Over at Marginal Revolutions, Alex Tabarrok notes that " Many of the major banks are insolvent", and links to NYTimes article here.
Actually, the article notes that if a 'stress test' is applied to banks, using 2 years of future losses, then these banks are near insolvent. They pulled in a credit firm, CreditSights, to generate the stress test. They produced some really large losses, like which generates some really large numbers, like $119B for Wells Fargo, which isn't inconceivable given they have about $1.3 Terabucks on their balance sheet.
But invsolvency is not "insolvency under a stress test", any more than you can say someone is "dead" if they are dead assuming they were run over by a bus. Further, without any details on the stress test, you might as well say "I pulled this number out of a monkey's butt--but I can still give you three significant digits!"
What where the default and recovery rate assumptions? What assets had mark-to-market losses in the stress tests. Break them out by product type to sufficient level of granularity (eg, a subprime loan where the owner has no equity is different than a conforming mortgage where the owner has a 50% equity stake). Without such additional information, their numbers are meaningless.