Why Are ARMs So Expensive?

A friend of mine is shopping for a mortgage right now, and I just had a very frustrating conversation with her mortgage broker. What I'd like to do is be able to choose between a fixed-rate mortgage and an adjustable-rate mortgage. If I take the adjustable-rate mortgage, I expect to pay a lower interest rate in return for taking on interest-rate risk. But it seems that the only ARMs on offer are all "teaser rate" products, where the mortgage resets to a significantly higher spread once the initial teaser period is over. And even the teaser rates, on closer examination, don't look particularly attractive compared to the fixed-rate mortgage on offer.

The broker offered three ARMs to my friend: a 7/1 ARM at 6%, a 5/1 ARM at 5.875%, and (after I asked about it specifically) a 1/1 ARM at 5.75%. All three of them, he said, reset to 225bp over one-year Libor at the end of the initial period.

The reason I asked about the 1/1 ARM, of course, was to get an idea of what happens to the spread over Libor. At the moment, 1-year Libor is 4.47%, which means that the 1/1 ARM starts off for the first year at 128bp over Libor, and then jumps all the way up to 225bp over thereafter. If one-year interest rates stay where they are, that means my friend will be paying interest of 6.72%. Even the 30-year fixed-rate mortgage is much lower than that: just 6.125%. And of course my friend would be paying well over 7% once one-year rates go above 4.75%, which is entirely possible.

The broker was quite clear that you should never buy an adjustable-rate mortgage with an intial rate any longer than the amount of time you intend to own your home. If you're going to sell within five years, then get the 5/1 ARM: it's cheap money. But if you're intending to stay in your house and pay off the mortgage over time, then don't even think about it: you'll be killed once that adjustable rate of 225bp over Libor kicks in.

Libor doesn't go out beyond one year, but Treasury rates do; the one-year Treasury trades today at 3.49%. So let's look at spreads over Treasuries as opposed to spreads over Libor: that way we can compare all the options on a like-for-like basis. And let's assume that 225bp over one-year Libor is the same as 323bp over Treasuries.

Product	Initial spread	Spread after reset
1/1 ARM	226bp	323bp
5/1 ARM	216.5bp	323bp
7/1 ARM	210bp	323bp
30yr fixed	158.5bp	N/A

This is just incredibly counterintuitive to me: the spread curve on mortgages seems to be pretty steeply inverted. The more-floating and less-fixed the mortgage, the higher the spread is – even before you take into account the seemingly-penal interest rate once the initial period is over. Are these numbers remotely similar to the ones that Alan Greenspan looked at when he famously said that adjustable-rate mortgages made more sense than fixed-rate mortgages? How can it make sense for adjustable-rate mortgages to reset to 323bp over Treasuries, while a 30-year fixed-rate mortgage for the same borrower is quoted at 158.5bp over Treasuries?

Indeed, looking at this table, even the initial rates offered on the ARMs look pretty underwhelming: they're "teaser rates" only in comparison to the really high rate charged after they reset. If anybody can provide an explanation of what's going on here, I would be extremely grateful, because I can't make heads nor tails of it. Why is it that a borrower pays more when the borrower is taking interest-rate risk than when the lender takes interest-rate risk?

Update: The fixed interest rate on a 15-year fixed rate mortgage is 5.875% – the same as the teaser rate on the 5/1 ARM.

Comments

[User 123594:]

it's not a buyer's market for borrowers right now. and generally speaking, leaving aside arguments re: volumes and fees which aren't relevant right now anyway, lenders want the teaser rate to be as close to the conv/fixed rate as possible. they have that option now, and market conditions are forcing their hands anyway.

financing costs are up where it is available, yields demanded by the 2ndary market are up (including GSEs), and ARMs are essentially a dirty word for the investing side of things. look at the YOY drops in private label mbs issuance for October - that pretty much says it all - 90% for subprime, 70% alt-a, 55% overall.

2007 Nov 18 01:11 PM

[gordon:]

look at the chart here for the dismal lending future. Also, if foreigners demand higher 10-yr rates, even with real estate deflation, long rates will rise, a dbl-barreled recession. California is already in one.
www.signonsandiego.com...

2007 Nov 19 01:13 PM

[Malkiel:]

It looks like investors have priced in their newly-identified risk of buying commoditized mortgage products from lenders whose vetting process is suspect, which is a reasonable thing for them to do. The borrower is paying more because the creator of the mortgage is facing greater risks attempting to sell those products into the market (such as being stuck with holding the loan, tying up their own funds).

2007 Nov 19 05:29 PM

[AnotherGuy:]

It seems the lending markets have a reduced appetite ARMs right now, with rates (@3-5 yr) within 1/4% or so of a 30 yr fixed rather than a more "normal" 0.75-0.875% below the fixed. The more "normal" pricing would reflect the lower cost of funding the ARM off the nearer term yield curve (for say 3-5 year ARMS) versus the longer term cost of the 10 year Treasury (off of which 30 yr mortgages are priced).

Nomenclature point of order: Properly speaking, those rates you quote aren't "teaser" rates; teasers are significantly lower rates for a short period of time (like a 1% or 2% rate for the first 3 or 6 months) which then reset to the fully indexed rate, or to an initial fixed rate for some additional interim period. The rates you list are plain old "hybrid ARM" or "initial fixed rate converting to floating" ARM rates.

Your friend should stay away LIBOR based ARMs. Most of the time, the typical pricing (LIBOR+225bp vs. 1-yr Treas+275) is very close to the same rate when you hit the adjustment. But about once or twice a decade when credit markets blow up (like recently) the spread of LIBOR-to-Treasury widens substantially, and if your adjustment happens to occur during such a period you can pay quite a bit more than if you were tied to the Treasury. (Also, now that I look at it I think that is what is wrong with your table, which I didn't quite understand; after a return to "normalcy" at some point that 323 will pull in closer to 275).

2007 Nov 19 06:56 PM