What's missing from this argument is that the lander is taxed on the rental income (after the deductions mentioned) but the homeowner is not taxed on the imputed rental income from the home.
The real tax break that homeowners get is not the interest deduction, but the fact that the imputed rental income (what the home would rent for) is not taxed.
This is because I did not write clearly, not because I hadn't considered it. Bill Wheaton's point was that the deduction for depreciation for rental units (about 3.3 percent per year) offsets the non-taxation of imputed rent. Most owners who get lots of net income from imputed rent are the elderly, who have paid off their mortgages. Because they generally have low cash incomes, they are in low tax brackets, which means that the value of the tax benefit are small. Owners of rental units are often in higher tax brackets, which means the value of the depreciation deduction might be high.
How it all washes out is an empirical question, of course...