The FED MODEL doesn't make much sense to me either.
If I remember correctly, Ed Yardeni "discovered" the model in some fed document and continues to promote it even today. It graphs the yield on the 10-Year Treasury Note vs the earnings yield (E/P or reciprocal of the price to earnings ratio, P/E) of the S&P500. Whichever has the higher yield is the better buy.
So what's wrong with that?
The note/bond is a nominal asset (the yield already includes an adjustment for future inflation). The earnings "coupons" on the stock aren't static, they are expected to have both inflationary and real growth over time. The Fed model thus favors the note/bond in a comparison by ignoring the growth inherent in a stock's "coupon".
The risk on the two assets are very different. Besides the obvious Treasury vs Corporate credit risk, there is duration. This is a bond term but it can also be applied to stocks as well. Duration is the answer to the following question: "How much money will I lose (relative, as a percentage of the total) if interest rates increase by an absolute 1%?". The note/bond has a duration of less than 10, the stock, I believe, is something greater than 50 currently. Stocks are risker (duh!). Given the same total return wouldn't you take the one with the lower risk? In this case the Fed model favors stocks by ignoring their inherent risks. Duration is a tough subject for me, it involves calculus and first derivatives :^), I'm still working on it.
Actually for stocks it is the increase in the REAL not nominal interest rate that matters. If interest rates rise solely for inflationary reasons stock prices like TIPS (inflation protected notes/bonds) should not be effected. On the other hand, if Mr. Market should suddenly demand a higher REAL total return then both stocks and bonds (nominal & real) will get be trashed!