Breaking Down the Multi-Manager Playbook: How This $19B CIO Thinks About Alpha | Sean McGould | The Lighthouse Group

Sean McGould is CIO of Lighthouse Group, a multi-manager hedge fund with approximately 19 billion dollars in assets under management, roughly 30 years of his own long-short experience, and more than 20 years of proprietary hedge fund data collected by the firm. The portfolio allocates roughly two thirds of risk to global equities, approximately 25 percent to event-driven strategies such as merger arbitrage and SPACs, and approximately 15 percent to liquid macro, deliberately excluding real estate, private equity, and venture capital to preserve the liquidity that the multi-manager model requires. McGould frames the multi-manager structure as the institutional successor to the proprietary trading desks that operated inside investment banks before post-GFC regulations shut them down, drawing a direct parallel to how private credit filled the void left by post-GFC bank lending restrictions.

The theoretical ideal portfolio McGould describes is one hundred uncorrelated return streams with zero correlation to each other, each optimally levered to produce the smoothest possible aggregate return. He is explicit that a riskless portfolio produces no return, so deliberate risk-taking is required even within market-neutral strategies. A portfolio that appears market neutral at the index level can carry significant hidden factor exposure, such as a concentrated long AI position paired against short sleepy industrial companies, which he considers unacceptable factor concentration. The most persistent classical factors are momentum, value, and quality, but thematic factor baskets including stay-at-home versus return-to-work, Republican versus Democrat, and AI have emerged as additional factors requiring ongoing monitoring, and the speed at which themes move from perceived alpha to recognized factor beta has accelerated particularly since COVID.

McGould describes the current trading backdrop as favorable, citing tariff negotiations, an M&A resurgence, and equity capital market activity. He points to the SpaceX and Google recent offerings as record issuances that created market dislocations requiring some investors to sell existing positions to buy new ones, and notes both offerings have gone up in value since issuance. He contrasts this with the 2020 to 2021 SPAC cycle, which involved unprofitable, unscaled private technology companies, some without revenue, and which largely did not work out. Companies coming public today such as Anthropic and OpenAI do have revenue even if not yet profitable, and current issuers like Google are raising capital for a clear use case, specifically investment in more data centers, rather than for opportunistic financing.

Japan and Korea represent McGould's clearest examples of regulatory change as an underappreciated source of alpha. The Tokyo Stock Exchange revised its corporate governance code in June 2021 to emphasize return on equity and capital allocation, accelerated those rules in 2023, and new NISA guidelines released in 2024 encouraged Japanese households to move money from zero-yield deposit accounts into equities, growing NISA accounts from approximately 14 million in 2023 to 28 million currently. The Nikkei moved from around 28000 in January 2021 to approximately 69000 today, outperforming the S&P 500 by approximately 8 percent per year over that period, with trading volume on the Tokyo Stock Exchange now approximately five times what it was three or four years ago. Korea launched a corporate value-up program in January 2024 modeled after Japan, introduced a requirement for boards to consider minority shareholder rights in 2025, and offered 100 percent capital gains tax relief on repatriated capital through May 2026. The KOSPI is up approximately 234 percent since January 2024 versus the S&P up about 63 percent, though Samsung and SK Hynix accounted for 60 percent of KOSPI gains, and McGould attributes outperformance in Taiwan and Korea specifically to significant AI capital expenditure exposure rather than primarily to capital repatriation or dollar dumping. Because regulatory changes cannot be predicted in advance, Lighthouse avoids large country bets such as long Japan short US even when structural improvements seem inevitable in timing.

Long-short alpha in Asian markets is generated by identifying relative mispricing between securities within the same sector rather than by picking absolute index winners, and an individual stock can have a range of roughly 30 percent within a single year, creating relative valuation opportunities. Implied correlations are about as low as they have ever been, making the current environment favorable for stock pickers and relative value traders. McGould cautions that shorting small-cap stocks is dangerous because covering a short can trigger a short squeeze or find no sellers, producing highly non-asymmetric returns, and that raw data advantages get arbitraged away quickly once a data provider sells a new dataset to multiple competitors, leaving proprietary methods of collecting and using data as the durable edge.

On AI adoption, McGould notes that Anthropic stated the financial industry is the second fastest adopter of AI behind high-tech firms. He believes AI will not eliminate human workers but will allow individuals to handle higher volumes, and that human oversight remains necessary due to real money, compliance requirements, and common sense needs. He still believes in the specialist model, arguing that a specialist who has studied a particular industry for a long time has an advantage over a generalist, and that AI tools help specialists get a broad view quickly while allowing them to apply deeper domain knowledge that generalists would not know where to apply. Data quality remains a significant issue when using AI tools in investing.