During the whole bull run from March 2009 until May 2015, the bulls kept telling us equities were cheap, based on their favorite measure: the price-to-earnings (P/E) ratio. Their argument was that even though the S&P kept ascending, P/E ratios looked relatively modest versus their historical averages. That argument never made much sense to anyone not looking for a rationale to sell folks more stock, whatever the price. But it passed for science on Wall Street.

Today, the folly of using the current P/E to judge whether stocks are cheap or pricey is being unmasked, in a pretty spectacular fashion. How can P/Es be a great measure of value when they’re higher now, after the big selloff, than at their peak?

In the quarter ended September 30, 2014, S&P 500 earnings per share for the trailing 12 months reached their all-time peak of $106. At that point, the S&P’s P/E multiple stood at 18.6. The bulls argued that stocks were still reasonably priced, since they kept predicting that profits would rise briskly from those already elevated levels. The problem was that earnings were already in a near-bubble, measured as a percentage of sales, GDP, or any other metric. The inevitable trend was not up, but down.

And that’s what happened. From the fall 2014 earnings summit to the market’s May 2015 zenith, profits dropped by 4.8%; but over the same period, the S&P kept rising, reaching an all-time record of 2135 on May 20, 2015. The disconnect between falling earnings and rising prices pushed the P/E to 21.2. At a number that rich, it was hard to argue that stocks were a solid buy. So the Wall Street crowd trotted out a different metric, claiming multiples were ‘reasonable’ when calculated on analysts’ forecasts of profits for the next twelve months. Since analysts’ estimates are always inflated, and even more outré than usual at the peak, the forward P/E measure turned out to be what it’s been historically: worthless.

Since cresting last May, the S&P has dropped to as low as 1810 and was standing at 1926 midday on February 17, a nearly 10% drop from the peak. But profits fell a lot more than 10%, shrinking 18.4%. So the market’s P/E actually rose from 21.2 to 22.3. So if you believe in current P/Es, stocks are 5% more expensive now than when their prices were 10% lower.

Now, the bulls are in a bind. They’d like to argue that stocks are a fantastic buy after such a sharp correction, especially since they were supposed to be a bargain before they sank. But their favorite yardstick is showing just the opposite.

The lesson is that earnings are so volatile that they’re constantly distorting P/Es. The inflated profits of recent years made shares look modestly valued when they were actually overpriced. That’s why it’s so important to use a measure that averages earnings over a span of several years or more. The Cyclically Adjusted Price-to-Earnings ratio (aka CAPE, aka the Shiller P/E) will give a much better picture of whether shares are a buy or sell. When the peaks and valleys are removed, P/Es are extremely useful. Unfortunately, that methodology predicts that earnings are still at unsustainably high levels, and will probably keep falling. Hence, today’s “adjusted” P/E is really in the mid-20s.

Rhetorically, the bulls are a powerful force. We’ll soon see new arguments for why this recent bounce is a sign the bear market scare has ended and a great buying opportunity is at hand. Don’t believe them.