AI has grown from a niche bet to a core component of benchmarks in just a few years, yet Aziz Hamzaogullari, founder, chief investment officer, and portfolio manager for the growth equity strategies team at Loomis, Sayles & Company, says that theme-driven investing is likely to leave most investors disappointed.
“This is true going back to radio when it became a big thing, or the telecom bubble – it’s the same thing with AI and the disruption that we are having today,” Hamzaogullari said during a recent podcast with Top1000funds.com.
“I would argue maybe five or six companies will benefit disproportionately from AI and everyone else that’s riding this in terms of price momentum will pay a price because, at the end of the day, they’re not going to justify those inflated expectations. But then those few winners are most probably underpriced because people are underappreciating how long these companies will be able to sustain their growth.”
The Magnificent 7 – which make up about one-third of the S&P 500 index – continue to invest hundreds of billions of dollars in AI strategies. Some of those names are also held in Hamzaogullari’s fund: Nvidia, Alphabet, Tesla, Meta, and Amazon are its five biggest holdings within a concentrated portfolio of typically 30 to 40 stocks.
But those familiar positions can obscure a portfolio that Hamzaogullari argues is still materially different from its peer group. He points to a broader problem across the active management industry – most managers are active in name only and largely mimic market-cap-weighted indices.
“It’s not what you own, it’s how you own it,” he said.
“If you look at a company like Amazon, we’ve owned it since our inception in 2006. Since then, there have been roughly 9,750 funds that owned Amazon at one time or another, but only 57 – or 0.6 per cent – of them owned it for the entirety of that two-decade time period.
“The second important aspect is it’s not just if you own Amazon for 20 years or not, but what was your weight in that company over that timeframe? The median ownership in those 57 managers that owned Amazon was only 1.6 per cent – our position was around 6.3 percent.”
Why diversification isn’t achieved by holding more stocks
The tendency of most active managers to be swayed by behavioural biases such as price momentum and recent events also encourages benchmark hugging. Perhaps not surprisingly, active fund managers tend to do well in either rising or falling markets – but very rarely in both, according to a Loomis, Sayles & Company analysis of 154 active managers with a long-term track record.
“We realised that the biases are being created by these managers at the initial stage of their process,” Hamzaogullari said. “Our approach is different – rather than having a reactionary approach to up or down markets, we construct portfolios with different business drivers that we believe can do well in both types of environments.”
The fund’s maximum exposure to any given business driver is about 15-to-20 per cent, while its 30 to 40 name portfolio diversifies away about 85 per cent to 90 per cent of diversifiable risk.
The portfolio may include significant exposure to some Mag 7 companies, but it is highly diversified, with other stocks such as energy drink company Monster Beverage and biotechnology company Vertex, which develops treatments for cystic fibrosis.
“The correlation of a business like Vertex to a company like Nvidia, which is driven by AI, is very low.”
The discipline, he says, comes from a stringent selection criteria that evaluates ideas through three lenses – quality, growth and valuation. It enables the fund to exploit short-term mispricings.
“In the near term, we believe that innate behavioural biases such as herding, overconfidence, and loss aversion influence investment decisions and create asset pricing anomalies, and these pricing inefficiencies converge towards intrinsic value over time.”
Quality companies are difficult to replicate by competitors even with time and capital – a high bar that encompasses only around 200 companies globally. Sustainable and profitable growth is part of that equation, with fewer than 1 per cent of all companies consistently achieving above average cash flow growth over a decade.
Buy rarely, hold for years, sell with discipline
When those companies trade at a significant discount to intrinsic value, the fund buys them.
“That doesn’t happen every day, so that’s the reason why we typically make two or three decisions in a given year,” he said.
“Many businesses will fail the test of quality, growth, and valuation, and that selectivity and focus helps us understand what we own. A long-term time horizon is a very crucial part because our turnover in the last two decades has been around 10 per cent which means our holding period is a decade or longer.”
Hamzaogullari also argues that the benefit of such a long holding period is measurable. He points to research suggesting investors can add about 200 basis points of alpha by backing low-turnover managers and avoiding high-turnover strategies, where higher costs compound over time.
That discipline extends to sell decisions. He exits holdings for three reasons: when a stock reaches intrinsic value, when a better risk-reward opportunity emerges, or when the original thesis is broken.
Of the 64 companies sold in the strategy over the past two decades, about half were sold because they reached intrinsic value. About one-quarter were sold because a better risk-reward opportunity emerged, and one-quarter because the initial investment thesis was wrong.
“Ninety per cent of those companies that we sold underperformed our portfolio. I believe that the transparency and having a clear process of uncovering mistakes is very crucial in our investment process.”
