Dear Valued Clients and Friends,
In this week’s Dividend Cafe, we are going to cover one of the most important dynamics in all of investing – the reality of range-bound markets. A “flattish” market can be a description of the whole S&P 500, or it can be used to describe a given stock or sector. But no matter what the description is being applied to, one thing is constantly true … The description only works in hindsight – for when one is living in it, they rarely know it. Our subject today will be why that is, and what it means to the moment we are in now.
The lessons of history. The realities of investing. The approach that investors need. All of this and more in today’s Dividend Cafe. Let’s jump in …
| Subscribe on |
Starting with the Easy Part
The S&P 500 was “down” from 2000-2009, though the -25% drop in price from the starting point to the ending point ignores dividends, which is utterly dishonest. With dividends reinvested, the return gets much closer to flat (not exactly, but I will round up from -0.95% per year and just pretend it was flat even though it was actually down -10% over the decade, depending on your exact start and end date).
Now, if we wake up at the end of that decade and say, hey, my S&P 500 fund is basically where it was when the decade started (or for that matter, do so now, as we are actually doing), what we have said is correct as a matter of math and calculated investment return. But for 929 calendar days, from March 24, 2000 until October 9, 2002, the S&P was down -49.1%. And from October 2002 until October 2007, the S&P 500 was up +101%. And then from the beginning of the subprime trauma until the market bottom of March 2009, it was down a stunning -56.8%. I will spare you all the spurts and sparts and zigs and zags that absolutely happened with all three of those periods (and I assure you there were plenty), as the decade is really quite simple to describe mathematically: The market dropped -50%, then went up +100%, then dropped -50%. Voila – no return for a decade. Very simple.
Does a “down -50%” market (let alone two of them) sound like a “flat” market to you? Does an “up +100% market” sound “flat”? In any practical sense the only way to think of the first decade of this new century as “flattish” is to draw a line between the start point and the end point … But alas, that IS the point, isn’t it? THAT line from start to end WAS flat. But no one on that line’s journey would have felt like it.
I understand why this particular decade and market index may just feel too extreme to people. I mean, ten years of the most widely held market index in the world seems like fair representative sample, but let’s see what else history may offer us.
The Famous One
Okay. This is more like it. Here you see the Dow right around 1,000 in 1966 just as the nation was about to get through its more embarrassing era, and ending right as Ronald Reagan was about to come win the cold war and cut marginal taxes. Now, that 1973-74 sell-off was doozy (-45% bear market), but with a lot of zigs up and a lot of zigs down the end-to-end result was “flattish.” Nothing about it felt flat in the moment, but the “point A to point B” line was just that.
And it happened with a volatile period (up and down and up and down and up and down 10-20% a few times from 1966-1972, a new high being made, then that big drop (1973-74), then a big rally (+38% in 1975 and Gerald Ford still lost to Jimmy Carter in 1976!!!), then a chop-chop-chop few years to bring us to 1981 when a new secular bull market would finally start.
A Note from Our Sponsor
I suppose I could turn this into a plug for how dividends become very important during flat, choppy, range-bound markets, and it would all be true (including for reasons not commonly understood or captured in the headline math). It is true that the “flat” Dow of 1966-1981 actually was up +3.94% per year in that period with dividends (a negative price return, plus dividends, equaled a positive total return, as dividends amounted to over 100% of the total return). Indeed, my own Profit from the Profit book made the point (on p. 57) that there are decades where dividends are all one has (see 1970’s, 2000’s) …
But no, that very important point, is not the point I am here to make today.
Some Math
The S&P 500 has averaged +10% for the last one hundred years. It has also averaged +10% for the last thirty years (despite having the aforementioned negative decade in the midst of that). It has averaged +15% for the last fifteen years (which is how you get +10% over thirty years when there was a negative decade in the middle of it). But it has averaged +16% over the last seven years, and +22.9% per year for the last three years. This leaves us with a grand total of two options:
- The long-term historical average of the S&P 500 (+10-11%) is going to become higher, OR
- There has to be something negative to revert the +22.9% and +16% we have seen in recent years towards the mean of +10-11%.
I am not making a market call. I am not predicting what or when. I am doing math. The only way +23% becomes +11% is if there are some lower than +10% numbers in there to “bring it down.” Again, math. How does +23 and X get to 10, 11, or 12? Well, X either does a big negative to get you there quickly, or, X does a slow, moderated pace of something not negative, but below the mean. History has precedents for both. And I am merely trying to explain the math of how average returns are generated, and how history has often “gotten us to the mean” …
Same Principle with Individual Stocks
Let’s start with a question: What public company is the poster-child of the AI moment – the name you most think of when it comes to success in the years since ChatGPT launched? For almost anyone it would be Nvidia (ticker: NVDA). And, for good reason! The stock has skyrocketed since late 2021 when ChatGPT was announced to the world. Their chips power much of the AI world and they sit in the most enviable position of being the primary supplier (and, ummmm, financier) to the AI world. No one can complain about Nvidia results since the AI story became investible, and their revenue growth and market share and margins and all of that and more are, well, Cisco-1990’s level of amazing!
So then, with companies like Anthropic and OpenAI getting ready to go public, and a company like SpaceX having already gone public, and with the largest level of capex announced this year in world history (much of it in the form of orders for Nvidia to fill), and with AI being the predominant investment, markets, economic, technology, and news story on the planet, in a year where Nvidia’s year-over-year revenue growth is going to be somewhere around +100%, what do you think the stock has done over this booming year?
Does it feel like it? I am the first person to say no. And when the stock has rallies like it had in April or in August, the math is almost hard to believe. But that is because the math is always hard to believe in these end-to-end flat periods, because they are not flat while you are in them – they are up and down, zigging and zagging – but the result is all the same. If you went to bed in November 2025 and took a nap for a year (which sometimes I feel like some people do a lot) you would have woken up at the same place you started.
I am not picking on Nvidia. They are in a much better position than CoreWeave, who is down -60% over the last year, and down -70% since June of 2025, despite +112% revenue growth. Palantir’s revenue is also up +100% year-over-year. Their stock is down -7% from its late 2025 high (despite recovering +75% in the last 3 or 4 months alone).
I did not pick “Pets.com” equivalents to the AI story to straw man my point. Nvidia, CoreWeave, and Palantir are absolutely defining companies in the AI revolution – leaders, success stories, and examples of true technological brilliance and innovation. And they are flat, or down big, or down a little, with a ton of zigs and zags along the way, over the last year. These stories are systemic, and I believe will be systemic for years to come.
What Could Slow Down This AI Train?
Well, first of all, it is hardly hypothetical, as we have seen with many parts of the story. But It is not longer a 5% or 10% chance that the AI capex story gets called into question. You may not believe it is the base case yet (I do), but it has become a statistically real scenario that the narrative around hyper-scaler expenditures will change – that the cost of capital has increased, that data center builds at a level to make this whole thing go are just too prohibitive (politically, financially, and logistically), and that the expected returns on invested capital do not warrant the continuation of this capex boom at the levels we have seen or that some have anticipated. The whole story can remain robust, but any moderation of its robustness slows the train, perhaps substantially.
I already wrote a Dividend Cafe laying out my belief that, regardless of timing, it will be capital markets that ultimately “slow the train.” We are only getting teasers of it so far, but things like Meta’s credit default swaps blowing out Thursday essentially catalyzed a -700 point (-2.6%) drop in the Nasdaq Thursday (I recognize it’s usually not mono-causal, but this was at the very least primary). One day is not my point. Whether it’s already begun or remains pending – whether it happens suddenly or slowly – some form of capital constraint either in access to capital or more burdensome cost of capital strikes me as the most logical catalyst to pending challenges in this extraordinary moment.
The spread between return on invested capital and cost of capital, always and forever, raises its ugly (or beautiful) head – eventually. Never forget it.
For Those Who Believe Something Very Bad is Impossible
One of the hardest parts of being a skeptic about the mania of the last couple of years in the so-called AI space is that skepticism around the investment merits and sustainability of a particular thing are wrongly interpreted as either:
- Doubt in the efficacy and relevance of the technology, itself
- Predictions of some sort of March 2000 Nasdaq full-blown implosion
Now, it is entirely true that I do not know that the AI story will end up doing all it has been talked about doing, but since some of the stuff its biggest proponents talk about it doing are destroying humanity and starting nuclear war, maybe that is a good thing? But for all of my investment analysis I have been entirely accommodating of the most bullish and generous forecasts of the transformative capabilities of AI technology.
It is also entirely true that I do not know a March 2000 Nasdaq implosion repeat is off the table. I don’t think that specific set-up being totally repeated is very likely, simply because my study of history tells me that “it rarely repeats but often rhymes,” and so all of the most particular violent and despairing parts of what investors went through in that dynamic don’t seem very likely to be totally replicated. But who knows?
However, what I would say to those who either believe that we have EITHER years of bliss ahead with double-digit compounding of magical new normal returns, OR the apocalypse 2000 bust scenario – there is, in fact, another scenario on the table that does not fit into either bucket. And it is the scenario I think is BEST CASE, and NOT VERY GOOD. The one I have been writing about here today.
This Time it’s Different
I do not believe AI is going to be transformative because the big tech tycoons tell me it will be. In many ways, I believe it despite their saying so. The LLM world has become the thing, and the potential it offers technologically, and hopefully commercially, is tremendous. But of course, this comes after we were promised the metaverse was the thing. And web 3.0. And crypto. And non-fungible tokens.
If you hear those things mentioned now it is usually in the context of someone’s arrest.
I talk a lot about the 1994-1997 internet era in which, all at once, the world was changed forever, a whole bunch of people predicted it was being changed forever, and not a single person you can find in recorded print accurately knew or understood how. The business models, the revenue paths, the particular companies who would survive and thrive – all of that was a mystery, even as the “high level” forecast was exactly right (unless you were Paul Krugman). AI is very obviously not the metaverse or NFT’s. But the fallibility of fallible humans to understand how it will all play out is still present.
Is the media any help here? When it comes to zigs and zags no one has this better than the media. Consider what Ben Carlson pointed out this week …
*A Wealth of Common Sense, Oct. 20, 2026
And when this many trillions of dollars are on the line unknowability becomes a great facilitator of “flattish, range-bound, choppy” markets. The inability of all of us to know how the regulatory environment will develop is a reality. Where China fits in, or American politics, or power demand, or interest rates, or any number of other factors all remain not only unknown but unknowable.
The “Muted Return” Scenario so Many Face
In conclusion, I want to suggest that the dual reality of unknowability and excessive valuations make the possibility of a perfectly satisfactory company and sector result being accompanied by a very unsatisfactory stock market result very, very high. This scenario is not rooted to doom and gloom forecasts, some of which could well materialize. This is not based on a bubble burst, or regulatory implosion, or unexpected intervention from Washington DC, or shocking failure of a key company in the ecosystem, or competitive threat from China, or technological development that destroys the revenue model – though I would not rule a single one of those things out – this is based on all going well. In benign circumstances, reality often provides “flattish” investment results despite monumental risk (possibility of bad outcome) and volatility (up and down movement) along the way.
And that can (and does) happen, all the time. The rallies that happen within these periods desensitize us to the math, and even distort our perception of what is taking place. It is completely understandable.
My proposition is that dividend growth provides you a benefit to such choppy volatility, an enhanced return in a range-bound market, the peace of mind of sustained, durable, and growing income if you are income-oriented in your financial objectives, and one other thing, too … Much less sensitivity to the left tail risk of the dynamic I have spent this Dividend Cafe describing.
To that end we work.
Chart of the Week
This chart elegantly captures the most amazing story of 2026. Companies with unprecedented levels of operating cash flows for years and years have suddenly seen free cash flow go negative as capex levels grew even faster than operating earnings.
Quote of the Week
“I have never met a hater who was more successful than the person they were hating on.”
~ Chris Rock
More to Chew on
With bond yields dropping in October as quickly as they rose in September, I do want to make sure any who missed this oldie but goodie (two whole weeks old, to be exact) have the chance to re-visit it. I have a feeling it is going to age well, but of course you never know.
Make it a wonderful weekend, my friends. Markets can sometimes be range-bound, but a life well-lived, never is.
With regards,
David L. Bahnsen
Managing Partner, CIO
This week’s Dividend Cafe features research from S&P, Baird, Barclays, Goldman Sachs, and the IRN research platform of FactSet
