The S&P 500 just hit a dangerous benchmark—recalling the peak of the 2000 dotcom bubble

At first glance, a price/earnings ratio of 30 might not seem alarming. In June of last year, the basic earnings multiple—using the standard measure of trailing four quarters of GAAP net earnings—soared to 31, when shares were 26% cheaper than today. And at the end of the fourth quarter of 2020, the P/E stood at almost 40. That number hasn’t bothered the bulls at all, since the S&P has partied since then. But those pandemic-economy ratios are pretty meaningless in gauging whether stocks are expensive, cheap, or fairly valued. Although prices romped, it was the one-third earnings collapse from the fourth quarter of 2019 to the end of 2020 that most inflated the most famous and fundamental of valuation measures.

It was always clear that profits would rebound from COVID-stricken lows, increasing the denominator and putting that 40-plus PE in a more realistic range. To get an accurate reading on where stocks stand on the spectrum of super-pricey to bargains, and hence handicap where they’re headed, you need a yardstick for “normalized profits.” I define that level as the number the S&P companies collectively are most likely to post once the recovery takes hold. My choice is the Q4 2019 earnings of $139.47 per share, based on the total for the year.

I made that pick for two reasons. First, 2019 profits not only set a record, but also stood extremely high by historical standards when compared to sales and other measures. That reduces the chance that our normalized P/E is distorted upwards because we’re underestimating much higher sustainable earnings in the pipeline. Of course, the bulls will argue that our 30 multiple is way out of whack, because 2021 profits will dwarf even big-caps’ super performance in 2019.

Second, even in the strong economy of 2019, earnings per share (EPS) from quarter to quarter were essentially flat, for a basic reason. Profits had reached such high levels versus fundamentals that earnings lacked room to run. Even now, Wall Street analysts, who are typically ultra-optimistic, project annualized profits through the second quarter of 2021 that match the $139 registered at year-end 2019. Put simply, rescaling that summit is a likely scenario for profits in 2021.

Historically, the S&P multiple has exceeded 30 only at times when earnings dropped sharply or collapsed, with one exception prior to today. In the December quarter of 1999, the P/E reached 30.5 near the peak of the Internet stock craze. That number flashed red, because it came on top of the highest EPS ever recorded. In other words, investors were paying a gigantic $30 for every $1 in profits, and those profits were already elevated by historical standards. But EPS marched on, striking a new high of $53.70 in the September quarter of 2000.

As it turned out, EPS wouldn’t beat that mark again for four years. Those record profits coincided with an all-time peak for the benchmark index, which reached the summit of 1518 on Sept. 1, 2000. At that point, the index had surged 120% in four years. So the best single parallel to where the market stands today is its status three days before Labor Day in 2000. Once again, those are the only two times the S&P 500 achieved never-before-witnessed profits and prices at the same time. (Of course, that’s using our choice of 2019 profits for the comparison to today.)

On that date more than two decades ago, the S&P’s traditional multiple hovered at 26.75. By that yardstick, the index is 12% more expensive today at a P/E of 30. But the multiple doesn’t provide the full picture. Say that according to fundamentals at the time, EPS was a lot higher in 2000 than in 2021. In that case, the multiple could be understated because profits were bound to fall, and today’s P/E is overstated because EPS is still modest, presaging big increases to come.

The evidence is somewhat mixed, but on the whole, profits as of Q4 2019 appear at least as high, relative to the period, as at the apex of the tech bubble. For example, earnings for the full year of 2019 were equivalent to 5.4% of national income, far above the 4.4% for the four quarters ending in September 2000. As for operating profits as a share of sales, or “operating margins,” the S&P recorded 11.1% in 2019, beating the number in 2000 by around four percentage points. By comparison, the turn-of-the-century index did better in return on equity, garnering 17.5% to 2019’s 15.5%.

Since most but not all of the metrics point to higher relative profits in 2019 than 2000, we’ll take a conservative stance and rate them about equal for both periods. We’re left with a side-by-side look at the only two periods of all-time high earnings coincided with record valuations. What does this mean for today’s big-caps?

Bad news behind the boom

The boom that brought the P/E of 30 is great news if you’ve been riding the train. But it’s the worst possible tidings if you’re planning to jump in now, or want to leave your winnings in U.S. big-caps. The evidence: Look what happened in the period following Sept. 1, 2000. The S&P entered a long swoon, bottoming at 801 on March 11, 2003, for a loss of 47%. The index didn’t regain the September 2000 mountaintop for seven years, until September 2007; prices plunged again, of course, during the financial crisis, before climbing back above 1500 in March 2013. That adds up to over 12 long years of booking zero gains except for paltry dividends.

Of course, it’s impossible to predict whether we’ll experience a drop of anything like the aftermath of the dotcom catastrophe. The bulls forecast that two powerful tailwinds will converge for a full-rigged voyage to glory: fast-rising profits that will surmount the feats of 2019, and the support of super-low interest rates for years to come.

Both are possibilities, but unlikely. And benefiting from the confluence of both is an even longer shot. Let’s look first at the outlook for profits. At 5.4% of GDP, earnings are already outsize by historical standards. For EPS to keep outpacing GDP, as Wall Street predicts, profits would have to get even more outsize. It’s especially worrying that on April 16, the S&P’s valuation as a share of national income hit an astounding 161%, one-quarter above the 130% it reached in the 2000 rampage, and almost double the two-decade median.

To stay at 160%-plus, earnings would have to devour an even bigger share of the economy at a time when wages, depressed for years, are finally on the rise, the U.S. is facing long-term growth rates at below 2%, and corporate income taxes are likely to rise significantly as a prime source of revenue for President Biden’s ambitious spending agenda.

As for interest rates, the best response to the bull case that the 10-year Treasury yield will trail the trajectory of GDP in the years ahead is basic: It has seldom happened before. “Real” or inflation-adjusted rates normally track national income, and cases in which they’re in negative territory have never lasted long in recent decades. The most logical explanation for today’s sub-inflation yields is that they’re a rare phenomenon caused by COVID’s once-in-a-millennium blow to the economy.

Still, many seasoned investors believe rates will remain low for a long time. Louis Navellier, who runs a $1.8 billion portfolio for his family office, is a sober, insightful student of the markets. “We were hearing all this talk of inflation, and we saw stunning retail sales figures, and upward GDP revisions,” he told Fortune. “And yet bond yields fell. I can make the argument that where bond yields are now, stocks are still undervalued.”

Navellier makes a good case. I’ll stick with the takeaway from setting the only two times the S&P reached 30 on peak earnings cheek by jowl, and judging the resemblance. The aftermath of the dotcom fever ranked among the worst deluges of the past century. We’re reliving the warning signals and wild exuberance we saw then. The great times may keep rolling for a while. But for the roll to last would be a departure from history. And I’m betting on history, as chronicled by the craziness of late 2000 and the wreckage that followed.