Financial Econometrics and Empirical Market Microstructure
The Second Wave: HSBC’s February 2007 Loss
The second major jump in subprime volatility occurred on February 23, 2007, the day after HSBC announced a $10.5 bn loss in their US subprime holdings. It looked like a classic exogenous Black Swan shock, as AAA spreads instantaneously tripled from 11 to 31 bp (an unprecedented 12 sd daily outlier). Spreads soon stabilized
below 20 bp again, bolstered by “relative value” trades: traders bet that increasing subprime delinquencies would hurt the first loss equity or mezzanine tranches of CDOs, but thought that the AAA rated securities were immune. And they decided to make the trades “carry neutral” and some even went so far as to characterize this as a “hedge.” If BBB’s were trading at 200 bp and AAA’s were 20 bp, shorting $1 bn BBB meant buying $10 bn AAA. This was history’s most penny wise and pound foolish trade, and would later explain Howie Hubler’s record setting $9 bn loss at Morgan Stanley.[18]
Evidence of a housing bubble burst continued to mount, as prices fell and subprime delinquencies spiked. Subprime lender NEW’s default in March 2007 was no surprise to students of financial statements: 5 months earlier forensic accountant CFRA published a report that they had obfuscated rising delinquencies in their June 2006 earnings release. In May 2007, two Bear Stearns subprime bond hedge funds imploded, and the following month Merrill Lynch failed to sell their AAA rated CDO collateral (bonds they had sold to Bear and held in their own inventory). But even this failure could not shake the market’s confidence, as credit spreads and VIX continued to hover at historical lows.
Astonishingly, it took until the week of July 9th, 2007 for S&P, Moody’s and Fitch to announce their first subprime CDO downgrade. Only then did risk awareness enter mainstream consciousness, and waves of successive selling followed. Within a few weeks spreads rose to 150 bp before tightening once again to 50 bp, and then widening to 250 bp, and eventually spiking to 400 bp by early 2008.
See Fig. 10 of absolute spread levels, which reveals classic fractal amplification patterns different time scales: from daily ripples to weekly waves to monthly tsunamis.
When looking at the subprime spread chart above, ask yourself when risk was highest. According to VaR, risk peaked in 2008. But when considering hidden risk, the most dangerous time was 2006-2007 as the chase for ever narrowing spreads led banks to unprecedented leverage that would threaten the entire global economic system.
