It wasn’t the best of times for Rob Arnott, star investor for Pacific Investment Management Co. and early champion of the discipline known as smart beta.

Arnott, who oversees $38.1 billion for Pimco, was leaking cash. Clients pulled $15.8 billion from two of his mutual funds last year alone, and two big customers had been urging him to rethink the models that dictate his investments. They wanted more options than the fundamental indexes he designed in 2005, which seek to locate underpriced stocks by using weightings other than market cap. Value investing was out of vogue, and they wanted the latest hot alternatives.

As the 61-year-old tells it, he started looking into whether his clients were right, and what he discovered convinced him to turn them down. Smart beta, pioneered by Arnott and now the basis for almost one out of every five exchange-traded funds globally, had become a victim of its own popularity. Strong results often came from inflated valuations and many strategies were set to crash, he said.

“You can use smart beta smartly or you can use smart beta stupidly,” Arnott, chairman and chief executive officer of Pimco subadviser Research Affiliates LLC, said in a phone interview during a visit to Tokyo. “Strategies become so popular that the only way for them to succeed is to go from very expensive to more expensive. That’s a dangerous game to play.”

As a money manager, Arnott is wedded to indexes in a way most investors are not. His smart-beta approach has become wildly popular in the past decade, lauded by proponents as a way of squeezing bigger profits from indexes by choosing superior companies.

The problem as Arnott sees it is that many smart-beta products do no such thing. Rather, they’ve benefited from a herd effect, increasing primarily because investors piled into them and pushed up valuations. In his worldview, that’s a warning sign, because gains attributable to that kind of appreciation are doomed to reverse.

When Arnott published a report on his findings last month, he transformed from icon of an industry into one of its biggest critics. If he’s right that excess returns -- or alpha -- were mainly a “mirage,” it raises uncomfortable questions about the hundreds of billions of dollars invested in a business that exploded in the past 15 years.

“Reaction has been wide-ranging -- from people in the smart-beta community reacting angrily, from people in the consulting and asset-owner community reacting with enthusiasm and gratitude, and from competitors saying, well, ‘they’re just talking their own book, because value is what they do,”’ Arnott said, noting that the calculations would have identified his kind of investments as expensive in 2007. “Coming forward now with this research doesn’t guarantee that it’ll always point in the direction of value.”

Smart-beta managers tweak indexes, seeking an improvement on passive styles and greater consistency at lower fees than traditional active investors. While definitions of smart beta differ, it generally cuts the link between a company’s market value and how it is ranked in a gauge. Early examples include equal-weighting, where all stocks have the same prominence. Another variant is Arnott’s fundamental indexing, which uses criteria such as book value, cash flow and sales to select undervalued shares.

Then there are so-called factor-tilt approaches, which alter indexes based on specific characteristics expected to influence returns, anything from positive share-price momentum to high profitability or low volatility.

Arnott studied the performance of 12 strategies over a period from 1967 to 2015 and found many of them succeeded not by finding the best companies, but only by picking stocks that became more expensive. Higher valuation levels not only inflate past performance, they also reduce the potential for future returns and create increased risk of mean reversion, he said. Value is in its cheapest decile in history, according to Arnott’s analysis.

Soaring Popularity

“We think it’s reasonably likely a smart-beta crash will be a consequence of the soaring popularity of factor-tilt strategies,” Arnott and colleagues wrote in the report.

A second paper coming shortly will show that strategies targeting stocks with low volatility compared to the market are set to underperform as valuations approach all-time highs, while value-based approaches are among those with room to gain, Arnott said.

Funds focused on low-volatility stocks have exploded, with assets expanding to $20 billion from almost nothing five years ago. The iShares MSCI USA Minimum Volatility ETF has garnered $3.1 billion this year, or about a third of total assets, as the fund is up 4.3 percent, compared with a gain of 0.3 percent in the Standard & Poor’s 500 Index.

While valuations matter, predicting the success of smart-beta ETFs based on valuations can be misleading, according to Jay Jacobs, director of research at Global X Management Co., a New York-based ETF provider.

“When people don’t know the direction of the market, or they’re concerned about volatility, there will be a flight towards certain factors such as low volatility or quality,’’ Jacobs said. “That can drive up valuations and returns. If the market sells off tomorrow, those factors can still be really good to own.’’

Value, Momentum

Bargain hunting is showing signs of life after struggling as quantitative easing lifted all boats and traders stayed with winners regardless of price. While value lost to momentum by more than 50 percentage points in the U.S. in 2015 as megacap technology firms dominated, the wilder markets of 2016 are widening the differences between share prices, prompting famous stock pickers to declare their time has come.

“It’s easy to dismiss relative valuation and to chase performance, to chase fads,” Arnott said. “I’m not saying that smart beta is a fad,” he said. “The notion of systematic, disciplined approaches to investing that can reduce costs, increase transparency and improve investor results is a wonderful idea.”

For Arnott, academia gets some of the blame. The influence of the efficient-market theory has meant valuation isn’t considered important, he said. When someone comes up with a new smart-beta factor, it’s common to test if it’s truly original, but nobody bothers to check whether it worked in the past purely because it got more expensive, he said.

Marketing Ploy

While Jack Bogle, founder of traditional indexer Vanguard Group, has said smart beta is “baloney” and a marketing ploy, others see it as neither investing panacea nor something to be avoided. Rick Ferri, the founder of Portfolio Solutions LLC, a Troy, Michigan–based financial adviser, said in August that using the strategies required a “lifelong commitment,” because there would be times when they’d underperform.

Arnott’s two clients, meanwhile, reacted differently when he refused to emulate popular approaches. One thanked him, the other was angry. While time will tell which one was right, it’s clear that Arnott’s warning about his booming industry has ignited a debate about its future.

“Valuation matters,” Arnott said. “Why wouldn’t people for decades have posed the question: did this factor win by getting more expensive? It’s such an obvious natural question. The shock to me is that this wasn’t done decades ago.”