The Bubble You Can’t Exit | Dan Rasmussen on the Private Equity Trap

Episode Summary

Executive Summary: The discussion argues that private equity is entering a reckoning as weak performance, high fees, aggressive leverage, and private credit exposure expose fragile businesses, while AI may worsen risks for highly valued tech and subscale software holdings. The conversation then shifts to biotech, where Dan explains how Verdad found traditional quant models fail and rebuilt its framework around spending, specialist ownership, and peer momentum to better capture returns and risk.

Main Topics: Private equity underperformance and structural fragility: Dan argues the industry is finally facing the consequences of overpaying for low-margin, subscale businesses and relying on excessive leverage and private credit, with bankruptcies emerging first in the small end of the market. Institutional persistence despite weak results: LPs and allocators are said to be reducing exposure only when forced by capital calls or staffing constraints, while continuing to rationalize allocations by focusing on favored managers rather than the asset class in aggregate. Retailization of private assets via 401(k)s and interval funds: The guests are skeptical that bringing private equity into retirement accounts is truly democratizing; they warn about hidden fees, gating, liquidity mismatches, and likely lawsuits. AI as both innovation catalyst and bubble risk: The conversation frames bubbles as a necessary feature of innovation, but also warns that AI capex may be overbuilt and could destroy value among overvalued hyperscalers, subscale software holdings, and PE-owned tech assets. Portfolio positioning: diversification away from concentrated U.S. AI exposure: Dan recommends trimming extreme U.S. concentration toward a more balanced global allocation, favoring cheaper international markets and being cautious about staying overexposed to a single thematic trade. Biotech as a special case for quant investing: Verdad discovered its standard factor model failed badly in biotech, leading to a redesign using market-cap-plus-spend value measures, specialist ownership as a quality proxy, and peer momentum derived from clinical-trial similarity. Research culture and model humility: Dan emphasizes rapid iteration, intellectual humility, and willingness to throw out failing assumptions, arguing that successful quant firms must keep producing new ideas and updating their frameworks.

Key Arguments: Private equity performance remains poor because allocators expected unrealistic alpha after paying very high fees, while underlying businesses were overvalued, low margin, and often poor quality. Private credit has funded much of the PE system; rising bankruptcies at the small end may foreshadow broader distress in the middle and large segments. The industry may be in a delayed adjustment phase where public trustees/CIOs eventually force capital reallocation, but agency problems slow the unwind. Retail access to private assets through 401(k)s and ETFs is likely to create liquidity, valuation, and fee problems rather than genuine democratization. AI is likely to create winners and losers, but markets may already be pricing a lot of the upside into hyperscalers and chip-linked names; overvaluation can create “growth bankruptcies.” The best way to position around AI is not an all-or-nothing bet, but a move toward diversification, especially into cheaper international markets. Biotech requires different financial logic because cash is productive, spending can signal value creation, and specialist ownership provides useful quality information. Peer momentum based on scientific similarity and clinical-trial overlap is more informative in biotech than standard momentum or GICS-style industry grouping. Biotech remains highly idiosyncratic, making long-short and risk management especially important; many weak names can fail while a few strong ones can drive performance. A healthy quant process requires continuously challenging model assumptions, especially in sectors where traditional metrics break down.

Data Points: Private equity management fees: 400 to 600 basis points - Dan cites total fee burden for private equity and related interval fund structures. Expected PE alpha: 400 basis points net of fees - Allocator expectations described as unrealistic given the fee level. Private equity firms growth: Probably doubled in the last 10 years - Industry-level count used to illustrate crowding and competition. Hedge fund firm count change: Basically flat from 10 years ago - Used as a contrast to PE expansion. Small-end private credit distress: Sub-25 million EBITDA businesses - Segment where bankruptcy rates are rising first. PE tech exposure: About 40% of private equity capital - Dan says a large share of PE capital is deployed into software/tech-related assets. SP500 weight of Mag 7: 33% - Kai highlights concentration in U.S. large-cap growth. Mag 7 contribution to index returns: 75% of SP500 returns over the past few years - Illustrates concentration risk in U.S. equity benchmarks. US public companies trend: Number has declined - Dan notes fewer U.S. listed firms over time. International public companies trend: Number has increased - Contrasts with U.S.-specific decline in listed companies. Biotech model failure rate: 80% of worst prediction outcomes - Verdad found most bad outcomes in its factor model came from biotech. Biotech sector return characteristic: Median company lost money for shareholders - Yet the sector has historically outperformed overall. Biotech specialist ownership example: 70% owned by 10 top specialist funds - Illustrates specialist ownership as a quality signal. Biotech strategy allocation: 105% long and 5% short - Describes how some biotech hedge funds effectively de-emphasize shorting. OpenAI valuation/raise mentioned: $50 billion raise; $830 billion valuation - Used as an example of late-stage private-market financing.

Pivotal Quotes: "the reckoning is arriving" — Dan: Describing private equity’s lagging performance, excessive leverage, and growing bankruptcy risk. "the word democratizing, like the word reimagining, should set off alarm bells" — Dan: Expressing skepticism about 401(k) access to private equity and other private assets. "bubbles are good. We want bubbles" — Dan: Explaining that speculative excess can be socially useful during major innovation waves like railroads, the internet, and AI.

Implications: Listeners should expect more scrutiny of PE, private credit, and retail private-asset products, while AI may reward diversification and select beneficiary businesses rather than just hyperscalers. In biotech, specialized models and alternative data can matter more than standard factors.