Bubble fears have been looming over the stock market for months, driven largely by the artificial intelligence buildout and concerns that the leading AI companies are now planning to go public, with frontrunner Anthropic PBC filing last week for an IPO that could top US$2 trillion. New York-based economic researcher Charles Lister Smith has been tracking historic bubble patterns and says he has the formula for calculating when markets are in bubble territory, a theory he has detailed in a recent research paper. The Financial Post spoke to him about his calculations, what’s happening in the market today and why he thinks investors should be worried.
Financial Post: What got you interested in bubbles?
Charles Lister Smith: My career was largely dictated by events. I graduated college in 2002 and was lucky to get a job when the Financial Crisis hit. And then I got my MBA in 2009, right when that crash hit.
I know what it’s like to have your life wrecked every 10 years. My first two years at work I saw four rounds of layoff at Merrill Lynch (now part of Bank of America Corp.). It was a terrifying thing.
When I was running my hedge fund, I was looking at rates being near zero and wondering how we got into this mess in the late 2010s, and that’s when I learned that our monetary policy, in large degree, is built off an interpretation of what happened in the 1930s.
FP: How did looking at historical episodes of bubbles shape your views?
CLS: What I figured out was that what happened in the final movement of the Great Crash (of 1929) was an overhang in the IPO market. That was the first step for me, going, “Oh, wait a minute! I’ve seen this pattern before.”
The response of the U.S. government to the Great Crash was the rollout of the 1933 Securities Act, because FDR wanted to stop the unnecessary manufacturing of securities. So, all this legislation ends up being about controlling the new issues market. That’s part of the story here.
From ’70 to ’73, you have the largest IPO boom since 1929 and the markets peak in January of 1973. Then you have this huge cliff. In March of ’74, what you have is a dramatic drop that tracks the rise in the long-term rate.
I framed a bubble as a 40 per cent drop in equity prices. You always see this 40 per cent equity drop, no matter where the area of overinvestment is.
FP: What makes your paper, Architecture of a Bubble, different from existing literature?
CLS: Most of the work on bubbles has really just focused on the crash element and the panic element. They’re missing the fact that the real problem is starting earlier.
Once you define the bubble as 40 per cent (drop in equity prices), the only real literature that’s out there that’s useful at all is the Kindleberger-Minsky model (which describes a cycle of mania, panic and crash) and the Shiller model that focuses on price-to-earnings (PE) multiples. Because PE multiples aren’t predictive in fixed income bubbles and don’t tell you when to sell, they’re not really that informative.
That leaves you with this Kindleberger-Minsky model.
The first stage is called displacement, and that causes prices to arbitrarily move upwards. When stock prices go up suddenly, this is your classic boom phase, where you have some expansion and money supply that feeds stock price gains. That’s when people are rotating their savings into the market to catch price gains. Things start to get a little crazy and you have this phase in the model that’s called euphoria, which I redefined to euphoria in the new issues market.
From euphoria, the next phase in the original model is profit taking, which is the stage before panic, and I redefined that as market congestion, when bankers or underwriters bring out new stocks once stock prices are rising. You have this huge overshoot every time, and, importantly, an exhausted marketplace.
I have expressed the Kindleberger-Minsky model in an equation that effectively turns a static framework into a dynamic model. In simple terms, the formula compares what the market can sustain with the price you get once new supply arrives. The gap between the two is the decline.
Once you apply this five-step model to all these crises, the glaring takeaway is this huge overhang of securities, and that’s what the bulk of (my paper on the architecture of a bubble) set out to show.
My other big finding from the paper involves the long-term rate and the routine rise in market yields. If all these people are issuing capital at the same time, prices are going to go up. The spend from the issues (companies’ issuing new shares) is providing the boost to the economy and offsets the drag from the higher rates.
As folks sell out and prices come down, that encourages other people to sell other things, and so there’s a big crowding out that happens because all of these prices are in some respect built on the amount of supply that’s in hand.
What I’ve basically done with the equation in the paper (see an explanation of the equation below) is to show that double bump, smackdown supply sort of sequence. It’s meant to show how you create these declines. We can go back into the declines of the Great Crash (of 1929). We can recreate it and that means that we can say whether something is in bubble territory or not.
FP: What’s happening in markets right now?
CLS: You go from euphoric in the new issues market to market congestion, and that’s what we’re experiencing right now. You can see it in the marketplace today.
The market top is past us and it’s because all that supply that we know is coming on in less than 30 days.
The Anthropic IPO will be brought to market, unusually, in a falling price environment. The size of Anthropic’s IPO will be larger than I believe all the IPOs in most years on record. That supply is going to force turnover in the market.
Assuming the IPO of Anthropic is fully executed, the equation implies a decline of approximately 44 per cent in U.S. stock prices.
I think we have a situation where we’re going to have prices declining … at a time when the needs of the government are still increasing, and that’s going to cause the long-term (Treasury) rate to go up. As that long-term rate goes, it just starts pulling money from the stock market.
And the only thing that really changes this whole equation is what the Fed does. This (bubble) is not like the last two. … It’s way closer to what we saw in ’73, ’74, which means that we’re going to have an elevated period of rates that is going to cause a creeping crash. It’s going to come on slowly, and what ends up moving the needle there is the Treasury auctions. Because as they have to pull more money from the market, the rates go higher.
FP: What should investors do?
CLS: What we’re trying to overcome here is this mantra that’s pretty well known about buying the dip. We’ve never seen a huge outpouring of securities that hasn’t turned out badly.
Canadian investors would be well (advised) to rotate their stock portfolios into short-term government securities with durations between three and 12 months. This includes U.S. Treasury Bills or, more preferably, Canada Treasury Bills that are unencumbered by the same deficit risks presented in the U.S.
If you were to buy a 10-year note right now, and rates were to keep on going up, you would actually lose value in that because you wouldn’t be able to sell it to anyone at the same price because yields have changed. So that means that you’d have to hold that note for 10 years if you wanted to get that same price back. That’s why you don’t want to buy anything with a really long duration.
The safest thing is a three-month security … and to hold that until stock prices start to decline more. And what will happen at some point, is the Fed will say we’re not lifting rates. Once that happens, gold’s going to go through the moon. So, at some point you want to buy gold.
What folks are going to quickly see is that prices are going to slowly drift down. I’m not predicting a “crash” crash. I’m saying this is the market top. What happens beyond that initial phase, will depend upon how people react, how the government reacts.
Responses have been edited for clarity and length.
How Smith’s formula works
Smith reformulated the Kindleberger-Minsky model and expressed it as an equation, explained below:
Crash = [P_base × (1 + F)] / [P_peak × (1 + M × S)] − 1
P_base (20.6x) is the price-to-earnings multiple of the S&P 500 at the outset of the boom in January 2023. F (0.9 per cent) is the increase in investable capital, measured as the growth in the money (in the stock market) over the first two years of the boom, which in this case rose from US$21.3 trillion in December 2022 to US$21.5 trillion in December 2024.
P_peak (26.0x) is the multiple at the top of the market, from which the decline is measured, as of October 14, 2026.
S (7.1 per cent) is the overhang of new securities, measured as new supply against the market’s value: US$3.5 trillion of equity from the coming IPOs, led by Anthropic, plus the US$1.9 trillion deficit projected for fiscal 2027, (which indicates the crowding out amid budget deficits) against the Russell 3000’s US$75.6 trillion market value. (The Russell 3000 tracks the 3,000 largest publicly traded companies in the U.S.)
Editor’s Picks
M (5.9x) is the price impact of that supply. Economists Xavier Gabaix and Ralph Koijen estimate that each dollar moving into or out of the stock market changes its value by roughly US$4.70 to US$5.90. Smith said he uses the upper bound because panics mark the absence of the marginal buyer (who is willing to pay the maximum price) and outsized volumes commonly accompany steeper price drops.
He added that the figure is not intended for precise measurement but rather as an indicator of a speculative bubble, so a decline of more than 40 per cent is perhaps the better way to frame it.
Serah Louis is a Toronto-based reporter at the Financial Post who writes wealth-building stories for young people, explainers on complex financial topics and the latest news on the Canadian economy.
