This comes up often in the Mortgage Banking Industry, hence you get the MBA being a primary impetus on getting guidance on this topic. It's a bit funky, but I this is my understanding as well. The main issue here is that for the starting adjusted rate, there may be "price adjustment" fees or whatever to get to the starting rate, but for the base rate in such deals, the borrower is not actually paying specific amounts to reduce the interest rate. So, at this point in time there are no actual discount points, as points paid correlate with a decreased interest rate. So, there really are no "discount points" existing with the base rate. If you have the base loan, the borrower hasn't specifically negotiated to pay discount points to decrease the rate. He/she is just getting the only product available to them. If they want to decrease the rate further, then that is when they are actually paying discount points.
I have a couple of issues with the calculations presented though. I would treat the APOR/APR spread calculation in line with the statement above. So, for me, the starting adjustment rate does not include a "discount" at all. I would compare the APOR to 4.125% in the above example. The what they have stated it, the APR/APOR calculation guidance contradicts the Excluded Bona Fide Discount Points guidance. If the initial discount to get to the starting rate is to be considered, then it should be eligible for exclusion as well, if they really believe that the borrower has "paid" to decrease the rate. You can't have it both ways, imho. Also, the discount point paid is generally construed to have to equate to a .25% reduction in the rate. If the borrower paid 1.625 points, then the rate should be way lower. I would consider only .75 points eligible for exclusion here, as there was only a 3/8 percentage decrease in the interest rate after the discount points paid (4.125% to 3.75%).