High Valuations Aren’t Anything to Fear

With a booming 2024 now winding down, some pundits are starting to get fearful of heights based on lofty valuations. Across the board, many say price-to-earnings (P/E) ratios and other metrics are elevated. Yet valuations alone don’t—and can’t—predict where stocks will go, as history shows. High valuations can always go higher (and low ones lower).

Fronting the police lineup of allegedly worrisome valuations are the usual suspects: Trailing, cyclically adjusted and forward P/Es. As Exhibit 1 shows—and many pundits are pointing out—the S&P 500 sits at 31 times its last 12-month earnings, which approaches April 1999’s then-record high 34x (dashed red line). Given that mark preceded March 2000’s bull market peak, pundits sense foreboding now.

Exhibit 1: Trailing P/Es Don’t Foretell Stocks’ Future

Source: Multpl.com, as of 12/10/2024. S&P 500 price index and 12-month trailing P/E ratio, January 1926 – November 2024. Left-hand y-axis in base-2 logarithmic scale, which plots the same-sized percentage moves in equal increments graphically.

While stocks were arguably overvalued when the 2000 – 2002 bear market began, and spiking valuations were an indication of this, they weren’t—and aren’t—a timing tool. Turning bearish in April 1999 would have left hefty returns on the table, and staying bullish the next March because valuations were then falling would also have been folly.

Then, too, what about March 2002’s 47x, May 2009’s 124x and December 2020’s 39x? Those all occurred well into recessions when earnings wipeouts caused multiples to soar—and in the case of 2009 and 2020, sky-high trailing P/Es happened early in a new bull market. Forward-looking stocks didn’t care about past profits, which is why there has never been any particular trailing P/E that corresponds with bear (or bull) markets. Sharp P/E spikes can be telling about sentiment, but that is about it.

The even more backward-looking cyclically adjusted P/E, or CAPE, makes even less sense, as we recently covered. Rather than the last 12 months’ earnings, it uses the prior 10 years’ worth—and then bizarrely adjusts them for inflation. Averaging a decade’s earnings supposedly smooths (or “cyclically adjusts”) them so that past collapses—like 2002’s, 2009’s and 2020’s—don’t feature as prominently. But notice that when those earnings’ wipeouts roll off after 10 years, it boosts the denominator—arbitrarily, in our view. Why should decade-ago earnings influence today’s measure or tomorrow’s?

Then, by adjusting earnings for inflation—but NOT stock prices—CAPE skews the ratio higher. Not only is this inconsistent, it is unnecessary. Stock prices’ incorporation of inflation is a feature, not an artificial enhancement. Nominal stock prices typically navigate inflationary periods fine given corporations’ general pricing power, i.e., they can usually raise prices to cover costs. Deflating the denominator only casts stocks’ resilience as a flaw. To us, it is just conceptually strange.

What remains is a broken measure of “value.” Exhibit 2 shows CAPEs when bear markets begin range from 9x in 1980 to 44x in 2000. In recent years, CAPE levels overall drifted up—spending almost the entirety of the past three decades above the long-term average of 19x. If it seems like CAPE levels don’t matter, it is because they don’t.

Exhibit 2: Stocks Don’t Need a CAPE

Source: Multpl.com, as of 12/10/2024. S&P 500 price index and Cyclically Adjusted P/E (CAPE) ratio, January 1926 – November 2024. Left-hand y-axis in base-2 logarithmic scale, which plots the same-sized percentage moves in equal increments graphically.

So what about using P/Es based on analysts’ consensus expectations for earnings over the next 12 months? As Exhibit 3 shows, the S&P 500 at 22x forward earnings is still below December 1999’s 24x and August 2020’s 23x, but not by much. While we think this is a better approach conceptually—since it attempts to look ahead, like stocks do—it nevertheless faces a similar issue with trailing figures: Even forward estimates can’t say what actual earnings will be. We find forward earnings expectations tend to reflect prevailing sentiment, whereas stocks move mostly on the gap between those expectations and reality.

Exhibit 3: Forward P/Es Aren’t Telling, Either

Source: FactSet, as of 12/10/2024. S&P 500 price index and 12-month forward P/E ratio, 1/1/1996 – 12/9/2024. Left-hand y-axis in base-2 logarithmic scale, which plots the same-sized percentage moves in equal increments graphically.

Hence, the range of forward P/Es when bear markets begin is also pretty wide: from 24x in 2000 to 15x in 2007. 2020’s bear market, though brief, started at 19x. Not because that level was significant in any way. That is just what it happened to be when earnings expectations adjusted suddenly to lockdowns’ new reality.

Lastly, to show how far afield valuations can stray from stocks’ core driver, let us dissect the Buffett Indicator, which compares US stocks’ entire market capitalization to current-dollar (nominal) GDP. As Exhibit 4 shows, stocks have hit a new valuation record on this score. America’s market cap is now over 2.0 times nominal GDP, exceeding Q4 2021’s prior high, and much higher than Q4 2019’s 1.5x and Q1 2000’s 1.4x. Those readings preceded bear markets. But bear markets have also begun at 0.6x (1987), 0.5x (1990) and 1.0x (2007).

Exhibit 4: The Buffett Indicator Isn’t Too Oracular

Source: FactSet, as of 12/10/2024. Ratio of Wilshire 5000 Total Market Capitalization Index to nominal GDP, Q1 1985 – Q3 2024.

The reason there is no magic level: GDP provides only backward-looking snapshots of broad economic activity, much of which is irrelevant to stocks, which confer owners a slice of publicly traded companies’ future profits. GDP is also a flow indicator, imperfectly measuring activity in a given year. Market cap is a value or accumulated metric, share price times outstanding shares. Comparing the two is a stock-flow mismatch. GDP also comes out at a late lag, while stocks lead economic growth, typically turning up before recessions end and down ahead of contractions.

Buffett indicator peaks may seem to coincide with bear markets, but that is just because market caps were already shrinking—in part anticipating economic downturns. Meanwhile, new highs—or apparent peaks—don’t necessarily translate into bear markets (or recessions). After all, past prices—or market caps—never predict future ones.

In sum, valuations aren’t a reason to sell (or buy). Examining their underlying premises—and the historical evidence—shows they aren’t good guides to stocks’ path forward.