social Security income, assuming you have a decent lifespan (the crossover point is somewhere around 80 for most people).

And if you do that, you'll have a span of time when your effective tax rates will be much lower than otherwise, until you turn 70 and start taking Social Security and Required Minimum Withdrawals from your Traditional IRA's. This provides some interesting opportunities; for example, you can take advantage of that time to do Traditional-IRA-to-Roth conversions, which lowers the total amount of taxes you'll pay in retirement -- which keeps more money in your pocket! (Not to mention that it gives you more flexibility later, since your RMD's will be lower.)

So the problem is that I'm leaving money on the table (again).

Yes -- potentially a lot. I've seen it time and again -- optimizing your taxes in retirement can save you tens or even hundreds of thousands of today's dollars over the lifetime of your portfolio. I don't mind Uncle Sam or the insurance companies that sell annuities, but I'd rather the money stay with folks like you. You've earned your share; why not keep it?

Hey, speaking of taxes -- I've got a Roth IRA, a Traditional IRA, and a brokerage account. Which do I pull from and when?

The best way to answer that is to think in terms of goals. If you want to optimize your retirement, you want to (a) "smooth out" your taxes in potentially-higher-bracket years by taking advantage of lower-bracket years, (b) maximize your flexibility by reducing the amount of your RMD's, and (c) let tax-advantaged accounts grow untaxed for as long as possible.

Therefore, it follows that the optimal strategy is to do partial IRA-to-Roth conversions and Traditional IRA withdrawals in early years to the point where your RMD's won't result in a higher tax bracket in later years, and fund the rest of your living expenses via withdrawals from your brokerage account until it's depleted, at which point you'd switch over to funding your living expenses via Traditional and Roth IRA withdrawals.

Alright, let's say I want to optimize my annual withdrawals in retirement, so I'm not leaving money on the table but I'm also not running out of money. How do I do that?

There are several tools out there that run cash projections combined with Monte Carlo simulations or simulations using historical data. Many of them are tools for financial planners and cost hundreds of dollars a month, but you can do a decent projection with free tools like FIRECalc.

Sure, but what's a Monte Carlo simulation?

If you put on your statistics hat for a moment: you can pretty accurately model the returns of stocks and bonds as a Gaussian distribution. The average return of the S&P 500, for example, has been 9.69% annually with a standard deviation of 15.93% over the past 25 years.

But if you take that sequence of returns and rearrange the order in time, you get vastly different results at the end. As I mentioned earlier, a downturn at the wrong time or an upturn at the right time can make all the difference.

To take that into account, a Monte Carlo simulation randomizes the sequence of returns and calculates the outcome, over and over again -- say 1,000 times. This gives you an idea of the various portfolio outcomes you might experience.

Ah, OK. So I run the Monte Carlo simulation, and if the result is that all 1,000 simulations passed, then I'm good.

Not if you want to optimize. 1,000 simulations passing means that you're leaving money on the table!

Well, maybe, but if any of those simulations fail, then that means I run out of money. Isn't that worse?

It is worse, but that's not what a single failure means. If a single Monte Carlo run fails, that means that in this case, you went on autopilot and didn't make a single change even as your portfolio went to zero. You wouldn't actually do that. If you're running projections every year, you'd see the failure case coming 10+ years out, and you could make a small course correction to stay on track.

How do I see the failure case coming?

If, for example, you experienced a sequence of bad returns in real life, it would lower the success rate significantly -- instead of one simulation failing, you'd start to see hundreds of failures.

So I'm looking to keep the Monte Carlo simulations in a "Goldilocks range" -- not too high, not too low. I assume you have a range in mind?

80%-90% is a good range for many people in retirement. Above that, and you're likely leaving money on the table. Below that, and it's time to take action to bring that number up.

What, you mean retirement is like real life? Plan for the next few years, don't go on autopilot, make changes as necessary? Shocker!

Amazing, isn't it? If you've got any more questions, don't hesitate to leave a comment or send an e-mail.