Why did the U.S. Housing go bust in 2006-07?
The exact time when the home mortgage problems surfaced can now be pin-pointed as mid-2006 even though the housing problem was not fully acknowledged by the government and market players until almost summer of 2007. By mid-2006, there is now enough evidence that housing prices began to decrease significantly and default rates increased in some states such as California, Arizona etc.
There are essentially five theoretical models or frameworks that are used by economists to explain credit booms and busts. These are (1) changes in fundamentals over time; (2) irrational myopia as reflected in euphoric greed followed by fear or depressive panic; (3) implicit or explicit government subsidies and guarantees; (4) multiple equilibria or knife-edge problem; and (5) agency problems in assets management. Each of these frameworks is used to analyze the current housing problems, which triggered a U.S. financial meltdown and impacted a global economic crisis.
The ‘fundamentals’ framework emphasizes that credit cycles depend on evolving news and asymmetric information. Credit cycles reflect exogenous events which change rational expectations of future cash flows and risks among other things. There was no exogenous shock that triggered a credit crisis in 2007. There is no evidence that 9/11 attack on New York negatively impacted on credit for private investors and bankers continued to under-price risk and continue lending even larger amounts of money for mortgages during 2002-07.
Minsky-Kindleberger theories emphasized irrational myopia and herd-like behavior causing endogenous cycles as greed and fear dominate investment behavior rather than rational, long-term projection of fundamentals. While greed leads to under-pricing of risk, fear leads to over-pricing of risk. While the U.S. mortgage crisis can be partly explained by greed and fear, it does not explain the timing of the crisis, the duration of credit cycle in US and why it is of different durations in different countries, and how long will the current recession last.
The third framework explains the credit cycle as a result of government subsidies as reflected in (i) the existing tax-deductibility of home-mortgage interest payments; (ii) FHA programs which provide credit to first-time home buyers and permit up to 97 percent leverage at origination and also permit cash-out refinancing that resulted in 95percent leveraging (take-out of $520 billion each year by households through re-financing); and (iii) government financial subsidies through federal home loan bank lending for owning ‘an American dream’ and directing credit to low-income communities in line with the spirit of Community Reinvestment Act of 1977 and establishment of Frannie and Freddie. This argument points to factors exacerbating the credit crisis and mis-allocation of resources to sub-prime mortgages but does not explain why the crisis occurred in 2006-07 when all these subsidies were in place since 1977.
The multiple equilibria theory explains that credit booms and busts can occur with the same set of policies or factors, including endogenous liquidity scarcity. This does not explain the likelihood of crisis, timing or varying severity of crisis.
While each of the above theories gives a clue to the current crisis and partly explains the run-up to the crisis, a more comprehensive understanding is provided by the principal-agent framework for what happened to the credit cycle during 2001-08. Agency framework posits that asset managers (agents) bought and retained securitized instruments related to several asset classes, including mortgage-backed securities (MBS) and collateral debt obligations (CDOs) and willingly invested their clients’ (principals) money. Investors included institutional investors, pension funds, money market funds, banks and hedge funds.
The following is a short discussion on what went wrong this time around as a confluence of factors came together in a ‘perfect storm.’
- Sub-prime mortgages were the initial and main ‘bad’ assets. Rating agencies and originators of securities assumed low probability of default (PD) and low given default (LGD) which is also known as ‘severity of loss.’ Together the PD and LGD provide an ‘expected loss’ and the rating agencies and originators assumed low expected loss of 4.5 to 5 percent between 2001-04 and between 5.5 to 6.0 percent in 2006 (as indicated by Moody in 2007). In other words, assumptions of low expected loss and failure to revised it upward until early to mid-2007.
- This low expected loss (4.5 to 6 percent) was based on 2001-03 projected default loss records (ex-post actual losses for that period averaged 2% on sub-prime, with 3% on 2003 cohort of sub-prime mortgages). In addition, the information that ‘house prices never declined since World War II’ underpinned the second reasonable assumption of steadily increasing house prices.
- These two assumptions, which looked reasonable at that time, were critical in continued financing of sub-prime mortgages, especially when one considers that 80 percent was in the form of AAA debt, and 95 percent in A, AA etc by sub-prime backed MBS. There was Frannie and Freddie to buy up anything related to home mortgages.
- The reasonable assumptions however did not warrant the rapid increase during 2004-07 in sub-prime loans (based on no-docs, low-docs, ninja – no income, no assets).
- Despite negative news from some states since mid-2006, why did the rating agencies not revise upward the expected losses of 6% until mid-2007? Part of the answer is in the fact that Basel I and II as well as US government gave enormous power to rating agencies. In particular, SEC in line with Congress’ directives proposed ‘anti-notching’ regulation that gave rating agencies authority not to change rating of instruments, even if more powerful agencies, such as Moody and S & P strongly felt that the rating of a particular instrument should be changed.
- The 6% expected loss gave investors (principals) and asset managers (agents) the needed cover to continue the party – this is now called ‘plausible deniability.’ In particular, asset managers were placing someone else’s money at risk and were driven by large earnings, huge bonuses, and management fees. The asset managers were demanding more and more of securitized instruments, and the originators were only too glad to oblige. The asset managers should have been aware of the ‘race to bottom’ in terms of borrower quality and the unrealistic assumption of 6% expected loss.
- Regulation requiring banks not to hold junior tranches of securitization they originated enabled banks to reduce their loss exposure by disposing of their own securitized assets. In addition, worst sub-prime mortgages were used for newer securitized instruments (while retaining better sub-prime loans on their balance sheets). So, there was adverse selection of risks. AIG insurance of banks and financial firms for these asset-backed securities only exacerbated the agency problem.
- In contrast to credit card securitization, sub-prime mortgages were not checked for assumptions, borrower profile or quality, product type or state of the housing market (either in different states or even at the national level).
Tomorrow: Transmission of crisis from home mortgages to US credit freeze and global oil-price hike
- Advances in Development Economics