S&P earnings are very depressed due to the lack of financial earnings, which were zero or negative last year but in prior years were the biggest earners in the S&P 500. And I disagree with his using an inflation adjusted average, as earnings historically have grown at the faster nominal GDP growth rate. So I think Shiller underestimates it. Price to sales can be a more pure estimate. Price too book as well, but in recent years has grown obscured due to all the goodwill that has piled up on corporate balance sheets.
Longleaf has a recent piece that compares the P/E to risk free rates, check their website. The fed model would compare earnings yields on stocks to those on bonds. On any measure stocks look very cheap relative to treasuries, less so compared to corporate debt.