The real problem was refinancing

Maybe everyone knew this already, but I didn’t:

Three trends in the U.S. housing market combined to dramatically magnify the losses of homeowners between 2006 and 2008 and to increase the systemic risk in the financial system. Individually, these trends – rising home prices, falling mortgage rates, and more efficient refinancing – were neutral or positive for the economy. But together, they lured masses of homeowners to refinance their homes and extract equity at the same time (“cash-out” refinancing), increasing the risk in the financial system, according to Amir Khandani, Andrew Lo, and Robert Merton. Like a ratchet tool that could only adjust in one direction as home prices were rising, the system was unforgiving when prices fell. In Systemic Risk and the Refinancing Ratchet Effect (NBER Working Paper No. 15362), these researchers estimate that this refinancing ratchet effect could have generated potential losses of $1.5 trillion for mortgage lenders from June 2006 to December 2008 – more than five times the potential losses had homeowners avoided all those cash-out refinancing deals.

.   .   .

Using a model of the mortgage market, this study finds that had there been no cash-out refinancing, the total value of mortgages outstanding by December 2008 would have reached $4,105 billion on real estate worth $10,154 billion for an aggregate loan-to-value ratio of about 40 percent. With cash-out refinancing, loans ballooned to $12,018 billion on property worth $16,570 for a loan-to-value ratio of 72 percent.

.   .   .

This ratchet effect can create a dangerous feedback loop of higher-than-normal foreclosures, forced sales, and, ultimately, a market crash. With home values falling from the peak of the market in June 2006, the study’s simulation suggests that some 18 percent of homes were in negative-equity territory by December 2008. Without cash-out refinancing, that figure would have been only 3 percent.

Does this matter?  I am not so foolish as to try to make moral judgments about millions of people I have never met, so you won’t catch me saying; “Aha, they weren’t taking out these mortgages to put a roof over their children’s heads, but rather to get cash to buy a big screen TV and a boat.”  For all I know the cash-outs were to meet unforeseen medical expenses.

But here is how it might matter.  Earlier I mentioned that in a world where first best policies are politically impossible  (i.e. eliminating FDIC, TBTF, Fannie and Freddie, tax deductibility of mortgage interest) then maybe we should consider second best policies such as a requirement that people put at least 20% down on mortgages.  At the same time I recognized that even that sort of regulation would be impossible to get through Congress.  It is true that countries such as Canada and Denmark make it tough to get sub-prime mortgages, but their political systems are probably much less corrupt than ours.  Perhaps if we formed 50 separate countries, as I advocated earlier, we might get sensible reforms in states (like Minnesota?) with cultures similar to Canada and Denmark.

But this NBER study suggests there might be a slightly more politically palatable alternative, which could be almost as effective.  What about simply requiring at least 20% down on all refinancing, but continue to allow sub-prime loans on home purchases?

I suppose people would try to evade these regulations by moving next door.  Of course that sort of subterfuge is very costly, a factor that cuts both ways when considering the pros and cons of this sort of regulation.  Or how about exempting only first time home buyers from the 20% down.  Would that be more politically feasible?  Would it be too costly to enforce?

My first choice is still for second best policies that require everyone put at least 20% down on all mortgage loans, as this sort of regulation would also address another government failure, the enormous disincentives to save built into our fiscal policies.

Tags: ,


26 Responses to “The real problem was refinancing”

  1. Gravatar of marcus nunes marcus nunes
    14. December 2009 at 12:43

    The “at least 20% downpayment” is a no brainer. But go tell that to all the government people that think this would be discriminatory. Political Correctness requires that results cannot be disparate! (and low income households would be penalized under this rule).

  2. Gravatar of StatsGuy StatsGuy
    14. December 2009 at 12:46

    “What about simply requiring at least 20% down on all refinancing, but continue to allow sub-prime loans on home purchases?”

    It’s pro-cyclical. One of the chief problems right now is that people who bought with 20% down can’t refinance… Even if refinancing to a lower/fixed rate makes the difference between staying/losing their homes. Sure, they made a lousy gamble – but there are a lot of people in that situation. It’s also somewhat self-fulfilling.

    Consider: If the banks managed to refinance a lot of these people, it could stabilize prices and prevent foreclosures, which is virtuous in the sense that foreclosures force down sale prices which – yep – prevents sales to escape foreclosure. (BTW, one of the problems with refinancings is that owners of the loan can’t directly renegotiate, even if it would be in their interest, because loans were securitized and are being managed by management agencies… which have no interest in refinancing, let alone bearing the cost of this without remuneration.)

    Conceivably, you could prevent/restrict “cash-out” refinancings in which people take out more in loans than the value of the house, but then you are making moral judgements on how people use that money.

    One possibly stabilizing rule would be to limit cash-out refinancings by preventing cash-extraction on refinance loans of more than 80% (or even 90%) of the value of the 15-year trend-stabilized value of house. Relaxation of the rule could occur for hardship circumstances.

  3. Gravatar of Philo Philo
    14. December 2009 at 12:55

    I suspect the income-tax deductability of mortgage interest payments stimulated a lot more refinancing than would otherwise have occurred.

  4. Gravatar of dWj dWj
    14. December 2009 at 12:58

    I think the issue raised by Statsguy also creates more political trouble; if people are refinancing because interest rates drop (or, as was common with subprime loans, taking ARMs with the expectation that they will refinance when rates go up), it’s exactly these poor, defenseless illiterates who wanted nothing more than the right to own a house without actually having the money to buy one who are going to find themselves stuck with these unfavorable terms.

  5. Gravatar of Scott Sumner Scott Sumner
    14. December 2009 at 13:08

    Marcus, I agree. Here’s how I’d put it:

    The PC people say “looks what happens when you have no regulation, banks make lots of risky subprime loans. It shows we need regulations.” Then if someone proposed doing something about subprime loans, they’d say “We can’t do that, it would make it harder for people who can’t afford houses to actually buy houses.”

    Statsguy, You said;

    “It’s pro-cyclical. One of the chief problems right now is that people who bought with 20% down can’t refinance… Even if refinancing to a lower/fixed rate makes the difference between staying/losing their homes. Sure, they made a lousy gamble – but there are a lot of people in that situation.”

    But this makes no sense. Or should I say it would make sense if I was proposing that we reduce the minimum from 30% to 20%. But I am proposing increasing it from 0% to 20%. I agree that even with 20% down some would have trouble refinancing, but surely less than with 0% down.

    In addition, this reduces risk to the banking system, which is supposed the big public policy problem that we face.

    You said;

    “Conceivably, you could prevent/restrict “cash-out” refinancings in which people take out more in loans than the value of the house, but then you are making moral judgements on how people use that money.”

    I am not making any moral judgments, it is others who make those judgments. I don’t care if people borrow 100% and blow it all in Vegas. It is others who care. I am simply proposing public policies that might have a prayer of getting through our corrupt Congress, and yet also might partially offset the moral hazard problem in our banking system. I make no distinctions at all in regards to how people use their money, I am merely suggesting policies aimed at reducing systemic risk that others might consider politically feasible. I political feasibility is driven by (possibly flawed) ideas about morality, that’s not something I can control.

    Regarding your last suggestion, I don’t see the value of the 15 year average price, isn’t the current price always more relevant? In Japan the 15 year lagging average is far above the current market value of a house. In 12 years that may be true here.

    Again, I was talking about refinancing, not buying homes for the first time, so I just don’t see the downside in requiring 20% down. It would not stop a poor person from buying their first house with only 3% down, and then gradually building up equity that could be used to buy a bigger house later.

    After what we have just gone through do we really want to people who are refinancing for the purpose of extracting cash to be able to borrow 100% of the value of their house? Especially with what is essentially taxpayer money?

  6. Gravatar of Scott Sumner Scott Sumner
    14. December 2009 at 13:13

    Philo, I agree.

    dWj, I don’t like making generalizations about people who default. There are undoubtedly a wide variaty of situations. But I do agree that the last thing we want to do right now is make it easier for people to borrow close to 100% of the value of a house.

    I don’t blame people for taking out loans that looked like good deals at the time. If a bank offered me a deal that was very bad for them, but seemed good for me (at least the odds favored me) I’d jump at it. I’d figure it’s up to the bank to look out for itself.

  7. Gravatar of Joe Joe
    14. December 2009 at 13:21

    But why stop a bank from making a 0% down loan, other than the fact that we as a society rescue them en mass when they blow up? Why not let the banks make these loans if they think they are good investments. And why be mad at people when the value of the option, most zero down loans are options if you look at payoffs, makes it better for them to put the house back to the bank? Sounds like the kind of financial transactions that happen every day.

  8. Gravatar of mlb mlb
    14. December 2009 at 13:21

    How about we let banks fail. Then they will think twice before making such loans.

  9. Gravatar of David Pearson David Pearson
    14. December 2009 at 13:31

    I’m not sure an effective 20%-down rule would require legislation. The simple solution would be for all GSE-guaranteed mortgages to be written thus: “the mortgagee agrees to immediately re-pay the mortgage in full if the property is further encumbered by additional mortgages”. This is a matter of Fannie and Freddie loan documentation and underwriting standards. Both of these are under the control of a government-controlled management and board. Fannie, Freddie, and Ginnie/FHA today guarantee over 90% of all mortgages. In other words, the taxpayer, effectively, guarantees 80% of all mortgages. This is not a “free market” vs. “regulation” issue. It is simply a matter of protecting the taxpayer’s interest.

  10. Gravatar of David Pearson David Pearson
    14. December 2009 at 13:32

    Sorry — it should have read “the taxpayer guarantees 90% of all mortgages” (not 80%).

  11. Gravatar of StatsGuy StatsGuy
    14. December 2009 at 13:59

    “I am not making any moral judgments”

    I know, I meant the generic “you”. You were quite clear about this in your original post. Sorry.

    Regarding the pro-cyclical issue, consider this:

    A bunch of people buy houses with 20% down. If real interest rates fall (why they fall is another question) and house prices rise 10%, then these people can all refinance happily and take cash out of their newly-discovered equity. Yay.

    If real interest rates rise (even as nominal rates fall) and prices fall (let’s say there’s an exogenous effect that increases demand for liquidity), and house prices drop 10%, then no one can refinance (which would reduce their mortgage payments).

    And this does not even take into account option ARMs, which have interest rate cliffs; we can criticize homeowners for taking out short/variable rate mortgages, but this was _exactly_ the trend in commercial paper and government debt (a move toward shorter maturities with higher rate risk but lower average rates).

    Now if the price drop is 30% or more, this is compounded by a foreclosure crisis, and we have that right now.

    What you _really_ want is to prevent people from _ever_ taking out more than 20% equity on their homes (or buying homes with less than 20% equity), AND since house prices may be “frothy” in “local markets” (a la Mr. Greenspan), you also want to make sure that the equity calculations are not hokey (e.g. use historically stabilized price values).

    But, to avoid pro-cyclical dynamics, we should still allow non-20% equity refinancings to lower rates (which makes the mortgage more affordable) so long as equity is not taken out of the home.

    Is that more clear?

  12. Gravatar of Master of None Master of None
    14. December 2009 at 14:51

    “the system was unforgiving when prices fell”

    or perhaps prices were unforgiving when the system failed?

    We will get a good feel for how much healing there has been in the housing markets next year if long rates rise as much as many economists predict, making current prices less affordable. Of course that could be offset by lower lending standards if the banks are strong-armed into lending out their excess reserves.

  13. Gravatar of Doc Merlin Doc Merlin
    14. December 2009 at 15:25

    Refi is one of the major ways that money actually gets out into the system when the cost of borrowing drops. Requiring an extra down payment on refi makes inflationary monetary policy due to interest rate reductions very slow.

    If you believe in the effectiveness of monetary policy in general and wish the fed to actually be able to be effective with it, you must make the barriers to refi, very small.

    And most people I know who refinanced did so to save money. Interest rates were absurdly low (My student loans for example, were consolidated at 2.4% interest) and people took advantage of that to save money and lock in low rates because they knew rates weren’t going to stay this low. That was even the sales pitch with the refi company, “rates are going to raise next month so get in while you can.” This meant that everyone who had good credit then was set. People with bad credit who were in ARMs however got shafted when the rates rose.

  14. Gravatar of Doc Merlin Doc Merlin
    14. December 2009 at 15:33

    A second note:

    When buying a house, no one looks at the sticker price, what they actually look at is the monthly payment and duration of the loan. This leads me to say this:
    I think you are placing too much importance on refi for the bubble.
    Here is just looking at the prime rate, I’m using the prime rate because I can’t find the actual 30-year mortgage rate, and they usually track each other:

    May 2000: 9.5%
    May 2005: 6%

    This means that if you bought a house for the same nominal price in May 2005 as in May 2000, the monthly payment on the 2005 house would be almost 1/3 as much.
    This alone can account for most of the bubble in the first half of the 00’s. Then as the rates rose again, housing prices fell, and people got into trouble.

  15. Gravatar of Don the libertarian Democrat Don the libertarian Democrat
    14. December 2009 at 15:49

    “Lower interest rates are an incentive for investment. They take place in a context of competing incentives and disincentives. But an incentive is just that. Raising interest rates on the entire economy seems a rather blunt instrument to stop a housing bubble. The bubble part came from housing prices going up in an extraordinary amount as against most other purchases and parts of the economy. Had down payments been 20% or based on a reasonable percentage of income, then this crisis might not have taken place. Why focus on interest rates if there are other possible cures for the disease? For all we know, raising interest rates could have also meant that many potentially valuable businesses might never come into existence.

    Perhaps the Fed should have done things differently, but why not focus on the industry itself where the problem existed?”–blame_the_fed_the_government_in_general_and_the_economists–richard_a_p.php

    That was my view on May 19th. I then added this on May 30th:

    “Let me give you a particular proposal: As housing prices go up relative to other goods, and, hence, become More Expensive, you raise the down-payment requirement. Would that have had any effect? You are indeed reigning in a part of the economy, but it’s a particular part of the economy, as opposed to the whole economy.”–may_16_to_may_26.php

    After that, Richard Thaler made a similar point:

    “While imperfect, financial markets are still the best way to allocate capital. Even so, knowing that prices can be wrong suggests that governments could usefully adopt automatic stabilising activity, such as linking the down-payment for mortgages to a measure of real estate frothiness or ensuring that bank reserve requirements are set dynamically according to market conditions. After all, the market price is not always right.”

    And then on Aug. 18th on Baseline Scenario:

    “For Texas, you also have to consider the relative absence of land use regulations. My own view, also now advocated by Richard Thaler, is to increase the down payment as home prices rise relative to other goods. There should always be a 20% minimum down payment.”

    So, basically, I agree. But my comments also had to do with the proposal for the Fed to Lean Against the Wind, in which case I believe that it will be a Leaner of Last Resort, and only burst the bubble when it’s bigger than our heads. I also said this:

    “If this was me, you misread my point. My argument was that there were other, more concentrated, means of bursting the housing bubble without having to use the Fed. While you were freeing capital from housing,that might well have been better spent, you could well have been destroying other businesses with higher interest rates. If you see that it’s a housing bubble, then you should be able to deal with it as a problem of the housing sector. I’m not even sure why you believe that the Fed has the sagacity to notice the bubble, but can’t bring the problem to the attention of regulators and lawmakers, and suggest that they deal with it as a problem with the housing sector.

    Quite frankly, since the Fed has to use a blunt instrument, I would imagine that it would be a laggard in dealing with the problem, waiting until everybody else has attempted to deal with it, before they apply the breaks to those not involved in the bubble area as well. This reliance on the Fed seems ill-founded.”–may_16_to_may_26.php

    Posner responded to me by saying this:

    “One comment makes the interesting suggestion that raising interest rates in an effort to burst the bubble before it became so large that its collapse did grave harm to the economy would have been a blunt instrument, since higher interest rates reduce productive activity as well as pricking asset-price bubbles, by restricting credit. But higher interest rates, by bursting the bubble early, would have freed up a lot of capital for productive investment–capital that instead went into the purchase of houses at ever higher prices.”

    This was the general response that I got. But then, I read this:

    “The events of the past couple of years have clearly tipped the balance in favour of taking preemptive
    action. And in a second-best world, where monetary policy is the only instrument available to cool a credit/asset-price boom, then that may well make sense. Indeed, an
    inflation-targeting central bank ought to be willing to undershoot its target in the medium term, if it thereby improves its chance of meeting the target further out through the avoidance of a disruptive bust (Bean, 2003). But monetary policy is a blunt weapon for this purpose, and
    raising interest rates enough to cool a credit/asset-price boom that is in full swing is likely to involve substantial collateral damage to real activity.”

    I’m sorry that I’ve gone on so long, but I’m trying to make the case that the housing bubble would have been best addressed by focusing on the housing bubble, and not slowing down the entire economy. Since this blog focuses on Monetary Policy quite a bit, I’m trying to argue against Monetary Policy being the cause and solution of the housing bubble.

  16. Gravatar of JimP JimP
    14. December 2009 at 17:02

    Off topic – but –

    In this post –

    Krugman links to this paper –

    which is entirely clear – monetary policy does work at the zero bound – and it works entirely due to changes in inflation expectations. The effect is very powerful. As powerful (or more) than fiscal policy – if the central bank can credibly commit to it. See pages 23 and following.

  17. Gravatar of marcus nunes marcus nunes
    14. December 2009 at 17:34

    Your theme revolves around: “Raising interest rates on the entire economy seems a rather blunt instrument to stop a housing bubble”.
    No wonder Charles Bean agrees with you. A little over 10 years ago, the BOE raised interest rates with the argument that the housing market in the South East of England was “heating up”. At the time I thought it strange that MP could (or should) target a local market development that would have (and had) negative spillovers over the rest of the country.
    A couple of years later I read the following in an article; “where do you think the French Silicon Valley is located”? Exactly: In the London-Dover corridor. While in France the marginal tax rate was 60%, in England it was “only” 40%. Certainly mobility in Europe is not as easy as in the US, but the EU made it much easier and taxes are a great incentive!
    That would explain an increase in house demand in the SE of England. In the short run supply is inelastic but in time contractors and builders from the other parts of the country would divert resources to the area. Sur enough, prices came down with the increase in supply of houses.

  18. Gravatar of Matthew Yglesias Matthew Yglesias
    14. December 2009 at 18:24

    Part of the idea of the CFPA in the Frank regulatory reform bill is precisely to keep a lid on things like no money down “no doc” loans and such. It’s being fought by financial services industry lobbyists. I see no evidence that people being “PC” is the problem.

  19. Gravatar of Jon Jon
    14. December 2009 at 21:28

    Isn’t the claim actually that the banking system depends on the overall pool of equity. It isn’t clear that if people keep ratching their equity down to 20% that things are still stable. i.e., if there was no refinancing what is the initial equity percentage needed to ensure the system has adequate equity overall to prevent a systemic default?

  20. Gravatar of Greg Ransom Greg Ransom
    14. December 2009 at 23:01

    Scott, I live in Ladera Ranch CA, the #1 foreclosure zip code in Orange County, which is ground zero of the housing boom and bust in percentage and dollars terms.

    With housing prices doubling in 2 or 3 years, almost everyone took moneymout to domlandscaping, buy TVs, put in a pool, etc.

  21. Gravatar of StatsGuy StatsGuy
    15. December 2009 at 05:50


    “With housing prices doubling in 2 or 3 years, almost everyone took moneymout to domlandscaping, buy TVs, put in a pool, etc.”

    Hence the need for a ~15 year moving average (or trend based) price basis for establishing a 20% equity non-cash-out refinancing rule…

    It also has other beneficial effects:

    – Increasing the savings rate
    – Helping anchor inflation expectations

  22. Gravatar of Bill Woolsey Bill Woolsey
    15. December 2009 at 06:12

    With refinancing, 20% down doesn’t make sense. The rule would be, only borrow against 80% of the value of the house.

    And that means that if the house price drops up to 20%, then the loans can be paid.

    The problem is that people lent against the security of overvalued collateral.

    I think the fundamental problem is that people acted on the assumption that past capital gains on housing provided a good reason to expect future capital gains in housing.

    It was bad enough that borrowers made that mistake, but so did lenders. I think the “solution” is for lenders to take losses. It is an entrepreneurial error. Misjudging the value of the collateral.

    That the government promoted (and lenders accepted) the plan of using rapid housing appreciation to get poor people into homes they couldn’t otherwise afford wasn’t the key problem.

  23. Gravatar of Scott Sumner Scott Sumner
    15. December 2009 at 06:19

    Joe, Your question is answered in the very first sentence you wrote. Again, in a free market I would not support this policy, I think it is a second best policy we should consider, to offset the harm done by other bad government policies.

    mlb, That is my preferred policy. If we do that we don’t need any regulation. But we also must end FDIC, otherwise banks will take risks and pass the cost on to taxpayers.

    David, You may be right, but when Bush tried to tighten up the lending standards of the GSEs, people like Barney Frank stopped him. So it may require legislation. I am not certain.

    Statsguy, I think what threw me off is that I interpreted your first comment as meaning more procyclical than current policy. Actually it is less procyclical than 0% down, but still somewhat procyclical. If that is your view than we agree.

    Master of None, Yes, that is true if long rates rise. But I disagree with other economists. The Japanese case makes me doubt long rates will rise very much.

    Doc Merlin, I think you confuse money and credit. Monetary policy can work just fine in an economy with no credit markets at all. So it certainly doesn’t need refinancing.

    Doc Merlin, Your quarrel here isn’t with me, it’s with the NBER study. But generally NBER studies do a pretty sophisticated job of analyzing the data. It doesn’t mean they are right. But I don’t think your example disproves their study. Surely they knew about the fall in rates when they estimated the impact of refis vs new home sales.

    Don, I think we agree. Monetary policy is a blunt instrument that affects other industries as well. better to target regulations to address the specific problem in housing. Is that your conclusion? I am glad to see the other people you mention also advocating 20% down.

    JimP, Thanks, see my new post.

    Marcus, Interesting story. And of course that area is close to the train tunnel to France.

    Matthew, Would you agree that people being PC was a problem when the Bush administration tried to tighten up the GSE standards a few years back, and people like Barney Frank opposed them on grounds that it would reduce lending to low income individuals? I can’t say whether you are right about the current political environment, and I do agree with your implication that corruption is the greater problem. Indeed I mentioned corruption in my post, and only mentioned PCism in a reply when a commenter brought it up. But I also believe the following:

    1. There are still many Congressman on the left would would oppose anything close to a 20% downpayment requirement for essentially PC reasons. (Maybe PC is the wrong term, let’s say for ideological reasons.)

    2. The biggest problem in Congress is corruption, not PCism

    3. I don’t think Frank is corrupt, just misguided.

    4. Frank is not a typical Congressman.

    5. The Obama administration has adopted a new program that actually has the FHA subsidize no income loans with effectively no money down. I did a post on this program a few weeks back. It discussed a WSJ article by Pozen.

    6. I’m no expert here and could be wrong. Indeed I hope I am wrong.

    Jon, 20% down doesn’t solve the problem, but it makes it smaller than 0% down, indeed much smaller if this study is to be believed. Remember that lots of older people have much more than 20% equity in their homes. I am talking about a minimum, not an average. They estimate that refis lowered the equity stake from 60% to 28%. With 20% down, the average equity stake would have been far higher than 28%.

    Greg, Yes, I think that is exactly the problem this study highlights.

  24. Gravatar of Jon Jon
    15. December 2009 at 08:37

    Right Scott… and my question was whether the average bolstered by that minimum was enough. The monotonicity is clear; what’s in question in the magnitudes needed….

  25. Gravatar of Scott Sumner Scott Sumner
    15. December 2009 at 17:54

    Bill, I see your point, but did you notice that their estimates imply that the refis didn’t just hurt by making more housing go underwater, they also claim that all the refinancing was a huge factor in driving the housing bubble. Notice how the estimate of the total value of the housing stock is far lower without all the refinancing. That seems important as well. I think a 20% down requirement (or at least 80% equity, as you say) would have reduced that, even given all the caveats that you point out.

    Jon, Yes, and I can’t answer that. But also look at my response to Bill. I should have discussed in the post their estimates of how much refis helped blow up the housing prices. So it isn’t just ratios at issue, it is the total size of the problem.

  26. Gravatar of Don the libertarian Democrat Don the libertarian Democrat
    15. December 2009 at 18:10


    Yes. We agree.

    Take care, Don

Leave a Reply