The general consensus on what caused the Great Recession can be summed up as "bad banks full of bad bankers did bad things." This column argues, however, that this narrative doesn't fit the facts. And worse, it diverts attention from the real problem, which was regular use of a bad banking system—a banking system built to fail.
Everyone knows what caused the Great Recession. Bad banks issued bad mortgages. Bad bankers over leveraged. Bad shadow banks evaded regulators. Bad rating companies over-rated securities. Bad regulators slept at the wheel. Bad households drove up house prices. Bad derivatives expanded. Bad traders overtraded. In sum, bad banks full of bad bankers did bad things.
Some bad banking is a constant. But this time was different. Virtually all outstanding mortgages were subprime and virtually all subprimes were fraudulent no-doc, liar, and NINJA (Note 1) loans. Bank leverage reached record levels. Massively bribed rating companies gave triple As to securities that were triple Fs. Regulators were totally outgunned, outnumbered, and out of touch. House prices soared forming an incredible bubble. Derivatives became "weapons of mass destruction (Note 2)." Trading grew exponentially. And well-greased politicians looked the other way. The Financial Crisis Inquiry Commission summed it all up in two words—"pervasive permissiveness" (FCIC 2011).
There's just one problem with this narrative. As I argue in Kotlikoff (2018), it doesn't fit the facts. Worse, it diverts attention from the real problem. The real problem wasn't rampant misuse of a good banking system; it was regular use of a bad banking system—a banking system built to fail.
Structural failures have structural causes. The Hindenberg had a short circuit. The Challenger had faulty O-rings. The Titanic had unsealed bulkheads. The I-35 Mississippi Bridge had inadequate gusset plates. Our banking system had leverage and opacity. It failed colossally. It was bailed out, rebuilt to original spec, and is set to die another day.
All structural banking models feature multiple equilibrium, broadly defined. Whether it is people running on banks or banks running on banks, bank runs trigger firing runs. Firing runs referencefiring someone else's customers for fear others are firing yours. Firing runs exacerbate bank runs, producing a vicious cycle and flipping the economy from a good to a bad state (equilibrium).
The FCIC didn't reference a single economic analysis of banking, let alone multiple equilibrium. There was no mention of Keynes (1936), Bryant (1980), Diamond (1982), Diamond and Dybvig (1983), Shell (1989), Peck and Shell (2003), Bikhchandani et. al. (1992), Cooper (1999), Chamley (2004), Goldstein and Pauzner (2005), Bebchuk and Goldstein (2011), or any other classic multiple equilibrium study that would have screamed "SYSTEM, NOT OPERATORS!" No surprise. The FCIC was built to miss and retain the forest. It had ten commissioners. Only one was an economist and he had no background in banking.
If the Great Recession's accused villains held smoking guns, everyone would know who shot the sheriff. They didn't. Lo (2012), after reviewing 21 Great Recession books, concluded that none agreed who done it. Makes sense. The facts clear them all.
Subprime losses were too small to produce a recession, let alone a ‘great’ one. Indeed, during the Great Recession, the subprime foreclosure rate peaked at only 15% (Note 3). Since at most only 14% of outstanding mortgages during the Great Recession were subprime, at most only 2.1% (.15 x .14) of all mortgages during the recession represented foreclosed subprimes.
Furthermore, if subprimes were terrible assets, the Fed's $29 billion purchase of Bear Stearns' worst subprimes—a clear overpayment to bribe JP Morgan to buy Bear—would have produced a major loss. Despite the subsequent Great Recession, these worst-of-class subprimes didn't lose a penny. Indeed, the investment repaid $31.5 billion (Note 4).
The ‘housing price bubble’
The FCIC included "an unsustainable rise in housing prices" in its top-seven causes of the Great Recession. Prior to 2007, real house prices rose steadily for 32 years. The rise was slow—64% compared to 170% for real GDP. Yes, roughly a third of this increase occurred in from 2003-2007. But during this period, real house prices rose by only 2 percentage points more per year than did real GDP. One can construct models with real housing prices staying fixed and the quantity of houses rising or with the opposite. Given the increasing urbanisation of the country, the fixed supply of central city land, and the remarkable foreign demand for US housing (Note 5), an irrational bubble isn't needed to explain the pre-Great Recession real price rise. Indeed, the FCIC's "housing-price bubble" could be called "normal increases following years of abnormally low increases." Moreover, housing price changes simply redistribute between sellers and buyers, but have no impact on the economy's real wealth. Hence, the alleged bursting of the ‘housing price bubble’ did not constitute a real shock to the economy.
The FCIC report states that failures of the big-three rating companies were "key enablers of the financial meltdown." But Benmelech and Dlugosz (2010), who studied the rating of 180,000 CDO tranches, concluded, "It is not clear that rating shopping led to the ratings collapse as the majority of the tranches in our sample are rated by two or three agencies" (p. 203). Since structured-finance securities represented only 35% of the US bond market in 2008, since only 7% of these securities were re-rated, and since, at most, 20% were over-rated due to ratings shopping, overrating affected less than one half of one percent of the US bond market. In addition, this figure overstates the importance of ratings shopping as the downgrades were caused by the Great Recession—i.e. why re-rate and lose the next bribe unless you are forced to by the market?
A stable fable regarding the run-up to the Great Recession is that banks dramatically increased their leverage. Not so. Fed data show bank leverage falling from 1988 through 2008 (Note 6). Equity rose from 6% of bank assets in Q1 1988 to 10% in Q1 2008. Leverage was also not particularly high in either Bear Stearns or Lehman (Note 7).
According to Christopher Cox, former Chair of the Securities and Exchange Commission, Bear Stearns was well capitalised when it failed, with a capital ratio over 13% and a debt-equity ratio of 6 to 1 (Note 8). Cox stated:
"The fate of Bear Stearns was the result of a lack of confidence, not a lack of capital. … at all times until its agreement to be acquired by JP Morgan Chase …, the firm had a capital cushion well above what is required to meet supervisory standards… (Note 9)."
In the event, Bear's actual capital ratio didn't matter. Multiple equilibrium mattered. Creditors, past and prospective, came to believe, based on innocent and guilty rumours, that other creditors were pulling the plug. This did likewise.
Lehman was also well capitalised prior to its demise. It had tier-1 capital of 11% when its creditors pulled the plug (Note 10). An 11% capital ratio is close to the current banking system's tier-1 capital ratio of 12.3% according to the Fed's recent stress tests. Thus, today's banking system is no safer than the day Lehman was driven out of business (Note 11).
Another Great Recession ‘smoking gun’ is the pre-recession run up of mortgage debt, which roughly doubled between 2002 and 2007 (Note 12). But the increase in borrowing to purchase homes wasn't associated with a rise in household consumption relative to GDP. Instead, Americans borrowed to invest. And although the ratio of mortgage debt to household net wealth rose, the rise was minor. So too was the rise in debt payments relative to personal income (Note 13).
The reigning narrative—that derivatives were overrated, complex securities sold to naïve investors—doesn't jive with Ospinal and Uhlig (2018). They examined 8,615 2007-2013 residential mortgage-backed securities (RMBS), almost all of which were rated AAA. Three quarters of the AAA-rated RMBS had essentially zero losses through 2013. On a principal-weighted basis, the average loss rate was only 0.42%. Most striking, AAA-rated RMBS out-performed the universe of AAA-rated securities.
Repos are charged with helping dramatically raise bank leverage ahead of the Great Recession. But, again, bank leverage didn't rise. Yes, repos rose—by roughly $27 billion—in the months prior to the recession (Note 14). But the rise was trivial—just 0.3% of outstanding total bank liabilities (Note 15).
No skin in the game
Fahlenbrach and Stulz (2011) examined pre-Great Recession executive compensation contracts of 95 banks. The stock and option compensation in these contracts exceeded wages by a factor of eight. The authors also state:
"Banks with higher option (and bonus) compensation … for their CEOs did not perform worse during the crisis. Bank CEOs did not reduce their holdings of shares in anticipation of the crisis or during the crisis. Consequently, they suffered extremely large wealth losses in the wake of the crisis."
Jimmy Cayne, Bear's CEO, is an example. He lost $1 billion. In short, bankers had plenty of skin in the game.
The main job of bank regulators is to oversee bank leverage. Since bank leverage was not historically high, indeed it fell in the advent to the Great Recession, regulators did their job. What they didn't, and couldn't, do is prevent our unstable economy from switching equilibriums.
Democratisation of finance
Politicians, some say, forced Fanny and Freddie to support too much risky loans to the poor. But for this to be a major cause of the Great Recession, losses from subprime mortgage foreclosures would need to have been much larger.
Fed interest rate policy
In the 1990s, the expected real 30-year mortgage rate averaged 7.91%. It averaged 6.27% between January 2000 and December 2007 (Note 16). This decline is too small to matter. Furthermore, the Fed doesn't directly control long-term nominal rates, let alone real mortgage rates. As for 5/1-year adjustable-rate mortgages (ARMs), its real rate averaged between 5% and 6% in the two years preceding the Great Recession. Real rates of this magnitude are not low (Note 17).
Bank failures have a special midwife—opacity. Opacity permits misinformation to spread and to be spread. Bear was among the first to be picked off by short sellers because it was viewed as particularly opaque. According to Cohan (2010), no one on the street or inside the bank really knew what its assets were worth. The fact that Bear's stock was valued at $60 per share one week before it was sold for $2 per share says that its asset valuation was a matter of pure conjecture. Apparently, before it didn't, the market thought Bear's assets were worth something because everyone else thought its assets were worth something. Such self-fulfilling prophecies is the stuff of multiple equilibrium.
Lehman's CEO, Richard Fuld, publicly testified, "… what happened to Lehman Brothers could have happened to any financial institution (Note 18)." The facts support his view. Bankers didn't destroy the banking system. The banking system destroyed the banking system. It operated in the dark and it operated with leverage. That, plus rumours of rampant malfeasance, brought the entire house of cards tumbling down.
The take away is that banking can't be fixed with cosmetic reforms, such as the US Dodd-Frank reform or the UK's Vickers Commission Report (Note 19). What's needed is a system with zero leverage and full, government-supervised disclosure. Why zeroleverage and fulldisclosure? The answer is simple. You can't be a little bit pregnant. Any degree of leverage and opacity invites multiple equilibrium.
In Kotlikoff (2010) I provide a very simple means, called Limited Purpose Banking, to fix banking for good. Limited Purpose Banking would transform all financial corporations into 100% equity-financed mutual fund holding companies subject to full and real-time disclosure, effected by private firms working exclusively for the government.
This article first appeared on www.VoxEU.org on November 28, 2018. Reproduced with permission.