top of page

Sustainable, optimized portfolio income and Monte Carlo

OK, full disclosure: I'm a little twitchy around the phrase "passive income". It's like a slightly more socially acceptable form of "get rich quick", but with the same underlying idea: if you just find and implement the right scheme, you can be set for life, money just rolling in every month, without raising a finger ever again.

Let's set aside for a moment whether that's actually the happy ending that you might think it is, and focus on this: the term is misleading at best, and dangerous at worst. I see it most often applied to operating a rental real estate property, which can indeed be lucrative...but the primary reason (beyond luck) for this is because most real estate is leveraged. While this is not in and of itself necessarily a bad thing, if you're not careful, it can seriously bite you. Just ask popular personal finance author Dave Ramsey, who learned that lesson the hard way.

I also see it applied to portfolio construction, when people focus primarily on the income a portfolio generates, through interest and dividends, because they want the "passive income". Again, this is suboptimal at best, and dangerous at worst. I want to talk about that, and a better way to think about generating income from your portfolio.

The problems with "income-focused" portfolios

OK, so let's say you want to maximize the amount of income generated by your portfolio. It's actually pretty straightforward: you "reach for yield", finding the securities and asset classes that are generating income that is the highest percentage of their purchase price. There are a couple of problems with this, however, and while I've talked about this before, I'm happy to give the summary again here.

The first is that high-yield portfolios are generally tax-inefficient. Interest is taxed as ordinary income, unless it comes from municipal bonds - but those are lower yield than comparable bonds. Dividends, while sometimes taxed as capital gains, are often also taxed as ordinary income. And real estate income (including real estate investment trust income) is also taxed at ordinary income tax rates. Generally speaking, the difference between your long-term capital gains tax rate and your ordinary income tax rate is 10-15%, which can take a sizeable chunk out of your portfolio over time.

Another -- arguably more important -- problem is that public markets are extremely efficient: if there's a high yield, there's always a good reason why, and that reason is risk. (Another way to think about it: if income divided by price is high, it may not be because the income is high, but because the value is low!) Long-term bonds have higher yields because they have high interest rate risk: if rates rise, their value drops precipitously. High-yield corporate bonds have higher yields because they have high credit risk: their value fluctuates -- or even goes away entirely -- based on the probability that the company will actually be able to pay off the bond. And high-dividend stocks often become high-dividend stocks not because the company has increased the dividends out of the kindness of their hearts, but because the value of each share of the company has decreased, again due to some risk regarding that company's future.

By ignoring the income an asset generates, and instead focusing on its tax efficiency, expected returns, expected volatility, and correlations with other asset classes, you can better optimize your portfolio for the highest amount of return for the least amount of portfolio volatility.

You may say, "Well, I don't really care if the value of the portfolio changes, as long as I'm getting constant income out of it; I'm not planning on touching the principal anyway." I want to be clear: you won't get constant income out of your portfolio. Dividend payments change. Interest rates change. Companies go out of business. Even if you don't touch a red cent of your principal, the amount of income your portfolio generates will be variable.

But now I want to go back to that "I don't want to touch the principal" mentality. Why exactly is that? Is it because you want to leave something to your children? That's all well and good, but if you live a good long life, your children's inheritance will come to them after they're already retired themselves -- and isn't it a bit late then?

I certainly understand if you want to make absolutely sure that you don't run out of money...but there is a lot of space between "don't touch the principal" and "run out of money in ten years". If your income is going to be variable anyway -- see two paragraphs up -- then why not create a system for pulling a sustainable amount from your portfolio each year? Something that provides a good balance between the risk of dying poor and the risk of dying rich?

"Sure", you say, "but how do I know what 'sustainable' is?" A good question -- let's talk about that!

Safe withdrawal rates

Way back in 1994, a fellow by the name of Bill Bengen decided he wanted to know the answer to that question. Specifically, he wanted to know this: if you withdrew a certain percentage from your portfolio in the first year of retirement, and then increased that withdrawal by inflation over the entire course of retirement, what is the minimum possible percentage for that starting withdrawal rate that would work for any given historical period? He made a bunch of assumptions -- all assets held in tax-deferred accounts, a 30-year time horizon, a 50/50 stock/bond split, etc. -- and came up with an answer: 4.15% Another study in 1998, the "Trinity study", modified some of his assumptions and asserted that this value was 4%.

In either case, the basic idea was this: if you started off with a reasonably-diversified portfolio, withdrew 4% (or 4.15%) in the first year, and then increased your withdrawals by inflation every year after that, there is no point in recent history when you would ever have run out of money.

And that's not a bad place to start.

Flexible withdrawal strategies

But what if you were actually OK with occasionally lowering your standard of living once in a while? You know, like you do when you get laid off, or decide to work a lower paying but better job, or decide to move to a city with a higher cost of living, or have a child, or any one of a number of things you've probably already done? In retirement, that might look like travelling less, or maybe even taking on a part-time job doing contract work...which plenty of folks in tech have done simply because they want to keep putting their brains to use!

In that case, you can start with a higher initial withdrawal rate -- say 5% or more -- and implement a flexible withdrawal strategy that determines your spending for any given year. One of the most popular strategies is the Guyton-Klinger "guardrail" strategy, which uses your withdrawal rate ratio -- your current withdrawal rate divided by the initial withdrawal rate -- to determine when you might want to cut back on spending (or increase it).

And that's even better. If you're like a lot of the folks who read this blog, you likely have the flexibility to cut back on certain expenses or make more income if you want to, which means you can start off with a higher withdrawal rate, and thus retire earlier, and/or retire to a higher standard of living.

But let's take it a step further, shall we?

Monte Carlo-based withdrawal rates

You've likely heard me talk about Monte Carlo simulations before, but a quick recap: the idea is to run many, many simulations of what the market might do (based on historical data, or economic projections, or both), and then apply it to your personal financial projections to come up with a "success rate".

Now, "success rate" is another term that makes me twitchy, because it implies that e.g. a 90% success rate means a 10% chance of having to go live in a van by the river. A better term would be "probability of no forced change": a 90% success rate means you could put on a blindfold and proceed without ever changing anything in response to what the market does, no matter how awful, and you'd be fine.

We use this in financial planning quite frequently -- for example, it lets us make an apples-to-apples comparison between two different portfolios, one with a higher return but higher volatility, and one with lower return but lower volatility. Monte Carlo takes both into account, and thus lets us figure out how much bang we get for the buck, as it were.

As you can imagine, it's not that hard to come up with a withdrawal strategy based on Monte Carlo success rates. Simply set a range -- for example, 80-90%. If your success rate is within that range, keep on the way you are. If it's higher, spend more. If it's lower, spend less.

One of the more interesting advantages to this approach is that you get to set the target. As researcher Derek Tharp has pointed out, the target success rate doesn't actually change the withdrawal variability much; rather, the lower the target, the higher the initial withdrawal rate, but also the higher probability that you'll end up decreasing your spend sometime over the course of the years. Got more flexibility, and want to spend more or retire earlier? Set a lower target!

(Interestingly, Morningstar interviewed him the day I finished this article, and they talk about "risk-based guardrails" in the podcast. It's worth a listen!)

Also, this approach works well if, like nearly all of Seaborn's clients, your net income will change significantly during retirement -- maybe you want to retire before you're eligible for Social Security, or go to half-time now but work until 75, or pop in and out of the workforce until your nursing home days. Whatever the scenario, you can update your projections accordingly, see how it affects the Monte Carlo simulations, and go from there!

But...I just want to know what the best withdrawal strategy is!

As you can see, there's not a single best strategy for everyone, for the same reason that there's no "best portfolio" or "magic number" for everyone -- it all depends on you! How comfortable are you with decreasing your spending on occasion? How badly do you want to retire early? How likely is it that you'll end up generating income in retirement -- simply because you enjoy the work you're doing?

So while I hope this article is informative, it really has a secret purpose: to get you asking these questions, and the questions behind those questions. What do you want? What are you comfortable with? Do you even know how much of your spending you could cut, if you wanted to? And what could you do, if you knew that?

Answers are important, but sometimes the questions are even more so!

Britton is an engineer-turned-financial-planner in Austin, Texas. As such, he shies away from suits and commissions, and instead tends towards blue jeans, data-driven analysis, and a fee-only approach to financial planning.


bottom of page