The balance sheet will break out assets and liabilities on a specific basis and you can clearly see where the banks got burned - mortgages, mortgage-backed, and asset-backed securities, on both the assets and liabilities -- basically, assets which the banks clearly didn't know how to count. (See Merrill Lynch's 10-K as an example.) For ML, there were massive losses in securities financing transactions, mortgage/asset-backed securities, and considerable losses on derivatives in 2008. The summarized balance sheet clearly shows what happened -- high leverage levels means that it only takes a 3% drop to wipe out shareholder equity (for ML, it was barely 3% - $667.5b in assets against $20b in equity) and ML saw a 34.56% decline in assets FY08 ($1t in FY07 to $667.5b in FY08). They got the leverage to 13.18 in Q1 2009 (down to 13.18 on $569.8b assets, $529.6b liabilities, $40.2b equity) which gave them a 7% cushion, but with a 14.6% decline in assets during the quarter. Profits and share issuance can help raise the equity and counter a drop in assets, but you're pretty much screwed trying to make up a 34% decline.
The ratio for ML reached its peak at the end of 2008, as the subprime mortgage market cratered:
- 2004 - 20.02
- 2005 - 19.13
- 2006 - 20.57
- 2007 - 30.94
- 2008 - 32.37
- Q1 2009 - 13.18
There's always been a race between big financial firms to beat each other to the very last penny (the concept of "flash" trading, for instance, has a hint of desperation in it) so one by one they decided to out-leverage each other to bring in bigger profits faster. It's a risk management problem -- ML bet their future in FY06 on continued success in prime brokerage and securities financing, as well as commercial and residential mortgage loans and long-term debt, while ignoring a 18% drop in equity due to a net loss in continuing ops, stock repurchases, and dividend payments of $1.40/share.