Over at The Big Picture, Barry Ritholtz makes this point.

On a trailing one-year basis, that puts the Price to Earnings Ratio (P/E) at over 19 as of today.  This does not make the market cheap.

And what about 2009? Again, the analysts are in a race to find the bottom.

The current projections are for $42.26 for 2009. That makes the forward P/E 22. That doesn’t look like value at all, when the historical average is closer to 15.

In 2001, as-reported earnings were $24.67. Operating earnings in 2002 were $27.57. Does anyone think the current recession will be milder than the last one? Or shorter?

Is the market expensive at these levels?  Does it matter? 

I will get to the first question later.  As for the second, a better question is why would an investor value the market off cyclically low earnings?

The market bottomed in 2002 on October 10 at 768.63.  Trough earnings were $24.67.  Thus, the market bottomed at 31.1x trough earnings.  Earnings in the third quarter of 2002 were actually higher at $30.04, or 25.6x trailing 12 month earnings.  So, in fact, the market trading at 22x depressed earnings today looks cheap compared to the last bear market!

The reasons why an investor would view the market as expensive and thus avoid stocks at these levels are twofold - earnings are going to go a lot lower and stocks will follow or earnings are going to stay permanently depressed.

First, there have been occasions when the market bottomed at 8x trough earnings.  Is it prudent, however, to wait for cyclically low multiples on cyclically low earnings before buying?  As I explained last month, the answer is generally no as extreme valuations are rare.  Instead, a better strategy - in my opinion - is to buy as the market approaches an extreme low valuation and sell as the market approaches an extreme high valuation.  You can never catch the bottom or the top - or at least I cannot - but buying and selling as the market approaches extreme levels is generally a profitable strategy over time.

As for earnings staying permanently depressed, well, perhaps, but I would not bet on it.  Margins have averaged about 6% over the past two decades.

Applying a normalized profit margin, is the market expensive?

I use 7% as a normalized profit margin.  This is higher than the long-term average but in the past, lower-margin hardware companies accounted for a higher proportion of technology profits than higher-margin software companies.  Today, software companies generate a higher proportion of technology profits. 

Based on this measure of normalized profit margins, stocks look cheap.

On other measures, the market also looks inexpensive.

The price to sales of the market is approaching multi-decade lows ...

... while the price to book of the market is at multi-decade lows.

Even though I am very long stocks, I am skeptical that we have entered into a new bull market.  However, stocks are not expensive.

EDIT - I am currently reading Anatomy of the Bear and came across this line about the 1921 bottom in stocks on page 67.

The bulls of 1921 focused on the value in shares an key technical factors which suggested that market prices would recover.  There was little discussion on the immediate earnings outlook.  While the market bottomed in August 1921 reported earnings had a further 37% to decline before they reached bottom in December 1921.  An earnings recovery significantly lagged the recovery in the stock market. [Emphasis added.]