There has been such a flood of media coverage of the financial crisis that it is best to begin with some very simple, basic points.
Banks (broadly defined to include investment banks and the many other lenders) borrow--bank deposits, for example, are loans to banks--and then lend out what they have borrowed. As a result, their loans are much larger than their capital assets (cash, a building, etc.). If their capital shrinks in value, they have less protection against the possibility that the loans they make will not be repaid in full. If a bank's capital is 10, and it borrows 100 and lends 100, and the persons or firms it lends to return only 90, its net worth will fall to zero (10 [its capital] + 90 [the value of its loans] - 100 [the amount it owes its depositors] = 0.
Banks in recent years have increased the ratio of their loans to their capital because borrowing costs were low and financial experts thought they had discovered ways of reducing the risk of leverage (that is, of borrowing). Many of the loans were mortgage loans, and the value of those loans fell when the housing bubble burst. (Risky, and in some cases deceptive, mortgage practices had contributed to the bubble.) What made the situation worse was that rather than retaining the mortgages that they originated, banks (especially the major ones) sold the mortgages in exchange for securities backed by the mortgages. Those securities became a part of a bank's capital. The value of the securities depended on the value of the mortgages that the entity issuing the securities had bought; those mortgages were the entity's assets. As that value fell, the bank's capital fell.
The mortgage-backed securities achieved geographical diversification of mortgage risk. But the housing bubble, though not geographically uniform, was sufficiently widespread that geographical diversification did not reduce the risk of mortgage defaults sufficiently to avert the fall in the value of mortgage-backed securities.
A complicating factor was that the value of those securities was and is very difficult to determine, because each security represents a share in pieces of many different mortgages. The bank that owns the security cannot readily determine the value of all those different mortgages, since it has no direct relationship with the mortgagor, having sold the mortgage to the entity that issued the mortgage-backed securities.
Because the banking industry (and remember that I am defining "banking" very broadly, basically as all lending) was highly leveraged, and because much of its capital consisted of securities very difficult to value, the bursting of the housing bubble reduced the capital of the banks, but by an unknown amount. The reduction and uncertainty have curtailed lending by reducing the capital cushion that a bank needs to reduce to an acceptable level the risk that some of its loans will not be repaid. That is the "credit crunch,‚Äù and it is painful because so many individuals and businesses borrow to finance their activities.
Ordinarily one would expect a credit crunch to be self-correcting. As lending dropped because of the fall in bank capital, interest rates would rise and this would attract more capital to the financial markets. We have seen this process at work in Warren Buffett's $5 billion investment in Goldman Sachs. Buffett has capital, Goldman needs it, so Buffett gives it to Goldman in exchange for preferred stock (which is really a type of bond but one that does not have a term--it is never repaid) paying a handsome interest rate.
But Goldman is pretty healthy. Many lenders have so much of their capital tied up in mortgage-backed securities or other novel forms of capital that are difficult to value that they cannot attract new capital at a price that would enable the lender to continue in business. The sale of the securities would just expose their lack of value. The federal government, however, has essentially unlimited capital because of its taxing power. It is prepared at this writing to contribute perhaps as much as a trillion dollars to rebuild the capital of the banking industry. The Treasury wants to make this contribution in the form of buying the dubious securities, but that seems to be a mistake, unless pressure of time allows for no alternative. If the Treasury pays the actual value (if anyone can determine what that is) of the securities, it will not be injecting new capital into the banking industry, but merely swapping one form of capital for another. If the Treasury pays more than the securities are worth, then it is contributing capital to the industry all right, but it is also enriching the owners and managers of the banks, which creates the familiar moral hazard problem as well as upsetting people by rewarding careless management practices. The more it overpays, the most costly the bailout plan to the taxpayer.
A more palatable approach would be for the government to drive a Warren Buffett style hard bargain, in which, rather than buying anything from banks, the government would invest in them in a form, such as purchase of newly issued preferred stock, or bonds with a long maturity, that would augment the banks' capital and thus enable banks to make more loans. That would avoid conferring a windfall on the banks by overpaying them for their bad securities; no one thinks Buffett is conferring a windfall on Goldman Sachs. After the industry was back on its feet, the government could sell the bank stocks or bonds that it had acquired.