A Stupid Suggestion

A guest blogger on Megan McArdle's blog writes:

Here's my first such idea:

Abolish Mortgage-Backed Securities (and Offspring)

CDOs and credit default swaps don't kill financial systems, mortgages kill financial systems. There has been altogether too much opproprium directed at CDOs, credit default swaps and other structuring techniques that spread financial contagion, and not enough directed at the underlying collateral. The record seems to be, however, that Dick Pratt was correct when he called the mortgage "the neutron bomb of financial products."

This makes no sense.  I don't have time for a comprehensive argument, so here are a few bullet points:

  • His argument rests on the fact that mortgages have inherently hard-to-quantify risks.  I don't believe that, given how long the financial system worked just fine writing mortgages, but if this is really the case, shouldn't he be proposing to ban mortgages, not just mortgage-backed securities?
  • Holding the higher-quality tranches of an MBS simply cannot, by any mechanism I can fathom, be more risky than holding a lot of individual mortgages.  In fact, for a given bank, it should spread the risk geographically and to a larger number of mortgages.
  • The first actual problem with MBS's is that the default risks were under-estimated by those packaging the securities.  Basically, the top AAA tranches were too large (or too wide, I think the term is).  This is correctable, and likely already has been corrected (In fact it had more to do with the actions of the government-enforced credit rating oligopoly than with actions of bankers).
  • The second actual problem with MBS's is that the default risks were under-estimated by government regulators world-wide when in Basil II and the equivalent US law changes c. 1991, MBS's were given very preferential capital requirement treatment.  Basically, MBS were treated, for capital requirements, as if they were nearly as risk-free as US treasuries, providing incentives for banks to over-weight in them.
  • The largest problem was the reduction in credit requirements for mortgages.  Increasing LTV from 80% to 97% or 100% or even 100%+ hugely increased the risk of default, and no one really took that into account in MBS packaging or bank capital requirements.  Bank capital requirements for mortgages and MBS were set as if they were European style recourse loans with 80% LTV.  But the same regulations and requirements applied to MBS built on US-style non-recourse loans with 97% LTV, which is crazy.

Here is a better plan:

  1. Narrow the AAA tranches of MBS
  2. Fix bank capital requirements vis a vis mortgages and MBS
  3. Stop encouraging high loan to values on mortgages


  1. TJIC:

    For whatever reason, economics is one of those areas where it's easy to convince yourself that you understand things at least as well as the experts.

    Physics journals are apparently always getting submissions from borderline illiterates disproving relativity, or pi, or somesuch.

    In economics, the bar to publicity and getting link-love from lefties is much lower: just blog it.

    The particular delusion that most bugs me (well, OK, that's not fair - there are dozens that bug me equally)...but ONE delusion that bugs me is the logical fallacy "X was the proximal event to the problem, therefore X is the problem".

    Corporations lay off workers? Easy! Pass a law saying that corporations can't lay off workers.

    CDOs blew up? Easy! Outlaw CDOs!

    The Titantic got holes in the right hand side? Easy! Outlaw hull plates on the right hand side of ships!

    A building burned down? Easy! Outlaw fire!

    I'm not even talking about cost benefit analysis or unintended consequences. I'm talking about people who can't grok that the manner in which a problem exploded into view is not synonymous with the cause of the problem.

  2. Mark2:

    FHA still offers loans to 97% And you can game the system to still get up to 101% (ask for cash back from the seller, get a secondary loan for 2% of the house)

    And in addition to that FHA loans have better interest rate terms than conventional loans. And of course the requirements are incredibly low:

    "Two Years of steady employment, preferably with same employer. Last two years Income should be the same or increasing.Credit report should typically have less than two thirty day lates in last two years with a minimum credit score of 620 or higher or in some cases no credit score at all.Bankruptcy's must be at least two years old, with perfect credit since discharge. Foreclosure's must be at least three years old, with perfect credit since. Your new mortgage payment should be approximately 30% of your gross (before taxes) income."

    Has anything changed since the crash? No.

  3. me:

    Honestly, there is absolutely no problem with mortgages, MBSes of CDOs. There is a problem with risk/reward ratios caused by Too Big To Fail: if you're running a business and are responsible for losses and stand to benefit from the gains, you will control risk in a reasonable manner. And if you don't, well, maybe you get lucky and enjoy a huge payout. Or you get unlucky and you go bankrupt.

    In a system where you can take on insane risks, reap outrageous rewards if you get lucky and be rewarded only handsomely if you fail, the people holding the bag (taxpayers) are inevitable doomed, because even a rational player would always go batshit in that kind of incentive system.

    You can either resolve this structurally (by smashing any institution large enough into bits), but there is a risk that there won't be enough diversification regardless. Or you can address it by adding a cost to the folks responsible (hey, AFAIK, go all out and use our newfound powers to suspend any rights to fight "financial terrorism" by executing/imprisoning/seizing all the assets of the first four layers of management of any firm that is being bailed out).

    You can't fight it by continuing to play spin the wheel with a Billion dollar payout on red and a few million we'll get from your neighbors and give to you in any other case.

  4. Bram:

    Isn't the stock of a traditional bank (that doesn't sell their mortgages)a Mortgage-Backed Security?

  5. Anonzmous:


    I think a bit issue you miss is that the risk assessments of the tranches was modeled assuming that defaults were not correlated. In other words, that Alan's ability to pay was not connected to Bob's and Charlie's abilities to pay. Sure, Alan might get hit by a car and stop making mortgage payments while he's in the hospital, but Bob and Charlies are still making payments.

    Or to look at it another way, if one person with a credit score of 700 will default on average once every 100 years, then a group of 100 mortgages from people with credits scores of 700 see no more than one default every year.

    But what happened was a lot of people suddenly lost the ability to pay at the same time. Alan and Bob got laid off, and Charles was unable to find a buyer or renter for that condo he was flipping.

    So that group of 100 mortgages saw 10 or 20 defaults all in one year. Sure, each borrower on average defaults on average once every 100 years -- the total number of defaults was correct. They just didn't account for the distribution over time.

  6. Brad Warbiany:


    Have you read "The Big Short", by Michael Lewis? If not, I recommend it...

    One of the points that he made that dovetails nicely with the underestimation of risk was the way that these folks would game the ratings agencies. They knew that the ratings agencies were spending WAY too much effort looking at average credit score of the mortgage portfolio, and not enough at the distribution. So they'd package up "barbell" MBS packages with a bunch of extremely low-risk high-score mortgages with a bunch of garbage. The average score was still "low-risk", but the packages were full of a bunch of time-bomb mortgages waiting to go off.

    (I think I remember you actually blogging about this a while back though when you were considering a re-fi -- I could be wrong).

  7. Fred Z:

    A banker, as an individual, a person, who writes a bad mortgage, does not get fired, or sued, or smacked upside the head. When that ends, so will bad mortgages.

    I lend money to high risk people on mortgage security. Over 25 years my default rate has been low and my loss rate is zero %.

    It takes a huge amount of stupidity to screw up the lenders position in a mortgage security. Happily, the government is always there to supply that stupidity.

  8. NL_:

    If the MBS is risky then the CDO is risky.

    They kept repackaging the residuals into new packages, meaning the riskiest part of one product was repackaged with a bunch of other risky parts into another product. The squared product was often then further subdivided into primary and residuals. So some people holding the cubed product residuals were holding the riskiest part of the riskiest part of a bundle of somewhat risky subprime mortgages. Yet they still rated it AAA. The rating agencies failed to police this stuff, the banks failed to investigate their own investments properly, and the originators didn't care about high-risk mortgages because the banks were willing to buy anything.

    And the rating agencies based their predictions on the assumption of persistent growth in real estate. Since so many subprime mortgage holders had little money down, the main way to avoid ballooning interest payments was to refinance. But refinancing only makes sense if you have equity or the house keeps appreciating. If the market stopped appreciating, the ratings guys admitted that there were serious challenges to their model (thereby risking the AAA ratings). If there was a nationwide small downturn in housing prices, it would completely destroy the model that allowed the AAA ratings.

    Everybody uncritically accepted that this stuff was a good investment. If investors refuse to police their own investments, and to probe the assumptions underlying the ratings models, there's nothing the government can do to fix that. Nobody will care more about your money than you do. There was a real bandwagon problem here, with banks not wanting to be left out of the free money. Some notable investors made loads of money by shorting real estate and mortgages. But the majority of the market couldn't get enough MBS. The government isn't a good way to fix this sort of systemic misvaluation. If anything, the government is the best way to get even greater systemic errors, since the main benefit of the government is regulating players into uniformity.

  9. morganovich:

    i think you left out the biggest problem of all:

    many buyers treated stuff they knew was crap as AAA because they believed (correctly as it turns out) that the us government would make them whole on anything that said freddy or fannie on it.

    the GSE's were handed a license to steal with this. they took utter crap, and the crappier the better as the spread is wider and slapped "uncle sam" on it and voila! AAA. absent their participation, not even moody's would have given this stuff AAA with a straight face.

  10. morganovich:

    "If the MBS is risky then the CDO is risky."

    only for the guy who writes it. owning one is a perfectly quantifiable risk. it's just purchase price. where you get killed is writing them and suddenly discovering that the assets you thought were uncorrelated ARE correlated and that you are on the line at 200:1 leverage for a whole whack of debt.

  11. Smock Puppet, 10th Dan Snark Master:

    >>> the default risks were under-estimated by government regulators world-wide when in Basil II

    Wow, that's a long time. Basil II, was he the third or fifth Emperor of Byzantium...?



  12. A Friend:

    Agree with morganovich: the problem was US Gov't guarantees on MBSs. Without that, no mispriced debt available, and without that, no wild speculation in housing. An interesting fact that gets lost in all this: about half of all houses built in 2005/2006 were second houses, i.e. built to "invest" not live in.

  13. Ted Rado:

    The good old days when one needed good credit and a substantial down payment to get a home mortgage are gone. Instead, we have "anyone with a warm body is entitled to a loan of any size with almost nothing down".
    To add to the folly, our financial geniuses dream up all sorts of packaged securities, derivatives, derivatives of derivatives, ad nauseaum. Is it any surprise that we are in the toilet? Where do the bankers, MBA's, and politicians get all this crap? Why not just shoot ourselves?

  14. me:

    Actually, the problem wasn't that the government was guaranteeing mortgages etc - they didn't. The problem was they stepped in after the fact and did because a lot of people had expected them to. Hence, the mess we're in. If you want capitalism and the benefits of free markets, you must allow bad businesses to fail and bad financial decisions to eat the consequence.

  15. Ted Rado:


    There was a lot of arm twisting by the government to loan money to people who weren't good risks. Then, the USG set up agencies to buy shaky mortgages from the bankers. Thus the bankers were pushed two ways to make the loans. No banker, left to his own devices, would loan money to someone who is a poor credit risk or who wants to borrow more than he can afford to repay. Yes, many expected the pols to do this, but it is still a horrible idea. The pols are supposed to lead us down the correct path, not acquiesce to those wanting freebies.

    If you want capitalism, you must leave people alone to make wise decisions. No arm twisting or encouraging bad decisions.

  16. ParatrooperJJ:

    Require the bank that writes the loans to keep at least 25% of the loa in house.

  17. me:


    Not disagreeing with your point (in fact, rather strongly agreeing with the notion that overregulation is one of the largest problems the US faces this day)

    However, markets are wonderful, they don't even need that feature. All it really needs is that decisions be borne by those making them in the end. If they are getting bad incentives, it's still up to them to ignore those. Example: there are all sorts of programs paying me for letting characters live in my properties that I personally wouldn't want there. If I take that money and go against my own gut feel, who should be responsible for the consequences: me, or the government?

  18. Broccoli:

    It is very simple. End market distorting policies. Align incentives with risk vs. reward. Right now the system is entirely rigged so that the rewards are shared by the few and the risks are shared by the innocent.

    Two good ways to start:
    1) End TBTF (too big too fail)
    2) Bring back shareholder and director limited liability. Maximum liability up to a certain percentage of the average value of the shares over the last 10 years. Could be as low as 5% and I guarantee you shareholders would be extremely deligent in policing the risks of the companies they own.

  19. Mark:

    "is that the default risks were under-estimated by those packaging the securities."

    And that was one of the biggest systemic problems with the industry. Why did the "system" have the SELLERS of the bonds estimating (under or over) the risk? This is analogous of having the sellers of a home do the inspection and appraisal. Clearly, the sellers will pick the inspector and appraiser that will give them the best result for them. Almost every consumer understands that it is their job, as the buyer to inspect and assess the value. A home buyer hires their own inspector and their own appraiser, and pays their costs.

    But, in the world of mortgage bonds, the buyer of the bonds just took the word of the seller as to its risk and value. This is about as dumb of a situation as can be, and frankly, anyone who would be stupid enough to take the sellers word as to the overall risk of the bond deserves to go bankrupt. They were suckers, but suckers of their own making.

  20. John Moore:

    I think the problem is that the Wall Street "wizards" have made the system so complex that nobody understands it, and worse, that it is ever more subject to chaos and to black swan events.

    At some point, the increase in "efficiency" is simply too dangerous.

    While I don't know how to do this without getting government involved in a very dangerous way, I think that these big complexes need to no longer exist. If a bank is too big to fail, it is too big to exist. If systems are too complex to understand, they need to cease existing.

    MBS's move the burden of evaluating the mortgage risk from the issuer, who has the best knowledge, to "risk analysts" who these days are smart kids from the Ivy League who both have no skin in the game, and no practical experience. As long as this happens, we will have financial chaos, enormous principal agency issues, and dangerous and complex government interventions.