Greg Mankiw's Text... Keeps Getting Worse

On page 230-231, Mankiw calls it a "fact" that the Marginal Product of Capital is 12%. Thus, Manks says, the capital stock of the US is well below the golden rule level assuming economic growth of three percent and depreciation of four percent. Except, he could have chosen basically any number for depreciation and any number for MPK. The long-term average of the Dow is definitely less than 12% -- probably closer to about 6%. (The past 10 years, the stock market has returned closer to 0%...) I doubt too many firms in the US borrow at anywhere near 12%... The US government is now borrowing at basically 0%...

Of course, there are other problems with the whole Solow analysis, as I've pointed out -- namely, much investment in the US goes into housing, which isn't used in production (and depreciates more slowly than computers do...). Secondly, much of consumption is used later to produce -- such as consumption expenditures on food, health care, education -- so the whole concept is not that meaningful.

A second thought: The whole notion of "capital per effective worker" is both tricky for students to understand and adds nothing to the analysis of the Solow model. Why not drop it in favor of something more interesting? Like Kamarck/Crosby/Diamond/Sachs?