I agree with your second point that using Siegel’s analysis to prove the market is undervalued is misplaced. I don’t think he was really making that point in the piece. His argument for the market being cheap (in other writings of his) centers around normalized earnings of $90 or something like that, which is too high for my taste. I am just using the numbers that he pointed out and they can’t really be manipulated since it’s straight math.

I think the point here is more about the impact of the financials on earnings and not the valuation of the market. Hopefully we will see what normalized earnings wind up being in 2010 or 2011.

1. I think you are overestimating normalized profit margins for the S&P 500 with a $70-80 expectation of “normal” earnings. Jeremy Grantham has made this point frequently. Margins are mean reverting, and they can’t stay above average indefinitely. Using average historical margins gets you to around $60-65 in S&P normal earnings. put a historically average 14-15x multiple on that and you get…~900, which is where GMO puts approximate fair value on the S&P at. Remember that 2008 profits for the economy as a whole, certainly for Q1-Q3 were still being juiced by massive leverage outside the financials also – this included commodities, industrials, utilities, transports, and most importantly – consumer cyclicals. I will fully admit that in the grand scheme of things, fair value of 900 or fair value of 1050 is not that big of a difference. But for the near term, if you expect the economy to generate below average profit margins for a long time due to deleveraging not just the financial sector (as opposed to a steady state level of leverage, which will allow us to get back to average profit margins) as I do, there is no rush to put your money to work.

2. Siegel is being extraordinarily disingenuous with his analysis. Remember that S&P earnings have ALWAYS been calculated the way they are today. Given that valuation is an inherently comparative activity, you can’t compare the S&P valuation using Siegel’s method to the past because no one else ever used that method in the past. If you put together a time series of Siegel’s metric, you would probably come to a very different conclusion – that while the market is definitely undervalued, it’s not massively undervalued like Siegel has been arguing for a long time.

A better way of realizing that Siegel’s argument holds little water is historical dividend yields for the index – there’s no way to manipulate that. It’s the cash that ends up in the investor’s pocket holding the market basket. By those standards as well, the market is moderately cheap (10-15% undervalued) but not incredibly undervalued.