No Consensus on U.S. Stock Valuation

[HenryTo]

Interesting article on the various different predictions on where the S&P 500 "ought" to be. This is also the first time I've seen Ed Yardeni "articulate" his preferred model of gauging where the S&P 500 should be. Note that if the P/E ratio should decline to 16, we will see a 15% to 20% correction in the markets given no earnings growth - and approximately a 5% to 10% decline assuming 10% earnings growth (which is pretty aggressive) over the next 12 months.

Finally, note that Ed Yardeni's model NATURALLY assigns a highest "fair value" P/E ratio of 20.
------------------------------------------------------------------------------------
Quote: "Yardeni said his preferred model is taking 20 and subtracting the yield of 10-year Treasury bills to get the market's preferred P/E. Under that model, the P/E of the S&P 500, calculated using last year's earnings, should be roughly 16. Instead, it's about 19.5."

No Consensus on U.S. Stock Valuation
Sunday July 10, 1:24 am ET
By Ellen Simon, AP Business Writer
Underneath the U.S. Market's Placid Surface, There's Stubborn Disagreement Over Stock Valuation

NEW YORK (AP) -- Are stock prices too high, too low, or right where they should be? In the sometimes quarrelsome confines of Wall Street, there's no consensus.

"Valuation is like beauty -- it's in the eye of the beholder," said Ed Yardeni, chief investment strategist at Oak Associates.

Stocks have traded nearly flat for most of the year, reacting mildly after Thursday's terrorist attacks on London, the summer's increasing oil prices and the year's steady march of Federal Reserve interest rate hikes. If these events haven't pushed stock prices dramatically higher or lower, an observer could be forgiven for thinking the current prices must be correct.

But underneath the market's placid surface, there's stubborn disagreement about whether stocks are superb bargains or overblown.

"Wake up, equity market! I recommend 100 percent stocks," Edward Keon, chief investment strategist and director of quantitative research at Prudential Equity Group, wrote in a research report Tuesday, arguing stocks look "very cheap" compared to bonds.

Not so, says James Stack, president of InvesTech Research in Whitefish, Mont. "From a historical perspective, stocks are approximately 20 percent overvalued."

Who's right? Who knows. What we see is how investors gauge valuations, deciding whether individual stocks -- and the moment's markets -- are worth buying or selling.

Of course, if anyone had a magic model for valuing stocks, there would be no dispute over valuation and no risk in the market. So, we proceed with caution. "There have been countless attempts to develop models to determine whether the market is overvalued or undervalued that have been off the mark, sometimes wide off the mark," Yardeni said.

Some investors track the price of stock versus the price of gold, calculating how many ounces of gold you'd need to buy the Dow Jones industrial average. (You would need about 23.37 ounces at current values.)

Or, they compare the price of stocks to the Consumer Price Index. By this measure "we're at the peak of a cycle," said Gibbons Burke, editor of Markethistory.com, a financial research Web site. "We're looking at a big decline over the next 15 or 20 years. It's not great news, but that's what the chart says."

The standard metric for valuing individual stocks -- and markets -- is price-to-earnings ratios, or P/Es. The problem: Investors disagree on what a fair P/E is. In a report released Tuesday called "Nobody knows anything: Thoughts on equity market valuation," Keon wrote, "By now, shouldn't there be more widespread agreement on what constitutes a fair market P/E?"

Those who want to buy only when markets are cheapest look for P/Es below 15.7, the historical mean. "They love the market when the P/E is 8 to 10," Yardeni said. "The only problem is, (at that point) the rest of us are unemployed or worried about being unemployed."

Comparing the current P/E ratio for the Standard & Poor's 500, which is about 19.5, to the historical mean P/E is how InvesTech's Stack determined stocks were 20 percent overvalued.

Stack also watches P/Es in relation to interest rates. Higher valuations tend to be supported by lower interest rates, Stack said.

"The last two times we saw a single digit P/E for the S&P 500 were 1982 and 1974, both instances where short-term interest rates were 8 percent or higher" he said. When interest rates increase, money market funds and bonds become more attractive.

"Today's low level of interest rates are providing an important buoyancy to the market and to market valuations," he said.

Another popular model is called the "Fed Model," because the research behind it was done by Federal Reserve staff and because believers think it's something Fed Chairman Alan Greenspan considers as he soaks in the tub. The Fed Model takes the ratio of stocks' earnings-to-prices, which is their "earnings yield," and compares it with the 10-year Treasury bond yield. Under the model, when the earnings yield is higher than the bond yield, stocks are a bargain.

Fed Model computations were part of the reasoning behind Keon's "100 percent stocks" report. Keon computes his earnings yield using analysts' projection of earnings. When figured this way, the Fed Model is usually bullish on stocks, and Keon's computations are no exception. He finds the earnings yield for stocks is around 6.6 percent, while the yield on the 10-year Treasury bill has been hovering around 4 percent.

"Stocks look as cheap compared to bonds today as they looked expensive in early 2000 at the height of the equity bubble," he wrote. "Either stock valuation should rise or bond prices should fall or a combination of both, according to the model."

Yardeni said his preferred model is taking 20 and subtracting the yield of 10-year Treasury bills to get the market's preferred P/E. Under that model, the P/E of the S&P 500, calculated using last year's earnings, should be roughly 16. Instead, it's about 19.5.

"It hasn't been a perfect model relative to what the stock market has done, but it's been pretty good," he said. "I can't tell you why it works, but it shows why forecasting is so frustrating. The simple model is as good as any of the sophisticated models I've seen."

[nodoodahs]

I would think a "good" model of equity market PE should include at least:

1. Specific definition of the "market" be it S&P 500, DJIA, Wilshire 5000, etc.
2. Consideration of alternate investment yields, including treasuries, corporates, real estate, precious metals, commodities, and foreign equities
3. Consideration of inflationary expectations which might be included in 2. above
4. Consideration of recent monetary policy i.e. inflation properly defined.

Since GDP and CPI are moving targets (i.e. revised multiple times) any model that uses them is suspect IMO unless they reconstruct past figures to current calculation, or vice versa.

The argument for using projected earnings is that "past earnings aren't reflective of future performance." My response is that analyst's projections aren't reflective of future performance, either, so I'd rather use something that exists in reality, like TTM earnings. YMMV.

Comparing current PE to historic may not be germane as various items listed above could be different today from the other times S&P 500 PE was around 20. Consider that several historic charts of PE go from days of US dollar convertibility to gold for everyone, to convertibility for foreign governments only, to days of pure fiat. Without consideration of monetary policy a comparison of these different time periods may be flawed ...

I've got a model (which isn't "good" by my own definition LOL) that suggests today the market is undervalued. BTW it does show overvaluation from 1996-2000 in backtesting.

To those of us investing in individual stocks, the market's valuation is moot for the most part. I prefer to focus on market instability for that reason.
_________________
I haven’t seen a beatin’ like that since somebody stuck a banana in my pants and turned a monkey loose.

[Gizmo]

One of the best primers on p/e with charts from Carl Swenlin.

http://www.decisionpoint.com/TAC/SWENLIN.html
_________________
Gizmo

[Gizmo]

Still high by historical standards. Dips under 10 are the ultimate buy area. Given current p/e's anyone buying for the long term here will likely be dissapointed in the return. P/E expansion as in 95-99 not likely IMO.

http://www.generationaldynamics.com/ww2010/peratio.gif

In edit, another look shows that every time p/e's rose above 20 they subsequently dropped below 10.
_________________
Gizmo