It happens late in nearly every bull market: Complaints that value-fund managers are beginning to “cheat” on their mandates by sneaking growth companies into their portfolios, high valuations and all, goosing performance now but taking big risks on when the next bear market may arrive. Now those worries are back, with the twist that the tongue-wagging is concentrated on the FAANG names — Facebook, Amazon, Apple, Netflix and Google parent Alphabet, high-fliers that have led growth stocks to a decade of whipping value’s performance.

But there’s a smarter way to do value than just looking for cheap stocks, say experts led by Bill Miller, the unorthodox value investor whose 15-year streak (through 2005) of beating the Standard & Poor’s 500 index is still a benchmark no active manager can touch.

The key is to avoid the mistake of thinking “value” means nothing more than a low stock price. Now running his own fund after decades at Legg Mason Value Trust, Miller’s focus was — and is — on finding companies focused on high returns on invested capital and free-cash-flow growth, as well as large market opportunities. That helps small and large investors alike distinguish stocks that are undervalued from those that are simply cheap.

Twenty years ago that led Miller to Amazon, the first of the FANG stocks, excluding Apple, to go public. Along with early internet leaders, like America Online and Yahoo, Amazon helped Miller update what it meant to do value. And having just scooped up a 30 percent return by adding Facebook shares during the brouhaha over sharing of customers’ data last winter, he argues that those tactics remain smart.

To be sure, data assembled for CNBC.com by Morningstar shows little move into growth stocks by value funds, at least major ones. There are no reliable aggregate numbers for value funds holding growth stocks, but examinations of the largest value funds don’t show a pattern of widening their mandate to chase growth, Morningstar senior analyst Greg Carlson said.

Carlson checked the 20 largest buyers of each of the FAANG stocks, and no value fund was a top buyer of Amazon or Netflix between the the fund’s two most recent reports. Amazon and Netflix are the two most expensive of the FAANG companies, as measured by price-to-earnings multiples, making them the least suitable for value funds by many traditional metrics. No large value fund counted either Amazon or Netflix among its top 20 holdings either, Carlson said.

The only “value” fund among Facebook’s top 20 new buyers was Bill Nygren‘s $19.4 billion Oakmark Fund, which took the stake in the first quarter of this year. A related Oakmark fund also added a big position in Alphabet this year, while the flagship Oakmark Fund has 3 percent of its $19.4 billion fund in Google’s parent. Vanguard Wellington, an allocation fund that blends value stocks with a bond portfolio, is the only other top value-oriented fund that owns Alphabet, and it has had its position since 2014, Carlson said.

“Also, Oakmark Fund is in Morningstar’s large blend category, although Nygren considers himself a value investor,” Carlson said. Oakmark Fund has made about 1.5 percent total return this year, but 16 percent over the last 12 months.

Both value and growth funds have a much bigger problem than claims they are blurring their mandates — and that’s investors’ migration away from actively managed accounts to more passive funds, such as exchange-traded funds, Carlson said. Between the two categories, large value funds and large growth funds have shed more than $100 billion in assets over the last 12 months, he said.

Blurring the line between growth and value isn’t that uncommon — even Warren Buffett, as acclaimed a value investor as has ever lived, has taken a hugely profitable $46 billion position in Apple, the world’s most valuable technology company and one trading at a relatively rich 18.5 times this year’s estimated earnings. In the late-1990s bull market, Miller forged his streak by broadening the definition of value.

In an interview, Miller said the mistake critics make is to think value means nothing more than a low stock price. His focus on finding companies that were focused on high returns on invested capital and generating free-cash-flow growth helps distinguish stocks that are undervalued from those that are simply cheap.

That’s why Miller, like Oakmark, bought Facebook shares around $150 during the brouhaha over the social network’s privacy practices, which have so far had little impact on the company’s advertising sales or profits. At that price, Facebook was trading at just 14 times expected 2019 profits, Miller said — a bargain, unless the privacy controversy led to many more canceled user and advertiser accounts than early results suggested.

“Facebook at 14 times cash flow made no sense whatsoever unless you thought their entire business was in dire straits,” Miller said.

With Facebook trading around $200, Miller is sitting on a 30-percent-plus paper gain made in the course of a few months, he said. Overall, his fund is up 4.3 percent this year after beating the market’s big 2017 gains but trailing the S&P 500 from 2014 through 2016.

In its own way, the Facebook bet was similar to one Miller made on furniture retailer RH Inc., whose Restoration Hardware brand went through a painful conversion to larger stores two years ago that has paid off, he said. RH isn’t the long-term-growth company Facebook is, but its shares also got cheap as the company struggled to sell its conversion to Wall Street. Through its downturn, though, the company kept up its cash flow and bought back huge chunks of its own stock, looking to maximize the bump remaining shareholders would get when things came back.

In each case, the thing to watch was how well the company was generating cash and using it, Miller said. “RH shares fell as low as $25 early last year and have sextupled since, bolstered by an earnings report June 11 that made the stock jump 34 percent,” Miller said.

Other tech companies can deliver that kind of pop, even at relatively high stock prices, but it’s not automatic, he said. Miller is also still a big fan of Amazon, which he said always has been able to maximize cash flow (including operating cash flow topping $25 billion in the 12 months through March) and which has used its cash to grab the early leadership position in the cloud computing business. That business is fast growing — making Amazon look like a growth stock — but it’s also highly profitable, making Amazon attractive to value-oriented investors even now, he said. But he hasn’t bet on Alphabet recently, thinking there are better value plays than Google’s parent, at nearly 28 times this year’s profit estimates.

“To the extent that value is based on a simple calculation about ratios, that’s a very simplistic definition of value,” Miller said. “If people are buying things they haven’t analyzed … it’s not likely to end well.”

But Miller is hardly an ordinary value investor — for years he has benchmarked his now $1.7 billion fund against the S&P 500, a broad market measure, rather than the Russell 1000 Value subindex most value managers compare their performance to, Carlson said.

The top value fund this year has been the Copley Fund, returning nearly 12 percent year-to-date. Its top holdings include JPMorgan Chase, ExxonMobil, Chevron and Verizon Communications. Four share classes of the Federated MDT All Cap Core Fund are among the select group of funds that have have returned as much as 6 percent so far this year, according to Morningstar — its biggest holding is Apple, and it also owns open-source software company Red Hat and Alphabet stock.

A more representative value manager is T. Rowe Price’s John Linehan, who runs the mutual-fund company’s $21 billion Equity Income fund, which has returned 10.2 percent for the 12 months ending May 31.

He doesn’t own the FAANG stocks — his closest such play is Microsoft, which the fund bought before CEO Satya Nadella led Microsoft into the cloud computing business. That focus has moved Microsoft onto growth-stock screens even as the company milks cash from its mature Windows and Office franchises. To find a value stock, you can focus on either companies with low valuations or a secular business challenge it can probably solve — much as Nadella later fixed Microsoft’s stagnation, he said.

But most of Linehan’s big recent additions have been more traditional value plays — Wells Fargo, betting that the banking giant can bounce back from its fake-accounts scandal; utility company Southern Co., which has loaded up on debt since 2015 and is now looking to cut its burden through asset sales; real estate developer S.L. Green and Philip Morris.

“Our process hasn’t been modified at all,” Linehan said.

He argued that the talk of value managers chasing growth stocks, if anything, would be a sign of an approaching bottom for value plays late in an expansion. When value finally began outperforming growth stocks after the 1990s internet boom, “it happened very quickly,” he said.

“Kohl’s has more than doubled in the past year,” Linehan said. “Reports of value’s demise have been exaggerated. You ignore these companies at your peril.”

.