Intervista a Ben Felix: AI, Bolle, Decumulo e... Pepperoni! | ft. @BenFelixCSI

Finance-Focused Summary

AI and Financial Advice

• Ben Felix said PWL Capital uses AI to match meeting transcripts against a text database of 100+ financial- and tax-planning strategies, suggesting relevant strategies for advisers to consider. The goal is to improve efficiency while preserving human advice.
• He cautioned that general-purpose large language models can give incorrect or subtly misleading answers. PWL’s approach is to embed its advisers’ professional judgment rather than rely on AI to make client-specific recommendations independently.
• Felix expects AI to change work across industries, but said PWL is not currently seeing reduced staffing needs.
• The speakers discussed regulatory gaps around AI-generated personalized financial advice but did not reach a firm view on how it should be regulated.

Risk Tolerance, Advice, and Portfolio Construction

• PWL licenses a risk-tolerance questionnaire developed and tested by academics. Felix said it can assess risk tolerance, but composure during an actual market crash cannot be known until someone experiences one.
• Citing a conversation with Ken French, Felix said a crash can provide a new data point about an investor’s ability to tolerate volatility. That may justify changing an allocation—not to time the market, but because the investor has learned something about their own risk capacity.
• Some well-informed investors hire advisers for peace of mind, time savings, expertise, tax and retirement planning, or because professional support helps them tolerate a more suitable level of risk. PWL charges on an assets-under-management (AUM) basis.
• Felix views a low-cost, globally diversified index fund as a strong default for DIY investors. A moderate factor tilt is another approach he uses, but it can involve long periods of underperformance versus a broad index. He said benchmark-relative performance concerns can lead investors and institutions to abandon strategies prematurely.
• PWL uses Dimensional funds. Felix described them as more diversified and index-like than concentrated small-value funds, while still applying factor exposures and discretionary trading decisions.
• He noted that factor-based funds can also exclude companies with high prices and poor profitability—what he characterized as “junk” or “slop.” Factor strategies are not free of uncertainty: research results can depend on the data vintage and methodology. Felix said this uncertainty argues against extreme factor concentration.

Valuations, AI-Related Stocks, and Diversification

• Felix described US equity valuations as high, but said that alone does not establish that the market is in a bubble. High valuations can reduce expected returns, though he did not conclude that returns must be zero or negative.
• He cautioned against trying to time bubbles. Japan was cited as an example where valuations rose to roughly twice or more current US market valuations before the eventual decline; investors who exited too early could have missed substantial gains.
• Felix said PWL’s geographically diversified portfolios and factor tilts naturally reduce exposure to the largest, highest-priced companies. He cited the US “lost decade” from the dot-com bust to roughly 2010: US stocks had about 0% return over that period, while US small-cap value stocks performed much better in relative terms. He also said Japanese small-cap value outperformed the broader Japanese market over a long period.
• He argued that concerns focused on a handful of highly valued companies should be considered in the context of the broader market. Nvidia was mentioned as a very large company with a high valuation but also substantial fundamental growth. SpaceX and Anthropic were discussed as examples of companies attracting attention, with Anthropic described as not publicly traded at the time.
• As a concentration-risk example, Felix said Nortel once represented about one-third of Canada’s stock market and later fell to zero. Despite that, the Canadian market outperformed the US through the dot-com period, illustrating that other companies can offset the failure of a dominant constituent.
• The host cited the US as about 65% of the global market and the “Magnificent Seven” as about 30%. These figures were raised by the host, not independently verified by Felix in the exchange.
• On home-country bias, Felix said Canadian investors have tax reasons to favor domestic shares and cited research suggesting an allocation of around 30% to the home market, while noting limitations in the supporting studies. For Europe, he said the relevant reference may be currency exposure rather than an individual country, but he was not offering a Europe-specific recommendation.

Factors, Indexes, and Research

• Felix pushed back on treating factor funds as purely mechanical or fully passive. He said Dimensional and Avantis use trader discretion to manage transactions and have changed their methodologies over time.
• He also described the S&P 500 as historically discretionary: a committee selects “leading” companies rather than simply including the 500 largest US companies.
• The speakers discussed concerns about “factor-zoo” research and replication. Felix said many published factors appear to reduce to a smaller set of broader factors with plausible risk- or behavior-based explanations. However, he also cited research showing that results can vary out of sample and across versions of historical datasets.
• Felix said this uncertainty supports moderate rather than extreme factor tilts. He described PWL’s portfolios as much closer to broad index funds than to concentrated small-cap-value portfolios.

Alternative Investments and Risk Parity

• Felix said PWL generally does not recommend complex alternatives such as trend following, carry strategies, or managed futures to clients seeking straightforward market returns and advice on allocation, spending, and taxes.
• He was skeptical of risk parity, saying it can rely heavily on historical correlations, returns, and volatility, and may use leverage to reach a target return or risk level. He characterized many appealing results as backtest-dependent; hypothetical examples discussed included 50% or 100% leverage.
• Assets often included in risk-parity portfolios—such as gold, long-term bonds, and managed futures—may look attractive in historical tests, but Felix said he lacks sufficient confidence in the assumptions to recommend them to clients.
• He cautioned that correlations can rise in crises and that even gold has been an imperfect diversifier during equity-market stress.
• On the total portfolio approach, Felix said it may encourage adding assets with different economic exposures, but can also lead to complex allocations, including private assets, and may require leverage.
• He discussed Canada Pension Plan Investment Board (CPP Investments) as a prominent institutional investor. It shifted from index investing to active management in 2006 and, according to Felix, had underperformed its reference portfolio—roughly its former index-based benchmark—over the period discussed, while incurring substantial fees. He said its approach and future crisis performance remain open questions.
• Felix cited CPP Investments’ assets as about C$863 billion, explicitly noting that the figure came from an AI summary and might be inaccurate. He said the Canadian pension system is structured to remain able to pay benefits as demographics change.

Private Equity and Private Markets

• Felix said access to private equity for retail investors is not automatically democratizing: wider access may bring private-company valuations closer to public-market valuations and expose investors to high fees and illiquidity.
• He said IRR has limitations for comparing private funds with public markets. Public Market Equivalent (PME) is designed to address some of those issues. In the research he referenced, private equity’s gross-of-fee returns were high, while net-of-fee returns were broadly in line with public markets.
• He questioned whether investors reliably receive an illiquidity premium, noting that private-equity, private-real-estate, and private-credit returns have not made such a premium obvious. Smoother valuations may feel reassuring, but they also reduce liquidity.
• Felix disclosed that he retained a substantial equity stake following PWL’s acquisition by OneDigital in January of the prior year. He said he kept more than half of his PWL equity and requested a larger equity component than the overall deal structure. He cited Canadian tax deferral, confidence in OneDigital’s business, and his desire to retain an ownership interest as reasons, while acknowledging significant company-specific risk.
• He also said that CPP Investments was an investor in the company that acquired PWL, a relevant relationship to bear in mind when interpreting his comments about CPP.

Crypto, Prediction Markets, and ETF Products

• Felix said he bought Bitcoin and Ethereum while studying crypto, partly to understand the system rather than to pursue high returns. After the podcast series, he concluded that crypto was not the financial revolution he had initially been asked to examine. He later sold his holdings when an exchange-related issue required action; he said the amount was immaterial and that he did not sell because prices were down.
• On prediction markets, Felix said they may generate information, but absent special information or an edge, they are not a long-term investment strategy and are negative-sum for an individual after costs.
• He criticized the proliferation of single-stock ETFs, including leveraged, inverse, and covered-call products, calling the ecosystem messy. He cited a Morningstar article showing that most such products underperformed their underlying stocks, in part because volatility can erode compounding, and noted that fees can be high.
• He described some fintech platforms as using revenue from areas such as crypto, prediction markets, or individual-stock trading to subsidize low-cost ETF trading. He characterized this as a transfer from less-informed to more-informed participants and left the question of whether that is socially beneficial to regulators.

Retirement and Decumulation

• Felix cautioned that maximizing a safe withdrawal rate can optimize the wrong objective if it leads to excessive efforts to reduce volatility. He suggested that some flexibility in spending, where feasible, can be part of a retirement strategy.
• PWL may adjust a client’s asset allocation based on their starting situation, but Felix said it generally does not make sweeping portfolio changes simply because someone has retired. Changes may occur gradually before retirement; for some clients, the retirement allocation may be set when they begin working with PWL.
• Retirement planning also includes decisions about which accounts to draw from for tax purposes, when to take government pensions, and how those choices interact with portfolio withdrawals.
• PWL’s planning process includes identifying a client’s goals, assigning costs to them, and using a financial-planning projection model to compare spending or life choices now versus later.
• Felix said clients may overemphasize accumulating more wealth at the expense of current time and quality of life. Given PWL’s client base—which is financially strong—he said the firm more often helps clients see whether they can spend more on things that matter to them than tells them to save more.
• He cited cases where clients realized they could become financially independent and left jobs they did not enjoy.

Other Discussion

• On universal basic income, Felix said he had no opinion. The host expressed skepticism that automation would produce a sustainable, broadly adequate income for everyone, but those comments were the host’s personal view.

Key Figures, Instruments, and Entities Mentioned

• C$863 billion: Approximate CPP Investments assets, with Felix caveating the AI-generated source.
• 2006: CPP Investments’ shift from index-based investing toward active management.
• About 30%: Home-country allocation cited in research discussed by Felix, particularly in a Canadian context.
• About 0%: US stock return over the dot-com-bust-to-roughly-2010 period, as described by Felix.
• About one-third: Nortel’s former share of the Canadian stock market, according to Felix.
• Instruments and assets discussed: Global and US equity indexes; S&P 500, MSCI, and FTSE indexes; small-cap value and profitability-factor strategies; Canadian, international, and emerging-market equities; government and high-yield bonds; gold; managed futures; private equity, private credit, and private real estate; Bitcoin; Ethereum; and leveraged, inverse, and covered-call ETFs.
• Companies and entities discussed: Nvidia, SpaceX, Anthropic, Tesla, Nortel, Dimensional, Avantis, PWL Capital, OneDigital, and CPP Investments.

Disclosures and Cautions

• Felix disclosed his retained equity stake in OneDigital following PWL’s acquisition and acknowledged the associated concentration and company-specific risk. He also disclosed CPP Investments’ relationship to the acquiring company.
• PWL charges a percentage of assets under management.
• No explicit general “not financial advice” disclaimer was stated in the subtitles.

Presenters and Sources Mentioned

• Presenters: Andrea (host, associated with Mr. RIP) and Ben Felix.
• Organizations and programs: PWL Capital; Rational Reminder podcast; Dimensional Fund Advisors; Avantis; CPP Investments; OneDigital.
• Researchers, commentators, and other sources referenced: Ken French, Eugene Fama, Meir Statman, Victor Haghani, Scott Cederburg, Marcos López de Prado, Andrew Chen, Dimitri Morav? (name unclear in subtitles), Dan Bortolotti, David Blitzer, Ludovic Phalippou, Graham Stephan, Thomas Peterffy, and Morningstar.