Subprime recession outline 2010 (following Mizen)
John Eckalbar
I’m not interested here in who is more at fault for the current financial crisis—Democrats or Republicans. Anyone who tries to pin this on one party to the exclusion of the other is more interested in spin and political advantage than in facts. Both parties behaved badly as did a great many lenders and borrowers, the Fed, the regulators, the rating agencies, the banks, and so on. Just a couple examples on both sides are sufficient:
Democrats…
…downplaying housing bubble and financial risk… Listen for yourself to what they say:
http://www.youtube.com/watch?v=iW5qKYfqALE
http://www.youtube.com/watch?v=9HQWk1Wp3L4
http://www.youtube.com/watch?v=ahtuNt3AKCg
…or pushing banks to make loans to bad credit risk borrowers
Community Reinvestment Act of 1977 or support of Acorn
(See Peter Swan, “the Political Economy of the Subprime crisis,” European Journal of Political Economy, 25 (2009), 124-132.
Republicans
…They obviously have a great deal of blame, since they held the bulk of the power during the housing bubble after January 2001.
Enough of politics, now to the economics.
Consider this: In the old days, when a lending institution made a mortgage loan, it typically held the loan to maturity, so it had every incentive to investigate the borrower and estimate the probability that the loan would be repaid. Mizen calls this “originate and hold.” But that model gradually changed, as you see…
The process of “securitization” (bundling and re-selling loans or “originate and distribute, as Mizen calls it) began in 1968 when the Government National Mortgage Association (GNMA or Ginnie Mae) securitized FHA (Federal Housing Administration) and VA (Veterans Administration) mortgages, which were backed by the US govt. These mortgages were then bundled into securities and sold to investors. The advantage of this is that relatively small investors (plus some huge investors) could buy into a diversified pool of carefully selected, government backed mortgages, and this could aid the flow of funds into housing. The Federal National Mortgage Association (Fannie May) started issuing mortgage backed securities (MBS) in 1981, and soon other private entities were creating MSB using prime, but not government-backed, securities.
Says Mizen, p. 537: “Securitization was undertaken by commercial and investment banks through special purpose vehicles (SPVs), which are financial entities [separate companies, je] created for a specific purpose—usually to engage in investment activities using assets conferred on them by banks, but at arm’s length and, importantly, not under the direct control of the banks.
The advantage of their off-balance sheet status allows them to make use of assets for investment purposes without incurring risks of bankruptcy to the parent organization (see Gorton and Souleles, 2005). SPVs were established to create new asset-backed securities from complex mixtures of residential MBSs, credit card, and other debt receivables that they sold to investors elsewhere. By separating asset-backed securities into tranches (senior, mezzanine, and equity levels), the SPVs offering asset-backed securities could sell the products with different risk ratings for each level. In the event of default by a proportion
of the borrowers, the equity tranche would be the first to incur losses, followed by mezzanine
and finally by senior tranches. Senior tranches were rated AAA—equivalent to government debt. In addition, they were protected by third-party insurance…”
(from companies like AIG. The insurance was called “Credit Default Swaps.”)
“A market for collateralized debt obligations (CDOs) composed of asset-backed securities
emerged; these instruments also had claims of different seniority offering varying payments.
Banks held asset-backed securities in “warehouses” before reconstituting them as CDOs, so
although they were intermediating credit to end investors, they held some risky assets on their balance sheets in the interim. Some tranches of CDOs were then pooled and resold as CDOs of CDOs (the so-called CDOs-squared); CDOs-squared were even repackaged into CDOs-cubed. These were effectively funds-of-funds based on the original mortgage loans, pooled into asset-backed securities,…”
Many of these securities were bought by bank-formed entities called SIVs (Structured Investment Vehicles) and were held off the banks books, with little regulation or capital requirement…. Securitization of subprime mortgages started in the mid-1990s, by
which time markets had become accustomed to the properties of securitized prime mortgage products that had emerged in the 1980s, but unlike government or prime private-label securities, the underlying assets in the subprime category were quite diverse.”
So we ended up with a bunch of financial firms, pension funds, sovereign wealth funds, and so on holding securities backed by bundled loans, including very low quality sub-prime loans.
Definitions of types of loans: Alt-A mortgages do not conform to the Fannie Mae and Freddie Mac definitions, perhaps because a mortgagee has a higher loan-to-income ratio, higher loan-to-value ratio, or some other characteristic that increases the risk of default; and
(iv) subprime mortgages lie below Alt-A mortgages and typically, but not always, represent
mortgages to individuals with poor credit histories.
Why would a lender make a subprime loan to a borrower with low probability of repayment? Because the lender could make a quick buck selling the loan upstream through brokers, originators, arranger/issuer, SIV, Funds, End investors. Why would the end investors buy these? Because safer investment paid less and the investments were secured by houses, which always go up in value, right? Plus their financial mathematical models and the models of the rating agencies assumed that the correlation between the probability that borrow A will default is independent of the probability that borrower B will default. Which sadly turned out not to be true when the housing bubble burst, house prices started to fall, and mortgage default began to rise.
Leverage
Suppose you buy a financial asset for $100, and it rises by $5 or 5% to $105. Good. The worst case is it might have fallen to zero, and you’d have lost your entire $100. Now suppose you put up $100 and borrow $900 more, then buy a $1000 financial asset. (You are now said to leveraged at 9:1, or “nine to one.”) If your investment goes up 5%, or $50, you make a 50% return (before interest expense) on your $100 investment. Great. But leverage can cut both ways. If the $1000 financial instrument fell by only 10%, or $100, your investment would be wiped out entirely. A large part of the problem from the subprime crisis is that many financial firms were leveraged as much as 34:1.
Note that we are just considering one investment here. If you are leveraged and you take a loss larger than your original investment, you can pay the loss off and not be insolvent if you have sufficient capital at the start. Hence the importance of capital/asset ratios.
As parts of the developing world saw rapid growth rates in recent decades, their savings accelerated and they looked for investments in the developed world, including, obviously, the US. Some refer to this as the world savings glut. At the same time, the US savings rate was falling.
Mizen, p. 534.
Here’s an update for the US. Note that our savings rate has increased markedly during the recession.
And our debt to income ratios were rising. Mizen, p. 535.
And an update for US. Note here that individuals (but not government) are working their debt load downward now.
The Fed also shoved the federal funds rate very low and held it low in the early part of the 2000s, then ran it up prior to the housing collapse.
The low rates tended to drive investors toward riskier assets as they searched for yield.
Housing prices rose almost steadily from the late 1980 to 2006, and home ownership grew from 43% in WWII, to 62% in 1960 to 69% in 2006. Politicians of both parties regarded this as a good thing. And policies to enhance “affordable housing” were widely discussed.
Here’s a link to the Case-Shiller index:
http://www.macromarkets.com/csi_housing/sp_caseshiller.asp
Post 2001, sub-prime mortgages grew very quickly.
Here’s a graph of home ownership rates:
And how house prices relate to foreclosure:
From J. B. Taylor, “The Financial Crisis and the Policy Response,” NBER No. 14631, Jan. 2009.
Mizen, p. 537.
Both prime and sub-prime mortgages were bundled into mortgage backed securities (MSB). These MSBs were often both designed/built AND given risk ratings by the rating agencies: Moodys, Fitch, S&P. There is much discussion about conflict of interest here.
17. Rate of serious delinquency on residential mortgages,
by type of mortgage and type of interest rate, 2000–10
NOTE: The data are monthly and extend through January 2010. Seriously
delinquent loans are 90 days or more past due or in foreclosure. The prime
mortgage data are representative of all residential mortgages, not just those
held by commercial banks. The subprime mortgage data cover only
securitized loans.
SOURCE: For prime mortgages, McDash Analytics; for subprime
mortgages, LoanPerformance, a division of First American CoreLogic.
From Federal Reserve Bulletin, May 2010, pp. A1ff.
“Conduits and SIVs had funded their purchases of CDOs and other securitized assets by issuing their own asset-backed commercial paper (ABCP) at short maturities. The expansion of mortgage related ABCP issuance accounted for half the growth in the commercial paper market in recent years. The ABCP needed to roll over periodically, usually monthly, but as investors were less willing to purchase short-term paper in the capital markets, these entities could not obtain the necessary short-term funding from these markets. Figure 5 shows that ABCP issuance peaked in July 2007 and fell sharply in subsequent months.” Mizen, p. 542.
Mizen, p. 541
Mortgage delinquency rates for both the primes and subprimes started rising in 2005, though rates were much higher in subprimes.
Mizen, p. 540.
Mizen’s list of issues (pp. 550 ff):
“Poor Incentive Structures Under “Originate and Distribute” Banking.
First, brokers and agents of banks selling mortgages were motivated by up-front fee income
unadjusted for borrower quality. The bonuses rewarded growth of business over a short time
scale (typically a yearly cycle) with no penalties if subsequent developments revealed a lack of due care and attention in the origination process or losses to the originator.
Second, originators had no greater incentive to look more carefully than brokers at borrower
quality.
Third, the profits from securitization created incentives for originators to obtain new loans
regardless of their quality provided they met minimum standards for resale. As the quantity of new borrowers declined, lenders reduced their standards to maintain the volume of loans feeding into the securitization market.
Fourth, tranching enabled the SPVs to construct products with ratings suitable for certain
types of investors.
Fifth, ratings agencies made a large share of their profits from rating structured finance products; for example, Portes (2008) reports Moody’s generated 44 percent of its revenues from these activities. There was scope for conflict of interest within ratings agencies because they were paid an up-front fee by the issuer to provide a rating of the assets. At the same time, though, the same business would sell advice to clients (for another fee) on how to improve those ratings, identifying “tranching attachment points” to make sure the
securitized assets just attained the required rating for the intended investor group.
Sixth, fund managers, like brokers, were motivated by bonuses and usually on a competitive
basis relative to their peers. CDOs offered a simple means to enhance portfolio performance,
which generated bigger bonuses and improved the performance of funds offered to the public.
Provision of Information. In many respects, the provision of information and the regulations
concerning information lie at the root of the 2007-08 credit crunch. The observed change in
banking practice toward originate and distribute models has greatly altered the incentives facing the originators of loans, and information about the risks associated with the assets was lacking but regulators and investors were slow to pick this up. Not only does a lender who intends to sell the securitized loans face less incentive to diligently examine the quality of the borrower, or the collateral against which the loan is made, but there is an information asymmetry between the seller of the securitized assets and buyer that cannot easily be overcome by organizations such as the ratings agencies.
Complexity in the Assessment of Risk. It seems surprising that investment banks accustomed
to dealing with complex assets could be convinced that AAA-rated assets could command
returns that had such large spreads over risk-free assets such as Treasuries without being inherently more risky. Perhaps Chuck Prince was right that investment banks knew the risks but were prepared to continue to “dance” while money was being made. For a less sophisticated class of end investors, several factors made risk assessment more complex and difficult.
Reliance on Ratings to Assess Asset Quality.
Given the complexity of the products offered, investors relied on ratings provided by ratings
agencies such as Moody’s, Standard & Poor’s, and Fitch. These ratings indicate the likelihood of default on the product, and for the highest ratings—AAA—the likelihood was equivalent to government debt default for developed economies (i.e., negligible). The granting of AAA ratings to asset-backed securities meant many investors believed they were buying very safe assets, and certain organizations such as pension funds,
which face restrictions on the assets they are permitted to purchase, were able to buy these assets. These risks were not properly priced…”
Don’t forget to add leverage to this list.
And Credit Default Swaps (CDSs). CDSs involve “three parties. What happens is that if a borrower defaults, the third party purchases the debt, pays the remaining interest to the lender. So the risk of a bad loan is transferred from the lender to the third party and the third party charges a premium.” (From another glossary: http://www.soxfirst.com/50226711/subprime_explained_crunch_time_glossary.php) This is what brought down AIG (or the US taxpayer, depending upon how you look at it)…it insured MBS.
Models and risk…risk was assessed by mathematical models, which were no more accurate than their assumptions, as always. The critical erroneous assumption was that the underlying asset, houses, would not go down in value. In other words, the assumption was that there is no correlation between the probability that one borrower will default and the probability that another would default. In other words, the assumption was that if you have a pool of borrowers, the probability that one will default is independent of (not correlated with) the fact that another borrower just defaulted, i.e., there is no feedback loop from one default to another.