Two IMF economists have done a study of 124 systemic banking crises around the world between 1970 and 2007, looking in depth at 42 in 37 countries. Link to IMF Report.

The conclusions from this research are;

1) Speed matters. The faster the response, the cheaper the eventual cost. Best example of what not to do; Japan during the Heisei Malaise.

2) "Forebearance", i.e., permitting essentially insolvent banks to keep operating results in a deeper bad debt hole, crippling tax burdens and a more severe credit contraction--i.e., suspending mark-to-market accounting is not a good idea.

3) Recapitalization using public (taxpayer) money, which is often partly or fully recouped through subsequent re-privatization, usually means a smaller hit to taxpayers. The average cost of a systemic banking crisis is 6% of GDP, which in the US case would be $850 billion, not $700 billion.

4) Total fiscal costs are higher, averaging 13% of GDP (or $1.8 trillion for the US), with an average recovery of public (taxpayer) outlays of 18% gross, i.e., $324 billion in the US case.

5) Costs will be incurred regardless, either through fiscal (taxpayer costs), or reductions in GDP/output (i.e., recession, increased unemployment, etc.).

Thus the idea put forth by conservative Republicans that mark-to-market should be suspended is not a good idea, nor is buying toxic assets to remove them from bank books without nationalization or some form of equity/ownership by the government.
To the US voting public, the message should be "this is going to cost you, regardless if we use your tax dollars, or allow the credit markets to freeze up and companies to go bankrupt like dominoes, creating a deep recession and substantially higher unemployment."

A real-life example from the Great Depression comes from Meyer Mishkin, grandfather of Columbia economics professor Fredrick Mishkin (link here) . As the story goes, Meyer Mishkin owned a shop in New York that sold silk shirts to workingmen in 1929.