In the debate about whether the US stock market is currently overvalued, analysts have any number of valuation metrics to choose from. Most of them give comparable results (expensive, but not a bubble), so it surprising to see Goldman Sachs argue that it the current multiple is just 18.1x compared to a historic average of 17.2x (for comparison, the S&P 500’s Shiller PE ratio is currently at 24.91x compared to a historic average of 16.52x).

normalized S&P 500 earnings 0414

shiller pe ratio 0414 S&P 500

“Normalized valuation looks less stretched than other metrics such as P/E, cyclically adjusted P/E, and our macroeconomic model,” write Goldman Sachs analysts Stuart Kaiser, David J. Kostin, Amanda Sneider, and Ben Snider in an April 10 report.

S&P 500: Normalized meant to show long-term trends, much like CAPE

The idea behind normalized PE is the same as for CAPE – you want to see the underlying value of a company, not just what’s happening today. While CAPE does that by looking over a full business cycle, normalized PE assumes an underlying growth rate to the market and then uses that to adjust earnings. But this gives analysts an enormous amount of room to maneuver and push results up or down according to their biases. If the normalized PE is being calculated over multiple years, then the analyst also has to decide whether the underlying growth rate is constant or if they should choose a different rate for each year.

S&P 500: Normalized PEs can wash out important differences

Maybe normalized PE is useful for identifying certain types of trends, but it’s hard to imagine using it as the basis of investment decisions. In a pre-crisis article on Seeking Alpha, value investor Geoff Gannon walks through a typical example of using normalized PE to smooth out changes in the Dow Jones Industrial Average (INDEXDJX:.DJI), assuming an underlying growth rate of 6%. The result differs so wildly from actual results that it’s hard to imagine that it’s actually that useful (and this is from someone who thinks normalized PE ratios are a good tool).

In Gannon’s example, the percentage difference between the normalized earnings and actual earnings swings between -50% and more than 60%. CAPE valuations are popular among value investors because they smooth out temporary effects, but it seems like normalized PE ratios can wash them out entirely. The Goldman Sachs team uses a 15.5% ROE for the S&P 500 (INDEXSP:.INX), which is probably reasonable, but saying that the market is fairly priced by this metric doesn’t seem to tell us that much, since it is practically designed to pull valuations to the long-term average.

percent diff DOW norm actual 0414 S&P 500