The Madness of Crowds

There are few better reminders of the limits of intelligence in markets than Isaac Newton. One of history’s greatest scientific minds could explain the motion of planets, yet he was still vulnerable to one of the most powerful forces in investing: crowd behaviour.

Isaac Newton was an extraordinary mathematician who invented calculus to help explain to his astronomer friend Edmund Halley why planets’ orbits are elliptical rather than spherical. There were no IQ tests in the 17th century, but some estimates place his IQ between 170 and 190. At the upper end, that would put him among roughly one in 137 million people. 

Yet this same brilliant individual was drawn into one of the biggest frauds in England’s history. Adding to the irony, Newton invested not once, but twice, in the South Sea Company, and it was the second investment that caused him to lose almost all his savings as shown in the chart below. 

At one point the South Sea Company was worth 420 million pounds, or more than all the land in England at the time. Fraud was pervasive within the South Sea, including efforts by management to lobby the government to pass the “Bubble Act” in 1720, making it more difficult for its competitors—more than 200 of them—to form joint venture stock companies. 

In the end, this was the “Bubble Company” era, and almost all of the companies involved failed within a year.

So, how does this sad story of human behaviour run amok impact us today?  I believe we can find a connection in this quote from Mark Twain: “History doesn’t repeat itself, but it often rhymes.”

The chart above may be familiar to many readers. It shows the CAPE ratio since the 19th century.

The Shiller CAPE ratio is calculated by dividing the current price of a stock (or market index) by its average inflation-adjusted earnings over the previous 10 years. It is typically higher than the current P/E ratio, making it a more conservative valuation measure. It allows us to observe long-term trends in stock prices relative to long-term earnings. 

What’s troubling about the latest Shiller Cape index is that it is close to a record high at more than 40 as of June 2026 and 22% higher than stocks were in October 1929. 

Dividends at record lows 

When one looks at another metric, dividend yields, specifically the yield on the S&P 500, issues arise.

Until now, the lowest dividend yield was recorded for the S&P 500 at 1.1% in March 2000, just as the dotcom bubble burst. That miniscule payout has now dropped even lower, to 1.06% in May 2026. US 10-year Treasuries are yielding almost 4.5%—more than 4 times as much as the yield on stocks.

Why does this matter? When it comes to stability, reliability and predictability, an S&P 500 company’s dividend is four times less volatile than its price.

How do we know this? Our Public Market’s team completed a study of stock prices and dividend payouts on a monthly basis, from 1970 to 2026. Several observations emerge from that analysis.

The range of dividend yields has gone from a current low of 1.06% to a high of 6.37% in July 1982. It is important to remember that 10-year treasuries were paying almost 14% that month.

During this period, there were at least five bear markets. The table below highlights when bear markets occurred, the magnitude of the S&P 500 decline, and the associated dividend yield and cash dividends per unit of the index. The final measure is arguably the most important.   

^{1 Beginning of bear market 2 End of bear market}

What is surprising about these numbers is that even though stock prices fell almost 40% on average during five bear markets, the dividends being paid by those stocks rose by more than 4%. Making a bet on dividend payouts is much more reliable than making a bet on price movements, by a factor of four-to-one.  

Unfortunately, dividend yields have been falling consistently since the Global Financial Crisis, when they last peaked at 3.82%. Since then, they dropped regularly until they reached a new record low of 1.06% in May 2026. One of the culprits for low dividend payouts could be the cumulative impact of seven of the largest technology stocks on the S&P 500 known as the Magnificent Seven: Apple, Microsoft, Tesla, Meta, Nvidia, Amazon and Alphabet.  

The chart below shows how, over 25 years, these companies have grown from less than 5% of the index to 35% today:

Why is that a potential problem for investors? In two ways: For those who invest in passive market-weighted ETFs, it means that more than one-third of their money is tied up in seven names, and two-thirds is spread amongst the other 493. Investors are increasingly making a concentrated bet on a small number of companies.  

The second concern relates to dividends. In the table below, we show the annual current dividend yield of the Magnificent Seven vs the rest of the S&P 500. If we equally weigh the Magnificent Seven, their dividend payout is 0.29% per year versus 1.42% per year—almost 5 times as much—versus the rest of the S&P. And that yield of 1.42% is already near record lows.  

Record-low dividend yields do not necessarily mean that earnings are under pressure or dropping. But they do indicate that the relationship between the price of a stock and its earnings is not moving in sync. As of June 2026, the Magnificent Seven forward PE ratio was just over 50. Their growth has justified higher PE multiples, but over the last decade, that has averaged about 40.  At 50, an investor is paying $50 in price for every $1 in future earnings and receiving a dividend of under 15 cents annually for each $50 invested.

Even if current conditions are not enough to puncture enthusiasm for US equities, the addition of companies such as SpaceX, Anthropic and OpenAI could further amplify speculative behaviour. SpaceX is already a public company with a market Cap of $2.25 trillion, as of June 30, 2026, and Anthropic and Open AI are planning IPOs at the $900 billion mark this year.

Let me be clear: I am a strong supporter of technological innovation and the role it plays in driving productivity and creativity. I do believe that AI will be transformative. That does not automatically justify the prices some investors are willing to pay to own a piece of the “action.” By all the metrics I have looked at, it is hard to make a financial case for any of these three behemoths. Consider:    

• SpaceX is trading at 44x forward sales, 120x trailing-twelve months (TTM), while Nvidia is trading at 20x forward sales, and 20x TTM, (and makes money). SpaceX does not project when it will be profitable. 
• OpenAI does not expect to be profitable until 2029 and has an estimated value of about 40x sales. 
• Anthropic now makes a small profit but will come to market at close to 40x sales. 

My detractors would point out that all three companies have great potential to increase revenue and scale quickly enough to become profitable and justify their current valuations. AI has the potential to be among the most transformative technologies in human history. 

Some might argue that the invention of the steam engine and the creation of railways were also massively transformational, as was the invention of the Internet. Yet each of these technologies went through a very familiar pattern, from embryonic to boom, to bust, to sustained growth.

The six companies highlighted below all survived the dotcom bubble in 2000, and four of them are in the Magnificent Seven today. They are Amazon, Nvidia, Apple, Microsoft, Intel, and Oracle.

See table below for the drop in prices, from peak to trough. 

Let’s consider Cisco and Microsoft in greater detail.

A business can be excellent and still become a large portfolio risk if too much return depends on it. A theme can be real and still become crowded.

• Let’s bring it back to circa 1999.
• Cisco and Microsoft were the darlings in the 1990s.
• Microsoft had a cumulative gain of 8,400% or 57% annualized return.
• Cisco was even better, showing a 66,000% gain or a 93.5% annual compounded return.
• This may sound familiar with Nvidia’s 18,000% cumulative gain over the last 10 years, with a 68% annualized return.

After the dotcom market crashed, it took the high-quality Microsoft name 17 years until December 2016, to get back to its 1999 price. Cisco peaked in March 2000, and it took 25 years to get back to its previous high. 

MSFT was trading at 64x, and CSCO at its peak was over 100x. These were great companies with excessive valuations.

^{_{*As of June 2026}}

These were serious enterprises that, in some cases, had been in business for decades before the dotcom bubble burst. They are now among the leaders in US technology companies. But even they had to fall from grace when investors, in a blind euphoria during the mid-to-late 90s, could not get enough capital into them because they were going to transform the world. And in many ways, these companies did transform the world.

Just try to imagine how slow, backward, and unproductive the world would be today if the Internet did not exist. Companies that survived the dotcom bubble did so through a combination of hard work, money, innovation and luck. Right now, there are a number of players being rewarded for developing good ideas and technology. Those rewards only last as long as they can show that they can take that framework and make it into a revenue generating profitable enterprise. Only some will survive this ultimate test of creating value, and for those who do, it will not likely be at valuations that defy gravity.

The madness of crowds is not new. The discipline required to manage through it is not new either. If you have questions about how today’s market environment may affect your portfolio, speak with a Nicola Wealth Advisor.