As usual all the web sites are worthless and all the "real estate professionals" are morons, they just spout the bullshit that was in their training packet that says something about "points pay themselves off after N years so it depends on how long you will live there" ; great, thanks.
I tend to make mistakes on this stuff, so you can check me and let me know if I'm right in these calculations.
The big thing that everyone seems to fuck up is that there are two different factors changing the value of money :
1. inflation is making the mortgage payments actually cost less than they seem to, that is, later payments actually cost you much less in today's dollars, and
2. opportunity cost market appreciation is making points cost more than they seem to. That is, any dollar spent on points could go in the market and earn something (presumably slightly more than inflation).
Running these numbers :
30-year fixed rate loan 417k loan amount (maximum conforming loan) 4.50% base APR 0.125 APR per point points cost 1% of loan inflation 3% annual market appreciation 1.5% after inflation (same as APR) The total amount paid after 5,10,20,30 years , with 0.0-1.5 points : (in today's dollars)
The normal numbers that you see everywhere don't include inflation or appreciation and look something like this :
Which lead you to believe that points are actually a big savings after 15 years or whatever. In reality it seems to me that they hurt you up front and basically never make it back. (after 30 years they finally do, but the liquid investment has added utility which makes it still the winner)
As usual the exact right answer depends highly on the assumed figures, and there's huge uncertainty about them. However I think we can reliably say that points are grossly over-valued in the popular literature.
The next question is - if I have some extra cash, should I pay more down payment or put it in the market? (more down payment is equivalent to early prepayment, so this is equivalent to asking if you should early prepay).
Same assumptions as above. Extra cash available is 0.5-2.0% of loan amount Note : rows of this table are not directly comparable Blue columns : extra cash put towards reducing loan amount Black columns : extra cash saved and appreciated (counted as reduction of amount paid after N years)
Keeping the cash in the market is better over all time scales.
What if the market only matches inflation and doesn't beat it by that 1.5% ?
Now over 30 years paying off the loan early is better. But again it's so minor that if you count the utility of liquidity it just never makes sense.
One exception to all this would be if the market actually inverts for a while (returns less than inflation, or negative inflation-adjusted). It seems to me that is a real valid concern over the next 10-20 years, but when that happens pretty much all the "common wisdom" goes out the window.