In this environment, any rational card player would stop making bets until the true solvency position of each player is revealed and the bankrupt ones are expelled from the game. Having insolvent players sitting at the table spoils the game.

This is what is happening in the banking system -- only worse, because in poker you would only fail to collect the pot if you played with an insolvent player, while in the banking system you would lose your bets if you lend to an insolvent bank. Liquidity preference will not subside until the losses are made explicit, written off and absorbed.

To achieve this you simply let the illiquid borrowers and their financiers go bankrupt. This is how financial crises were solved in the 19th century. The method was expensive because commercial banks are central to the operation of the payments system and their uncontrolled failure can cause enormous damage to the economy.

Yet the alternative, continuing to play poker with failed players, also causes enormous damage. This is because the resources that could be used to spur economic recovery are not allocated due to lack of information. Worse still, resources are taken from the efficient to keep insolvent banks and companies operating.

The government must allow bank losses to come to the surface. The $700 billion rescue fund can be used for this purpose -- absorbing the losses and weeding out the failed players. How do we accomplish this? By forcing the failed commercial banks to write off their losses then asking the owners to compensate for the losses with new capital. If they don't do it, the government takes over and recapitalizes the banks by purchasing the bad loans at their nominal value (as if they had not been written off). This no longer benefits the previous owners, who have already abandoned the table. The government then sells the banks, making a loss equal to the write-offs.