Podpower Episode AtlasIn this episode of Monetary Matters, host Jack Farley interviews veteran macro trader Andy Constan, who posits that the current market, particularly in AI and technology, is experiencing a bubble. However, Constan distinguishes this from typical price bubbles, asserting that the 'E' (earnings expectations) rather than the 'P' (price-to-earnings ratio) is the true indicator of unsustainability. He draws parallels to historical bubbles he's witnessed throughout his career, including the dot-com era and the 2008 housing crisis, highlighting common underlying conditions such as new technologies, policy shifts, and parabolic price movements.
Constan explains that while valuations (P/E ratios) might not appear extreme due to soaring earnings, the projected earnings growth for AI companies and the S&P 500 as a whole far outstrips the available economic 'pie' (GDP growth). He points out that consensus estimates suggest an outsized portion of future S&P 500 earnings growth is expected to accrue to AI companies, a scenario he deems improbable without a significant rebalancing or a massive explosion in GDP. This imbalance, he argues, creates unsustainable expectations across the market.
The conversation also touches on the K-shaped economy, the role of policymakers in managing inflation, and the concentrated nature of AI investments among a few key players (semiconductors, frontier models, hyperscalers). Constan advises against shorting a bubble due to its unpredictable duration but suggests strategies like disciplined profit-taking and using options to manage risk in concentrated portfolios. He emphasizes that while he believes we are in a bubble, timing its collapse is notoriously difficult, and the eventual correction could be far more severe than a typical bear market.
Andy Constan's central thesis that the current market bubble is not primarily in elevated price-to-earnings (P/E) ratios, but rather in unsustainable and unrealistic earnings growth expectations (the 'E') for companies, especially in the AI sector.
Why careThis concept challenges the traditional view of bubbles and suggests that even if P/E ratios seem reasonable, the underlying earnings projections are fundamentally flawed and cannot be sustained by the broader economy.
The idea that the total Gross Domestic Product (GDP) represents the ultimate 'pie' from which all corporate earnings must be derived. Constan argues that current earnings expectations for AI and S&P 500 companies exceed what can realistically be supported by projected GDP growth.
Why careIt highlights the zero-sum nature of corporate earnings within a finite economic system, implying that if some companies gain an outsized share, others must lose, or the overall growth projections are simply too high.
Constan describes the AI investment landscape as a 'concentrated bet' among three major cohorts: semiconductor manufacturers, frontier model developers (like OpenAI), and hyperscalers (cloud providers like Amazon, Microsoft, Google). These entities are heavily investing in each other, creating a self-reinforcing cycle.
Why careThis concentration suggests a potential fragility, as the success of the entire ecosystem relies heavily on the ROI delivered by this interconnected group, and a failure in one area could cascade through the others.
The discussion draws a parallel between the K-shaped economic recovery (where the wealthy thrive while others struggle) and the corporate landscape, where a few dominant tech companies capture an increasingly large share of earnings, leaving less for the rest of the market.
Why careIt illustrates how market dynamics can mirror broader societal inequalities, with a small number of powerful entities disproportionately benefiting, potentially at the expense of smaller businesses or other sectors.
A scenario where aggressive actions by policymakers (e.g., significant budget deficit reduction, spending cuts, or central bank tightening) intentionally slow the economy to combat inflation or stabilize markets.
Why careConstan suggests this type of intervention is necessary for a 'garden variety' recession, but notes that current policymakers show little inclination to act aggressively, especially before elections, which could prolong unsustainable market conditions.
"If we're in a bubble, why aren't we suing crazy pees? It's not the P that's the bubble, it's the E."
"I'm not George Soros. I don't think any of us that are listening on this are anything like George Soros, nor have any of the type of information that George Soros has or had."
"The problem is there's not enough pie for the rest of the economy to eat in that circumstance."
"The GDP is the only thing available for companies to earn. And they can take more share. That's possible. Currently, they're running at 12% share. That's up from 4% years ago. So, they're already taking more share."
"You don't short a bubble. What do you think about collaring it?"