Finding the Way to Shambala : How to Intelligently Use the Tarp and Rescue the Financial System without Getting Lost in Myths

The implementation of the Troubled-Assets Relief Program - TARP will be a great challenge. Implementing authorities should focus on three issues: (a) conflicting policy orientations received from Congress, (b) the asymmetry of information between the government and the holders of troubled assets, and (c) the complex and sometimes perverse incentive structure under which government agents buying the assets will operate.This paper does two things. First, it questions the excessive faith in the isolated operation of four elements of the TARP: (1) the improved capacity of the Treasury to determine the value of troubled assets, (2) the use of auctions, (3) the external oversight and accountability clauses, and (4) the devices that allow the Treasury to hold in its portfolio troubled assets and shareholdings in financial institutions. Second, it proposes an implementation plan that emphasizes (a) a diversity and plurality of structures with more complex ownership, capital and control schemes; (b) sequencing; (c) mechanism and incentives scheme design.The implementation plan to be unveiled relies on a 4-stage timeframe and three different investment structures: (1) a government-managed fund, (2) government-sponsored hedge funds, and (3) leveraged private equity funds. Each structure should be deployed in the stage indicated by timeframe established in order to maximize gains and attain the policy goals sought. The first stage should seek to unfreeze and unclog the market, mildly recapitalize financial institutions, and create confidence; the second should seek to reignite the markets, make them move faster and correct some incentive problems; the third stage should seek to improve incentives, reduce taxpayers' exposure and increase private investors' participation, speed up the market, and boost liquidity; the fourth stage should seek to reduce risks for taxpayers and start recovering the investments/losses associated with the program