Is this the sound of a stock market bubble popping?

That depends who you ask. In September, super-bull Cathie Wood faced off with market skeptic Rob Arnott at a Morningstar investment conference. Their debate centered on whether the relentless rise of big-cap stocks to notch record after record, to reach levels seemingly out of whack versus the traditional fundamentals, signals a new normal (Wood) or that the S&P 500 is substantially overpriced (Arnott). Wood, chief of ARK Invest, declared, “This is no bubble,” arguing that advances in A.I., genome sequencing, and blockchain technology—“profound platform opportunities”—will keep equities surging for years to come. Arnott, founder of Research Affiliates, a firm that oversees $171 billion in mutual funds and ETFs, countered that today’s valuations can only make sense if corporate earnings keep waxing from already colossal, never-before-seen heights for years to come, a scenario he doesn’t see happening. As a prime example of today’s excess, Arnott cited Wood’s darling and biggest holding, Tesla, asserting that the EV pioneer’s shares are vastly inflated since “it was priced for stupendous growth, and got impressive growth.” Wood shot back that her investment choices are “looking forward,” while Arnott’s “questions are backward looking.”

Presumably, Wood was challenging Arnott for relying on the past trends that traditionally govern the markets, such as the tendency of price/earnings ratios and corporate earnings to retreat to historical levels following periods when they seem to defy economic gravity. For Wood, Tesla and the other great innovators will indeed deliver the epic profit expansion their towering market caps foretell. The bears just don’t get it, contends Wood. The “muscle memory” from the tech bubble and Great Financial Crisis makes the naysayers overly fearful of another crash, and blind to the golden road ahead.

The Wood-Arnott duel spotlights the most consequential, baffling issue in the investment world today. The gorilla in the arena, the question that dwarfs all others: Are we in a bubble? The answer will determine how Americans’ 401(k)s, college funds, stock and bond portfolios, and bets on such high-fliers as Bitcoin and Ethereum fare in the years to come.

The investment world’s wild atmospherics indeed point to a craze resembling the period preceding the tech crash of 2000. Gen Z zealots are driving beaten-down meme stocks such as GameStop and AMC, fallen names offering little hope of revival, to huge, seemingly unhinged valuations. The speculative fervor extends to cryptocurrencies; faith that someone will buy your Bitcoin for more than you paid has multiplied its price fourfold in the past year, though you can’t use the digital tokens for much of anything. Most of all, Wood’s view that “this time it’s different,” that breakthrough technologies are remaking the traditional bounds of profitability––even though in the past, competition ensures that doesn’t happen––is rallying hordes of new followers to get hip, junk the old-fashioned metrics, and jump on the M-for-momentum express.

For such market sages as Jeremy Grantham, Carl Icahn, and Jim Grant, the Wall Street crowd is living in the same fantasyland as in 1998 and 2006. They claim that valuations are unglued from the financial basics, that already off-the-charts prices that keep leapfrogging to ever-higher summits mark the late stages of a frenzy. For these veterans, the market’s on the cusp of a collapse.

The bulls, however, cite what sounds like a sound argument for why prices are reasonable and will advance from here. It’s that the most widely used indicator of whether big-cap valuations are rich or lean, the S&P’s price/earnings multiple, shows that stocks aren’t expensive at all, and enjoy plenty of room to run. Sure, the P/Es are a bit high by historical standards. But that premium is justified because today’s super-low interest rates make equities a great deal alongside the Treasuries and corporate bonds that are paying puny coupons.

Here’s the problem with the optimists’ case: P/Es only look this “normal” because earnings are stretched not just to the max, but beyond. The pandemic stirred special forces, notably trillions in stimulus spending and the stay-at-home economy’s gift to the tech titans, that greatly swelled corporate profits. While those trends do have legs, the explosive power they exerted in the lockdown and early recovery will fade over time. The bulls are really arguing that profits will keep devouring far bigger slices of the U.S. economy than ever before. In that scenario, profits would keep growing at the expense of wages, and consumers would keep buying ever more expensive goods instead of going with cheaper products. The workers, the legislators, family budgets, and most of all, the rigors of competition, won’t let it happen.

‘Absurdly elevated’

Put simply, today’s profit bonanza is a craze unto itself. Any convincing view on where stocks are headed must forecast that earnings will fall, or at least go flat, from here. “Corporate profits are in a bubble,” says Chris Brightman, CEO and chief investment officer at Research Affiliates. “They’re absurdly elevated. Investors have been extrapolating growth that can’t go on forever.” But Brightman acknowledges that the earnings bash has lasted much longer than he expected, and that the recent leaps from already seemingly unsustainable heights to fresh records is an especially big surprise. The ascent’s longevity and recent firepower makes it extremely difficult to predict when earnings will stall or fall, he cautions. “I’ve been saying the rate of corporate profit growth has been not sustainable and the level is in bubble territory, and that earnings have to revert for a decade,” says Brightman. “But profits keep going up.”

The popping of the tech bubble sent the S&P 500 tumbling 45% from mid-2000 to late 2002. This time, the damage is likely to be a lot less disastrous, but still moderately severe. So instead of calling today’s scenario a bubble market, I’ll invent a new B-term, and label it the Boxed-In Market. In the long term, profits must stay within guardrails fixed by such measures as their share of national income. Today, they have ventured beyond those limits, and the retreat in the pandemic windfall that got them there and rivals lured by their rich margins will force them back into the box.

That looming “big shrink” could come in the form of a sharp downturn in actual dollar earnings. Or they could flatline at current levels so that they keep falling relative to inflation. Either way, that we’ve hit “peak profits” portends substantially lower stock prices. We’ll get to our prediction shortly.

Keep in mind that our boxed-in, less-than-a-bubble outlook makes a pivotal, positive assumption: It’s that P/E multiples remain in the mid-20s, big numbers by historical standards. But they’ll only settle at that plateau if interest rates remain extremely low. If we suffer an outbreak of inflation that forces the Fed to cool the economy by raising “real” rates, P/Es would plummet. Then stocks could experience a deeper drop that in retrospect would make today’s backdrop look like a true bubble.

To gauge what today’s outsize earnings mean for the market’s future, this writer spoke to founder Arnott and two other experts at Research Affiliates, CEO Brightman and partner Vitali Kalesnik, an economics Ph.D. from UCLA. Their overall view is that right now, it’s these history-defying earnings that make valuations look reasonable, when in fact they’re anything but. The rub, they say, is that those super-rich profits can’t last. The complicating factor is that the earnings bandwagon could roll for a while, pushing prices still higher.

Says Arnott: “The earning surge is real and likely to be sustained for the next couple of years, [but since it’s temporary] the market is structurally overvalued.” For Kalesnik, today’s high prices are baking in still more earnings gains that over time won’t happen. “A long-term ‘real’ positive rate of return [for the overall stock market] is out of the picture because valuations are high,” he notes.

The incredible profit surge of 2021

The profit moonshot in the first half of this year ranks as a landmark in the annals of equity markets. In Q4 of 2019, S&P 500 earnings-per-share hit an all-time record of $139.47. That was 5% above the previous yearly pinnacle reached in 2018, and one-quarter better than any January to December number ever recorded. Yet after the pandemic crushed profits in the first half of 2020, they rebounded quickly, nearly regaining their 2019 marks in Q3 and Q4.

The takeoff started in Q1 of this year, when EPS hit $45.95, one-third better than any number ever recorded. The purple patch ran into Q2, when the big-caps earned $48.39 (based on S&P’s preliminary data). Wall Street is famously overoptimistic, and analysts typically forecast much higher results three or six months before the profits are counted than what the 500 delivers. But that’s not the trend this time. In fact, Howard Silverblatt, the senior index analyst at S&P Dow Jones, marvels that big-caps’ current performance is flummoxing even the bulls. “These results are incredible,” says Silverblatt. “On June 30, before the numbers for Q2 started coming in, the analysts were predicting EPS that turned out to be more than 18% lower than what companies ended up making. They’re predicting slightly lower earnings in Q3 than in Q2, but I wouldn’t be shocked if we get another record. That’s the way things have been going.”

The analysts polled by S&P indeed expect the good times to keep rolling, and according to Silverblatt the results may continue to be even better than they expect. They’re forecasting average EPS of around $46 to $47 for Q3 and Q4; if that happens, the S&P would notch $189 for the year, beating the all-time high in 2019 by 36%. That’s the same number you get by extrapolating Q1 and Q2 EPS for all of 2021. So we’ll use $189 as our benchmark for where S&P profits stand today.

Silverblatt attributes much of historic run to pricing power. “Profit margins to sales are at all-time highs,” he observes. “They’re now running at 13.6%, way above the average of 8.1% since 1993. Big companies are able to keep pushing price increases through to consumers. That will eventually stop. But even if they meet resistance and can’t push through as many increases, it’s a long way from almost 14% to 8%. So companies could be highly profitable for a long time.” For Silverblatt, a major tailwind for profits is that many Americans who kept their jobs and parked lots of savings in the pandemic are now spending freely. “If we see success over the variant, that trend will get even stronger,” says Silverblatt, who adds that for now, he’s so glad to get out to restaurants again that he’s little bothered by a high tab, and believes a lot of people feel the same way. “It would take a new virus that goes nuts, that closes schools, that had the same effect as, say, last March and April, to change this scenario,” he notes. “Otherwise, more good numbers are baked in.”

Earnings now hover far above the 2019 record, and stand at an all-time-high share of GDP

The Olympic-scale vault in earnings is masking a dangerously overpriced market. The 500’s P/E based on this year’s EPS of $189 is 23.6, about 8% above its average over the past three decades. The bulls keep saying that’s far from frothy. But inflated profits are making P/Es look deceivingly, artificially low, when stocks are in fact extremely expensive. We’re seeing slightly elevated valuations on top of hugely elevated profits. Profits on that scale can’t last, let alone keep growing, an outlook that spells big trouble for stocks.

How do we know profits will soon hit a wall? A good indicator of whether they have room to run or have reached their outer limits is their share of the U.S. economy. The current S&P 500 EPS mark of $189 is 36% higher than the record of $139.47 set in Q4 of 2019, and 72% above 2017. Big-cap earnings didn’t reach these incredible numbers because America’s overall economy boomed in the past several years. In fact, the Congressional Budget Office predicts that the recovery will hoist national income to $22.4 trillion this year. That’s a 14.6% increase over 2017. Hence, over those four years, S&P profits rose five times as fast as GDP.

Profits got there in large part because what companies paid workers stayed relatively flat, enabling producers to bank an ever-increasing portion of their revenues. The proof is the share of the GDP pie now flowing to corporate earnings. Since 1988, S&P 500 profits have averaged around 5% of GDP. In the flush 1998 and 2006 periods, their slice reached 5.5%. In the likely event that its members deliver that $189 a share, the S&P 500 will register total net profits of $1.69 trillion for 2021. That performance would put their earnings to GDP at a record of 6.9%, almost 40% above the 33-year average, and one-quarter higher than at other peak periods. Over long periods, earnings weave through peaks and valleys, but they tend to revert to their historical share of national income. Fierce competition and a push from labor to get a bigger piece of the pie have always hammered profits when they far exceeded the long-term norms. It will take a while, but the same forces will reassert their power this time, corralling profits back inside the guardrails.

The looming shrink in deficit spending

In my conversations with the folks at Research Affiliates, they cited four reasons that today’s Brobdingnagian earnings are destined to decline. The first is a booster that most observers might find surprising: the huge jump in deficit spending, chiefly tailored to counter the damage from the pandemic. In fiscal 2020, the U.S. ran a record shortfall of $3.13 trillion, more than triple the figure in 2019, and this year, the Congressional Budget Office is forecasting a budget gap of $3.0 trillion. Brightman explains that by pure math, the excess of spending over tax revenue lifts corporate profits. “The increased deficits must be offset by a combination of higher household savings, and corporate savings, and corporate savings equal earnings,” he says. “Though part of the increase in deficits went to the former, a lot went to the latter. That trend was a big factor in boosting profits.”

How does that spending get into people’s wallets, and get spent on a new laptop or outdoor deck for your home? “The government wires the money to people’s bank accounts, as via the CARES Act, a child tax credit, unemployment benefits, and other payments,” says Brightman. “Then people log in to Amazon and use the extra cash to buy products.” Part of the money goes to profits for Amazon, for the company that supplied the product to Amazon, and for others in the chain that made the product’s components.

Of course, we don’t know how much of the Biden administration’s $3.5 trillion “human infrastructure” spending bill will be enacted, or how much of the cost will be offset by tax increases. But it seems certain that deficit spending will drop dramatically in the years ahead from the awesome, $3 trillion gaps of 2020 and 2021. The CBO now foresees the shortfall declining by two-thirds to $1.15 trillion in 2022, and hitting less than $1 trillion the following year. “It’s impossible to predict what will happen in Washington, but if the deficits do decline, that will cause a hit to profits,” says Brightman.

Big Tech earners helped by the pandemic will lose momentum

Kalesnik notes that the COVID-19 crisis provided a giant push for tech titans, rendering the pre-pandemic superstars even more sumptuously profitable. “We saw a huge increase in earnings in big growth stocks, especially in tech, while value plays in such areas as energy, airlines, financials, and hospitality got hurt,” he says. “Those are the sectors that will profit most coming out of the crisis, while the tech companies that have gigantic profits will do a lot worse.” Amazon doubled its operating earnings in the past four quarters to $29 billion over 2018, and Apple waxed 50% over 2019 to $99 billion. “People were glued to their cell phones, benefiting Google as well,” he says. Adds Brightman, “People in the stay-at-home economy were buying a lot more laptops and other products from Apple and Microsoft to upgrade their home equipment.”

Big Tech’s profits will erode gradually rather than plummet, Kalesnik predicts. That’s because “spending that was postponed in the pandemic is still coming forward. People are now happy to go out and spend.” The jubilation resembles the huge parties in medieval times that followed quarantines from the plague. The consumption spike will eventually fade, and a bigger proportion of the dollars that went to the tech champs during lockdown will flow to such beaten-down sectors as hotels, restaurants, and energy that will benefit most from the reopening. “People will go back to stores and return to a lot of their old buying habits,” says Kalesnik. The net effect, he notes, will be a slow downward spiral, where S&P profits lag inflation.

Workers will get a bigger share of the pie, and U.S. regulators will attack profits

The third factor: The profit feast is coming at the expense of workers, and suddenly, labor is gaining a lot more clout. “We’re seeing tight labor markets,” says Kalesnik. “Hunting for talent is pretty tough.” Groups from truckers to construction workers to chefs to nurse practitioners are in short supply, and commanding big pay increases. The times when bigger and bigger shares of revenues go to profits is ending. Companies still reaping big margins must hire more people to expand, and the competition for those workers will drive wages up and earnings down. Says Kalesnik: “The times when profits rise much faster than wages can’t last forever. We’ll see a reversal in that trend to a period of lower profits and higher compensation.”

The fourth and final threat to corporate profits is government action. The Biden administration clearly wants to steer money from corporate bounty to working families. Just one example is its proposal to lift the corporate income tax rate from 21% to 26.5%. The extra revenue envisaged in the $3.5 trillion measure would in part flow from earnings to such programs as the child tax credit and free pre-K education. The administration’s regulators also take the position that today’s tech Goliaths exploit their dominance to quash rivals. It’s a stance embraced by FTC chief Lina Khan and Tim Wu, who heads technology and competition policy in the White House. “They reflect what’s called ‘hipster antitrust policy,’” says Brightman. “They believe that much of Big Tech is generating excessive, monopoly profits.” The administration is exploring ways to fix the “problem.” One tool is to break up big players that allegedly undermine competition, another to curb earnings by raising taxes. “The ‘unseemly’ level of corporate profits is becoming a big political issue,” says Brightman.

How far will the stock market fall, and when?

Answering that question requires two predictions. The first is forecasting the P/E ratio likely to prevail going forward. Since 1988, the median S&P 500 multiple is 21.75. By the way, that’s far above the 150-year average of around 16. But low real interest rates are likely to persist, justifying rich P/Es going forward. The CBO is predicting an inflation-adjusted rate on 10-year Treasuries gradually rising to 1% over the next several years, then going higher. We’ll take a conservative position and posit that the long-term real rate settles at 1%. We’ll assume that investors will want an “equity risk premium” or return that exceeds that number by three points, near the extra margin delivered over long periods.

I’ll spare you the math, but that outcome would put the future P/E at 25, we’ll above the 21.75 average over the past quarter century. To be sure, the real rate in five years could be a lot higher than 1%, and if that’s the case, the P/E would be a lot lower. Even at real rates near zero in Q4 of 2019, the S&P 500’s PE was lower at 23.2. But I’ll stick to an optimistic 25 as our benchmark for the market multiple.

Second, we need to predict where earnings will settle once the current liftoff ends and inevitably reverses. I’ll take the “soft landing” view that EPS won’t shrink in current dollars, but simply go flat, so that the number will fall each year in “real” terms by the projected 2% rate of inflation. In that case, earnings would register the same $189 a share in late 2026 as today. Once again, I won’t go through all the numbers. But the “present value” of $189 per share 2026 is $148, or 26% lower. At a 25 PE, a “fair value” for the S&P would be 3700. So at the market peak of 4537 on Sept. 2, U.S. big-caps were 23% overvalued. The over 5% decline since then to 4292 at mid-afternoon on Oct. 4 means they’re now 16% overpriced. That’s short of a bubble, but it’s the opposite of the sunny story that Wall Street is selling.

The good news: Though the overall S&P should decline, it’s the pricey growth stocks that will take the hit. Investors who concentrate in such value sectors as financials, energy, hospitality, and airlines could do fine.

As Brightman says, the part of the picture that’s in a bubble––the frenzy in corporate profits––has already gone on much longer than expected, and could inflate still more. The haymaker for the bulls: The more outsize earnings become, the bigger the target they raise for regulators, and the more workers will cry foul. But the most powerful force will be plain old competition, the one that the “hipster” regulators claim is lacking. Even if they don’t break up Big Tech, rivals will find a way to take big chunks of the sumptuous profits from the tech giants that flourished in the pandemic.

I frequently spoke with the legendary economist Milton Friedman before his passing––including on the subject of profits. I’d call Friedman, and if he didn’t answer, leave a message with his assistant. The octogenarian monetarist, who knew the value a dollar from all angles, would phone back collect. I’d hear the operator intone, “Will you accept the charges from Milton?” One time, I asked him if super-high earnings could last. “No,” said Friedman. “Profits cannot exceed their traditional levels of GDP for long periods.” It was the animal spirts, the capacity for disruption built into America’s industrial machine, that ensured when the spoils got too rich, as they are today, hungry rivals would raid the rulers’ castles. No, the world hasn’t changed as Cathie Wood suggests. It just goes bonkers sometimes. But as Friedman reassured me, the competitive dynamic, the drive to do it better, cheaper, and faster, always wins. And it will reign again.

Dive into stories from Fortune’s print edition:

• Roz Brewer on what it feels like to be 1 of 2 Black female CEOs in the Fortune 500
• Can new CEO Fidji Simo turn Instacart into more than just a delivery company?
• Are women on a collision course with the COVID ceiling?
• Hard right turn: Companies are toeing the political line in business-friendly red states
• These 9 powerful women are ones to watch