Friday, August 23, 2013

30 year fixed rate mortgages are awesome

My mom’s former house in Copenhagen, on which she had a 30 year fixed rate mortgage.  Look up ISIN DK0009274144 on your local terminal to look up her bond and loan terms. (Update: this particular bond is from when she refinanced into a 20 year mortgage. I will try to find the original bond.)

Matt Yglesias writes about the 30-year fixed rate mortgage in the United States and asks whether it would survive the end of government subsidies through Fannie Mae and Freddie Mac:
One is that if you look at other developed countries you generally don’t see the American-style 30-year fixed rate mortgage. Instead you either have short-term loans that need to be periodically refinanced at new interest rates, or else something more like an Adjustable Rate Mortgage.
From these two stylized facts you can draw two conclusions. One is that absent a specific government subsidy for the fixed rate mortgage it would vanish and we would become more like Canada. The other would be to say that the jumbo loan market shows that there’s enough path dependency in the marketplace than the fixed rate mortgage would persist, just at slightly higher cost.
As it happens, I come from a country where 30-year fixed rate mortgages have a long history, so I will now abuse this platform to explain how they work at excruciating length. (The United States and Denmark also share the dubious distinction of being the only two countries with a debt ceiling.) Until recently most residential mortgages were 30-year fixed, and they are still very popular. In Denmark, the mortgages are issued through specialized mortgage banks (which are often part of larger financial conglomerates) that do not have any explicit government backing. There may be an implicit guarantee because some of these institutions are very large, but that guarantee probably isn’t stronger or less implicit than the one that any other systemically important financial institution gets in a small country.

The industry likes to say that there has been a default on a Danish mortgage bond. A careful study the history of the market, which dates back to 1797 and started because of the Copenhagen Fire of 1795, would probably add a few asterisks to that claim. Nevertheless, Danish mortgage bonds are generally considered very safe, and they can be used as collateral for monetary operations at Danmarks Nationalbank with very small haircuts.

Why are 30 year fixed rate mortgages so awesome for borrowers?
  1. The interest rate is fixed. This is an obvious benefit, but that doesn’t mean we shouldn’t take it seriously. Most people can pay off their mortgage in 30 years, and having a fixed rate means they know exactly how much they will pay every quarter (for Danish mortgages) for the term of the mortgage. A homeowner with a fixed rate mortgage could still lose his home if he loses his job, for example, and can’t make the payments, but he is much less likely to lose it solely because of movements in interest rates.
  2. Residential mortgages have a prepayment option. This is useful if the homeowner needs to sell the house and cancel the mortgage. It’s also useful if interest rates fall, so bond prices rise. By refinancing, the homeowner can reduce the mortgage payment, or keep the same payment and reduce the term, or benefit in some other way.
  3. Danish homeowners also have a delivery option. Suppose interest rates go up after the mortgage was originated, so bond prices fall, perhaps significantly below par. Because mortgages are tied directly to the mortgage bonds that finance them (more on this below), homeowners can buy up bonds at market prices, deliver them to the mortgage bank, and have the loan canceled. She would probably need a new mortgage to fund that bond purchase, but will end up owing less on the mortgage.
What are the main disadvantages? The prepayment and delivery options aren’t free, and their price will be built into the yield and ultimately the mortgage payments. I don’t really consider an adjustable interest rate a big advantage per se because the typical homeowner probably shouldn’t try to make guesses about future interest rates. One big advantage of adjustable rate mortgages is that they amortize much faster, which can reduce the homeowner’s risk because he will normally owe less than the fixed rate borrower at any given point in time.

How does the Danish mortgage market pull this off without explicit government backing? I don’t really know. I can list the key features of the Danish mortgage market, but I can’t promise that 30 year fixed rate mortgages will be viable and common in any country that introduces them. The industry is actually in an effort to export the Danish mortgage model to Mexico, and we will see how that goes. In the end I would probably agree with Matt that there is a huge amount of path dependence in the consumer financial products that will exist in a country.

Some of the key features are:
  1. Mortgage loans are tied directly to bonds that are traded on the Copenhagen Stock Exchange—this is known as the “balance principle”. When a mortgage is originated, the mortgage bank will tap issue new bonds to fund it. Instead of having a different interest rate on mortgages that are originated on different days, the bonds generally have integer coupon rates such as 3% or 4%, and are then issued at a discount to par. Danish home buyers are actually expected to understand how all this works and the news media often reports on changes in prices of bonds with particular coupons, for example when a lower coupon bond starts trading below par and is therefore available for new issue. This structure standardizes the way mortgages are traded and eliminates any funding risk for the mortgage bank, because the loan is essentially securitized at the moment it is originated.
  2. The balance principle means that every borrower in a bond will have the same interest rate. Lenders can manage risk by reducing loan-to-value or refusing to lend in the first place. It also means that mortgage rates are more transparent because they are determined directly in a very liquid market, not just made up by the lender.
  3. The balance principle also means that borrowers can’t freely prepay at any time. It generally has to happen in connection with one of the quarterly payment dates.
  4. Mortgage bonds are claims on the issuing mortgage bank, not a separate trust set up for the purpose as is common in the US. Issuers are fully liable for the bonds, everything is on the balance sheet, and they have to satisfy Basel II capital requirements.
  5. For primary residences, loans can be issued up to 80% loan-to-value, and usually 60% for second homes. Borrowers who need additional financing will usually get a loan directly from a commercial bank that will keep the loan on its books.
  6. Mortgage banks make a profit by charging a spread over the coupon paid to bond investors, as well as through various fees. For example, many homeowners want to lock in their rate when they sign the contract for a home, so they will enter a forward contract for the bond they will use, and the mortgage bank charges a fee for this.
Why would investors buy these bonds, with all the weird features? Well, why not? Bond investing is never easy and is often outsourced to bond fund managers. There is very little credit risk. Prepayment risk is hard to handle, but the investor is compensated for it. The delivery option removes some of the upside for investors, but presumably they are compensated for that too.
The Danish mortgage market is so obsessed with determining mortgage rates directly in the bond market that when residential adjustable rate mortgages were introduced, they were also based on market rates. A mortgage whose rate adjusts once a year, known as “F1”, is financed with 1-year bonds. Every December, the mortgage bank holds an auction to sell new 1-year bonds. Unlike the auction rate securities infamous from the US, there is no backstop or maximum interest rate, and it was never entirely clear to me what would happen if one of these auctions fail. Fortunately, they have never failed, even during the recent financial crisis. There is speculation that the central bank, Danmarks Nationalbank, would have intervened. (For commercial mortgages, it’s more common to have loans tied to a reference rate such as Libor or Cita.)

36 comments:

  1. One thing my Danish colleagues complain about a lot (and Guan can correct me): there's some kind of mandatory down payment, like 20-25% -- and many Danes actually BORROW the money to make this down payment in order to take out a mortgage. This might be an old thing, I don't know, and I'm not sure how it ties into 30-years in particular, more a commentary about other regulations.

    ReplyDelete
    Replies
    1. Woops -- i just saw you covered that in the 80% loan-to-value parameter. The point about the "borrowing the down payment" phenomena still stands though.

      Delete
    2. Yes, it’s a major difference in the underwriting process between the US and Danish mortgage markets that US lenders look a lot more carefully at where you got the down payment from.

      You could see this as a feature rather than a bug: use the loan-to-value limits, which could be lower than 80% if the loan is particularly risky for whatever reason, to make sure all the mortgages that go into mortgage bonds are reasonably safe, so that system works well, then use an alternative funding mechanism for the rest.

      Traditionally mortgage underwriting in Denmark has focused more on the value of the house than the borrower’s ability to pay. That has changed in recent decades, and today a mortgage bank would scrutinize the financing of the down payment and the borrower’s household budget much more carefully.

      Delete
  2. When does a 30-year bond stop being called a 30-year bond? Might be a really elementary question but I'm curious, the issuance on that bond is 2005 right? And ends 2028 -- why is it still considered a 30-year?

    I guess the same question goes for Treasury markets. When does a 10-year move from 10-year classification to 7-year? Or does it ever?

    ReplyDelete
    Replies
    1. Almost immediately. In fact, when you look up US Treasury bond yields, you’ll often see the term “Constant Maturity Treasury”, which is a time series of the yield a hypothetical portfolio that is always invested in bonds with the same maturity.

      In the case of Danish mortgage bonds, even “30 year” bond is a lie. Mortgage bonds are typically open for new tap issue for 3 years. However, every borrower gets a 30-year mortgage; the last borrower to use that bond does not get a 27-year mortgage. This means that the bond actually exists for 33 years, and investors get a weighted average of the cash flows paid by borrowers.

      Delete
    2. Indulge me then, if you'd be so kind, in your mother's mortgage bond example:

      Real Kredit issued it in 2005, with a 2028 maturity date. So it had a 23-year term, but was still considered a 30-year mortgage? This is outside the 3-year window you referenced.

      Also, you referred to this as her "former" house, but the bond is still active/trading. Is this because she has turned it into a rental property or is she for some reason paying the mortgage still even though she sold the house?

      P.S. Funny I never had that term-structure question answered when I've taken debt/money market classes. Thanks for clearing it up.

      Delete
    3. She sold the house. This was the mortgage she had.

      And now that I think about it, she did refinance the mortgage once, probably into a 20-year mortgage, so I was wrong. I will try to find her original mortgage, which should still be extant.

      Delete
    4. Just double checked and the bond is dead too, not active. So that makes sense now too. Cheers.

      Delete
    5. It’s actually still being traded. My mom is out of the bond, but many of the other borrowers who used that bond still have their mortgages. I can see that it has 2.3 billion kroner outstanding.

      Delete
    6. woops, i had a space after the number. derp.

      mathematically how does that differ from a MBS though? it's pooling individual mortgages like your mother's with varying cash flows into one bond of averaged cash flow.

      Delete
  3. Hi Guan,

    As I see it the main thing that enables interest rates to be fixed for 30 years in both the US and Denmark is the financing model for mortgages, not government guarantees. In both cases, mortgages are financed with long-term bond issuance, either by the mortgage lenders themselves to fund lending directly (Denmark), or via sometimes complex off-balance sheet structures to remove loans from lenders' balance sheets (US). In both cases, the interest rate risk on fixed-rate mortgage lending is transferred to the capital markets where it is dissipated - in the US's case, by being exported around the world. This prevents lenders being bankrupted by interest rate rises.

    In the UK where I live, lenders generally keep mortgages on their own balance sheets and fund them with shorter-term debt, typically retail call and notice deposits. This leaves them exposed to interest rate risk and makes long-term fixed-rate lending extremely dangerous. The UK's experiment with US-style securitization to fund mortgage lending did not end well; the two banks that used it in a big way (HBOS and Northern Rock) both failed in the financial crisis and were nationalised, so it is not surprising that the general public are now very wary of mortgage-backed securities issued by banks. Though institutional investors might be more forgiving, especially if the Danish rather than US model were used.

    Alternatively, it should in theory be possible for lenders to issue long-term fixed-rate mortgages and dissipate the risk with interest rate swaps (though British public aren't any happier with derivatives than they are with mortgage-backed securities, and they don't understand hedging). But hedging a 30-year fixed rate mortgage portfolio would not be cheap, and the cost would inevitably be passed on to borrowers in the form of higher interest rates or additional fees. I'd guess that this is why lenders haven't offered long-term fixed-rate mortgages here: they would be expensive relative to standard variable-rate mortgages. The UK public is used to variable interest rates here and pretty price sensitive. Would they be prepared to pay the higher cost of a long-term fixed-rate mortgage? I don't know.

    ReplyDelete
    Replies
    1. Tellingly, Danske Bank Northern Ireland, which is owned by the largest bank in Denmark, does not offer 30-year fixed rate mortgages.

      Delete
  4. But I do think that fixed rate mortgages have contributed to the ZILB.

    1. They encourage higher prices and leverage (the greater risk implied by variable rate mortgages will act to discover borrowing to the maximum)
    2. They reduce the leverage of monetary policy - with variable rate mortgages, the central bank has a lever on disposable income. So the effects of monetary policy are more direct.

    ReplyDelete
    Replies
    1. Oops
      discover - should be discourage

      Delete
    2. Reason,

      I'm afraid this doesn't square with the evidence.

      The UK has extremely high house prices relative to average incomes, and prior to the financial crisis people were borrowing up to 6x income and up to 125% of the value of the house. Yet most UK mortages are variable rate, and if they do have a fixed portion it is fixed for less than 5 years.

      There is no evidence that monetary policy in the UK has been any more effective than in the US. Possibly the reverse: Adam Posen suggests that because far more corporations borrow in the capital markets in the US, whereas in the UK they are more dependent on banks, US monetary policy has actually been more effective than UK.

      Delete
    3. Anonymous10:16 AM

      In the UK, you have to look at London separate from other areas. London is extremely expensive, comparable to San Francisco or New York. And just like these areas the cost is because demand is always greater than supply. Outside of the London area, UK house prices are comparable to what you see in the US.

      Delete
  5. Zathras6:03 AM

    So how do the rates in Denmark compare to those in the US? If it's a bit higher, that could measure the effect of the government's input into the industry.

    ReplyDelete
  6. Anonymous9:03 AM

    you should look up Nykredit before you say that Danish mortgage market is not quasi state backed

    ReplyDelete
  7. Wonderful information about mortgage. I love your thoughts and I really appreciate that you share this kind of information. I am still studying more about mortgage.

    ReplyDelete
  8. Would borrowers be able to get financing from a system outside these bonds? What would prevent a borrower from doing that? Why couldn't Washington Mutual fund a 110% LTV, 2/28 Option-ARM to a Danish borrower?

    ReplyDelete
    Replies
    1. Why couldn't Washington Mutual fund a 110% LTV, 2/28 Option-ARM to a Danish borrower?

      Danes are too smart. :-)

      Delete
    2. In general you can. I think the law mainly restricts what you can do with that ARM. You definitely can’t securitize it in the Danish market, but maybe WaMu could fund it in the US.

      Delete
    3. I'm just wondering why, during the bubble, a US bank didn't throw some garbage products there that might initially give 1-2% lower rate and take market share. Why do Danes choose the mainstream system over a non-traditional avenue? Are there legal restrictions?

      Delete
  9. The delivery option would be a very marginal effect - if I buy up bonds and deliver them to get a cancellation of my mortgage that has no adverse effect on the continuing bond holders except that the most credit worthy borrowers might all do deliveryt reducing the credit quality of the continuing pool.

    I think one of the things that make the Danish market work is the high population density resulting in a presumably very liquid market for housing.

    ReplyDelete
    Replies
    1. Phil Koop6:49 PM

      Ah, thanks. I was wondering how the delivery "option" could have economic value. That's a good point about adverse selection, as it seems quite likely that anyone who can afford to deliver must have good credit. It does seem very marginal, though.

      Delete
  10. It should be noted that the proposals for what will replace Fannie and Freddie if any of the bipartisan bills going forward that would eliminate them are passes are very poorly developed. There is a serious chance that indeed 30 year fixed rate mortgages could disappear in the US in the near future, despite President Obama's claim that he wants to preserve them. As it is right now, it looks like there is more interest in rushing to kill Fannie and Freddie than there is in setting up any viable replacement for them.

    ReplyDelete
    Replies
    1. The real-estate and construction lobby are strong, especially in many house districts. I think they will keep the GSEs alive. The home-mortgage interest deduction costs us $70B a year in lost tax revenue, while the GSE bailout may end up 'making' money. I would not bet on the GSEs changing anytime soon, but I'm a terrible gambler.

      Delete
  11. It would be interesting to see a chart comparing Danish and American rates for:
    1) new thirty year mortgages;
    2) bond yields on new mortgage backed securities based on those thirty year mortgages.

    ReplyDelete
  12. Anonymous3:24 PM

    "...There is a serious chance that indeed 30 year fixed rate mortgages could disappear in the US in the near future..."

    --- there is an old saying (my translation): "the blessed seat (spot) is never empty"

    Construction industry, timber industry and many others would go belly up if that happens. I would short Lowes, Home Depot, KB Homes, Lennar, DR Horton and many others if it came to it.

    ReplyDelete
  13. Usually the counterparties for 30 year bonds are insurers and pension plans who have long term liabilities to offset.

    While shorter terms or adjustables are possible, they bring with them their own problems. Increases in rates can make them more difficult to afford, at least for a while, and declining values can make refinancing impossible unless lenders are required to do so.

    ReplyDelete
  14. I teach one of the largest personal finance courses in the country at the University of Arizona and am President and founder of National Personal Finance Education, one the larger licensed providers of a personal finance course required for people in bankruptcy.

    I've thought about this a lot.

    My opinion on this is that usually 30 years is too long, at least for most college graduates with a decent income. And if it's a Wal-Mart worker couple, they should be very careful about buying any home they can just barely afford with a 30 year mortgage. The focus should really be about trying to upgrade their skills and education, and especially the education of their children, so as not to be a Wal-Mart couple.

    Usually the 15 year fixed is best (and even then it's often best to make additional payments to end it even sooner). The interest rate is lower; the payments aren't that much higher, and it so valuable with the country so financially unsecure for families to get to no mortgage as quickly as possible.

    In addition, as usual, positional/context/prestige externalities are profound and huge. When a family thinks 15 year mortgage, they are likely to buy a smaller home, not get so much wood, granite, and stainless steel, buy less large and expensive vehicles, etc. But because they get used to that level -- and don't start getting used to a higher level of position, prestige, etc., the decreased utility is not that much, and the increased security of having no mortgage right when the kids are setting off for college can be huge. And in any case, the buying a smaller home, cars, etc., can mean that total payments aren't even higher.

    Interestingly, given the ginormous importance of positional externalities, if the government nudged forward 15 year mortgages, you might see people commonly owning their own homes twice as quickly with little loss of utility, as you wouldn't lose position, prestige, or context of quality if everyone else was also doing 15 year mortgages and so they also had commensurately less to buy a home with.

    ReplyDelete
    Replies
    1. Anonymous6:13 PM

      "I teach one of the largest personal finance courses in the country at the University of Arizona and am President and founder of National Personal Finance Education, one the larger licensed providers of a personal finance course required for people in bankruptcy."

      The one-eyed leading the blind. We are truly doomed!

      Delete
    2. Of course, we'd also have to control home equity predators from undoing this.

      Delete
    3. You are right that most personal finance education is very bad. The people often have little or no education in finance and economics, and there is such a focus on the trivial or relatively low importance. Elizabith Warren wrote very strongly about this, and I agree. She said it was a motivation for her to write her personal finance book, All Your Worth, which I think gave birth to good modern personal finance. I've used it cover to cover in my courses since it first came out in 2005.

      Delete
    4. I have thought about this a little not a lot. But a few things come quickly to mind. The first is that context matters *a lot*. Your example and your NPFE students show people who have one of, and probably more than one of: low income, low financial stability, bad financial stability habits. I hope, for their sakes, they live in low-cost locations.

      In a high-cost location, SF for example, it would be difficult to buy a home on a 15yr mortgage. Not buying would mean expensive renting (what I did), and a longer term loss of wealth. I suspect that a different solution is needed for that context.

      I have a pair of friends in my current high-cost location (Singapore, *very* high cost). Their context is different; when they bought their flat they both had: low income, high job stability, good habits, good likelihood of increasing income. For them a maxed out long term fixed mortgage makes sense - especially if prepaid when the increased income arrives. They are able to maximize equity, but their job security means that long term liabilities are ok. Just a thought about context.


      Delete
  15. Hansen Mortgages Svc is a mortgage lending institution that prides itself in providing high quality financial services to the public.

    ReplyDelete