Book Summary: House of Debt

Book: House of Debt

Author: Atif Mian and Amir Sufi

Key takeaways: Mian and Sufi talk about how household debt was a key factor in starting and sustaining the Great Recession.

HOW the Great Recession came to be: Both the Great Recession and the Great Depression were preceded by a large run-up in consumer debt, and started off with a mysteriously large drop in household spending (esp. durable good such as autos, furniture and appliances). Several studies now confirm that growth in household debt is one of the best predictors of decline in household spending during the recession.

It is well-known how consumer debt ran up before the Great Recession, but few know that the Great Depression had something similar–advent of installment financing in 1920s (for the first time in U.S. history, durable goods merchants began assuming that a potential buyer will use debt to make purchases).

Sometimes people wonder why a similar wealth destruction (dot-com bust destroyed $5T of wealth vs. $5.5 T in housing bust) did not lead to a similar effect on household spending, and the answer is that tech stocks were mostly owned by rich households with almost no leverage (6c of debt for every $1) whereas people who owned houses were heavily indebted.

Exactly WHO took on so much debt? An interesting question to ponder is: who is the marginal borrower who took on extra $7T of debt during the housing boom? As we know today, it was the so-called “subprime” population (aka those with low credit scores), and not just that, the debt that was offered increased while the income of the group declined. Furthermore, while housing prices for high credit-score home-owners didn’t grow much between 2002-07, it grew by 20% for low credit-score home-owners in elastic counties, and 70 PERCENT in inelastic counties (where new housing development is hard because of geography or other reasons).

But, but, but… Why did ANYONE lend to them anyway? This is an obvious question. It is very easy to grasp why people with low credit scores wanted to buy homes (well, because, who doesn’t?), but it is harder to understand why would anyone lend to them. This is multi-part answer with multiple timelines.

First, we have to start with the Asian Tiger crisis of the 90s, when following the collapse of the Thai bubble, foreigners started pulling out cash from the Asian countries en-masse, taking their forex (mostly USD) with them. Now, the Asian central banks, of course, couldn’t print USD, so they had to accept help from the IMF, whose help, again, of course, came with strings attached (and who likes that, ever?). SO. Asian central banks, to stave off any potential runs on their currency/banking sectors, decided to build a “war-chest” of USD, and build they did–between 1990 and 2001, central banks bought $100 bn USD annually (sounds big, but remember, world economy is a $50 T+ gig), and between 2002 and 2006, post the war-chest build resolve, the reserve accumulation septupled. Now, what does this have to do with anything? Well, all those $s have to go somewhere, and invested in securities that are deemed safe.

Now, let’s travel even more in past, and go 25-30 years before the Asian Tiger financial crisis reared its head. It’s before 1970, and the vast majority of mortgages are originated by the local banks, which means they’re susceptible to local real estate bubbles and thus risk-averse, and thus, less lender-friendly. 1970 rolls around, and that’s when the HUD (U.S. Housing and Urban Development) started its securitization program, under which the local banks sold their mortgages to couple of national entities called GSEs–Fannie/Freddie, and because the GSEs had loans from everywhere, they didn’t own risk for a given locale. GSEs packaged these loans into securities called mortgage-backed-securities (MBS), and further sold these much-less-risky MBSs to investors. To make sure that banks don’t sell awful loans to the GSEs, HUD required them to adhere to certain standards and loans made under these requirements were called “conforming loans”; GSEs also charged a fee to the banks to partially insure the MBS investors against any defaults.

Now, let’s jump forward to early 2000s again. The demand for safe U.S. investments is booming, and for a while, money is plowed into these MBSs. But because there are only so many conforming loans, there are only so many MBSs, and the Asian central bank daddies need more. So, Wall St goes to work. The problem that Wall St needs to solve is that it needs to package non-conforming loans (i.e. on average, loans made to people with lower credit score) into safe securities, because that’s what is being asked. This was achieved through, as we all know, tranching–pooling all the not-so-good loan into one pool and creating tranches where the senior tranche gets paid before lower tranches. The higher tranches are deemed safe, rated so by the rating agencies, and fed to the MBS investors who couldn’t get enough of the GSE MBSs. So now, what we have is a private-securitization market of the so-called subprime loans, and this market soars. In 2002, only 20% of all MBS market are these private-label not-so-good MBSs, and by 2006, they’re a staggering 50% of the entire market (goes to ZERO percent in ’07, btw). On the face of it, there’s nothing wrong with a private free-market solution, but this is where things go wrong: because demand of these private-label MBSs is so high, local banks are incentivized to make any kind of loans and sell it to investment banks who in turn package it and sell it to the investors in these “safe” tranches. As you can see, now local banks have no incentive to make sure that the loans are any good, and all the incentive to pump out as many loans as they can. Explains by itself how things went wrong, right?

One small thing I’d add: one could assume that the problem was simply that people with low credit-scores shouldn’t have been given the loans to begin with. That is not entirely true. In fact, actual default were correlated more with falsified documentation by the local originating banks. For examples, for some private-label MBSs, 620 was the cutoff credit score–to sort of maintain a semblance of perceived quality. Now, if someone has a credit score of 615, the originating bank has no incentive to lie for those loans because they can’t be packaged and sold away. Studies show that the default rate of people with credit score of 615 was MUCH LESS than the securitized tranches that were full of subprime loans with credit score of 620 and above. So, while it is very easy to blame the poor people with low credit-score for the whole fiasco, remember the huge part the originating banks played by simply falsifying documentation.

Debt Begat the carnage:

Mian and Sufi argue that debt played a huge part in this housing-price bubble. They cite Charles Kindleberger, who has noticed that main driver of asset-price bubbles was almost always an expansion in credit supply–an increased willingness by creditors to lend to borrowers with no discernible improvement in income growth.

Now, the question is: why is it that debt-financing is more detrimental than equity-financing? Authors say that this is because pessimist debt financiers who believe that value of something is only $8, when someone optimistic who’s willing to hold the equity at a valuation of $10 (and needs debt-financing), is still willing to finance the $10-valuation business because the debt-holders’ claim is ahead of the equity-holders. In other words, as a debt-financier, what do I care that you, equity-holder erroneously thinks that the value is more that what I think; my claim is before yours, so I am happy to finance it anyway. This is is sort of amazing. Even though debt-holder is being a pessimist here, (s)he is in effect, inflating the asset values! Also, in addition to facilitating bubbles, debt also helps sustain them–for a while–because it impacts that expectations that other buyers could use debt to finance the asset next.

There are some other points as well–

  • In many cases, expansion of credit resulted from development of substitutes for what previously had been the traditional monies: we saw that in the case of private-label MBSs didn’t we?
  • Per authors, debt instruments lead investors to focus on a very small part of potential set of outcomes. I don’t know how much I agree with that, but it’s probably undeniable that the potential set of outcomes certainly narrow–i.e. even if a large product goes bust, you’ll still probably make whole.

Policy Measures Post-Recession:

Without any external shock to the economy, today, there are the following views on why recessions happen –

  • Change in growth expectations: Apparently, people lose faith that technology will advance, or that their incomes will improve. This sounds like such a load of garbage that economists came up with post-hoc without any “correspondence between results of their theory and the facts of observation”, as Keynes would say.
  • Animal Spirits view: You see, how in previous theory, people are completely rational, but in this one they go completely mad for no reason, and are driven by irrational and volatile beliefs, leading to a self-fulfilling prophecy.
  • Banking View: This is the one that is accepted the world over by policy makers: if we could just start the flow of credit again, everything would be great again, no matter if the run-up in the debt is itself the problem. This is easily one of the first things that’s done by the central bankers (lower interest rates) and govt. officials (relax standards to let banks lend again). This is the view that says that letting Lehman go under was a colossal error, but this doesn’t sit well with the obvious data that spending incl. durable consumption and residential investment was declining for several quarters before Lehman collapsed.Often, those who support this view, cite the need of credit by small business owners as the main concern, but surveys from National Federation of Independent Business (NFIB) show that less than 5 PERCENT of small business owners were concerned with credit through the financial crisis. On the other hand, guys who cited poor sales as a concern, jumped from 10% to 35%, which shouldn’t be shocking.

    Furthermore, in a levered-losses scenario, allowing for more lending is weird: no one wants more debt and with the rising default rates, banks don’t want to lend more.

It also needs to remembered that there need not be any external shocks–many severe recessions are triggered without any obvious destruction of productive capacity.

What would have been the right thing to do:

According to authors, the government should have stepped in and helped under-water home-owners, and the MBS holders should have modified terms to allow flexibility to home-owners (this would have helped both the MBS holders and the now-underwater home-owners).

The private-sector couldn’t come to an agreement itself because:

  • Trust Indenture Act of ’39 required that modifying the “economic terms” required consent of 100% of holders, which meant that a few hold-outs could stop the changes (a good thing in most times, but because a handcuff in this case).
  • Securitization agreements didn’t compensate the servicer for cost of modification, but did do so for foreclosure–easy to see what services preferred. It has been shown that delinquent mortgages that were part of securitization pool were more likely to end in foreclosure (vs. if they were held by an individual bank), and these were, in turn, less likely to re-default and were more profitable for the lenders.
  • Servicers wanted to be “tough” to prevent more widespread defaults.

As for why the government didn’t help the underwater home-owners, we all remember the Rick Santelli rant on the CNBC floor with traders cheering him on (one of the more misguided and mean-spirited things I have seen on TV). U.S. never ended up doing anything and we are left to wonder what could have been. It’s particularly sad because U.S. did something like this in the past and it worked–it was called HOLC (Home Owners’ Loan Corp.) after the Great Depression of ’33. HOLC started giving 15-yr mortgages vs 5-yr mortgages that were common at the time, and wishing 5 years, 10% percent of U.S. homes were HOLC loans. Another thing that the U.S. did was to remove the ability of long-term debt contracts to demand gold (U.S. was obviously on the good standard at the time), and this abrogation was a in effect a debt-forgiveness of the order of entire GDP of the country.

What about fiscal stimulus? Well, first of all let’s agree that fiscal stimulus was 2/3rd tax cuts and only 1/3rd actual spend–so it was limited in its appeal and effectiveness to begin with. Furthermore, authors argue that an even better idea that fiscal stimulus should have been debt forgiveness for the most indebted home-owners as they have the most propensity to spend.

Usual Complaints About Supporting Home-owners:

People often have simplistic but misguided reasons for not helping home-owners:

  • Problem is not in my state so why should we pay?: Well, it’s one market. While auto-dealer losses were localized, auto manufacturing losses were wide-spread because spending dropped everywhere. This is to say that problems that happen outside of your borders (in a unified market) are/will become your problems too.
  • Wages will get pushed down, and people will move to where jobs are and it’ll fix itself: Just doesn’t happen. Wage-growth, adjusted for inflation, increased +1.1% between 2008 – 2011 (in part because of nominal-wage-rigidity). And people won’t be leaving their social safety net to take a bet on something complete new across the country.

What is the solution?

Authors suggest using a debt instrument that is not as inflexible as traditional debt, and where interest payments are tied to the economic well-being of the country. They recognize that it has a snowball’s chance in hell of getting enacted, but mention it as a start of a conversation.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.