Incentive Problems in the Dodd Bailout Bill.

Senator Dodd's bailout plan has some serious drafting and incentive problems.

If the drafters of the Dodd scheme were to create a game based on the scheme (assuming a liquid market and rational behavior) and play a few dozen rounds one evening, they would know that the Dodd scheme won't work unless people do things that are directly contrary to their interest in making a profit.

Section 2(c)(2)(A)(i) provides:

"The Secretary may not purchase, or make any commitment to purchase, any troubled asset unless the Secretary receives contingent shares in the financial institution from which such assets are to be purchased equal in value to the purchase price of the assets to be purchased."

A LITERAL READING

If the contingent shares must be equal in value to the purchase price of the assets, then why would most companies sell troubled assets to the government at their current estimated value?

For example, if the government pays a million dollars for some troubled assets, then the company must give the government the troubled assets plus contingent shares worth a second million dollars. If the contingency had a 50% chance of occurring, then that would mean that the company would have to turn over a contingent right to $2 million in company stock ($2 million x 50% = contingent shares valued at $1 million).

But the most that would be payable if the government lost all $1 million on a $1 million purchase would be $1,250,000 in stock (1.25 x the loss). So a 50% chance of that would be worth $625,000, not the $1 million required by the literal language of the statute. In this view, the government could buy only troubled assets that where the chance of their becoming worthless was at least 80%. This makes no sense. And what company would want to receive just a million dollars in return for a million dollars in assets plus contingent shares worth a million dollars?

A FIRST GAME

The literal reading of the statute I went through above is probably not what was intended by the drafters. What the drafters probably intended is that, in return for paying a million dollars, the government would receive the troubled assets plus a contingent right to shares of stock that would be worth a million dollars IF they were NOT subject to a contingency. Since the shares ARE subject to a contingency, these contingent shares are worth something, but not a million dollars.

Nonetheless, this approach still doesn't make much sense for reasons that are best shown by imagining a game.

The government is paying a company million dollars for 2 things: (a) a troubled asset and (b) a contingent right to $1 million in stock.

Assume a heuristic game.

In the first round of the game, let's arbitrarily assign a value to the contingent right of $300,000 (eg, a substantial chance of up to $1 million in stock). Accordingly, in return for a nominal purchase price of $1 million, the company would trade the government an asset worth $700,000 and a contingent right (to up to $1 million in stock) worth $300,000.

Assume that an hour later, the government sells the $700,000 troubled asset for the market price of $680,000. According to the way I read the statute, the government has just sustained a loss of $320,000 ($1 million - $680,000). This triggers a penalty clause that gives the government 125% of its loss in company stock, measured by the stock price in the 2 weeks before the original deal. Thus, the government now obtains $400,000 of the company's stock.

From what we've learned in the first round of the game, a contingent share right (to up to $1 million in stock) should probably be valued at significantly more than $400,000 in the second round of the game; let's say it's worth $600,000. In the second round, the government agrees with another company to pay $1 million for a contingent right worth $600,000 and troubled assets worth $400,000.

Assume that an hour later, the government sells the $400,000 troubled asset for the market price of $380,000. According to the way I read the statute, the government has just sustained a loss of $620,000 ($1 million nominal price - $380,000). Again, the government is entitled to 125% of its loss in company stock, measured by the stock price in the 2 weeks before the original deal. Thus, after the second round the government now obtains $775,000 of the company's stock.

So our valuation of $600,000 for the contingent right in the second round was too low. In the third round, let's assign a value of $800,000 to the contingent right and $200,000 to the troubled assets. Again the government pays $1 million. An hour later, the government sells the $200,000 in troubled assets for $190,000. It now is allowed to take all $1 million of company stock to make up for 125% of its $810,000 in nominal losses.

Anyone who understands how such a game progresses would never play even one round.

A SECOND HYPOTHETICAL (or GAME)

The prior example assumed that the government was selling the troubled assets in its portfolio for as high a price as it could and that it flipped the assets quickly.

In a second example, assume that the government holds the property for a few months and tries to maximize its own profits, not to maximize the resale price. Assume that the government bought a million dollars in troubled assets from each of two companies, ACME and ZED. In the two months since the purchase, ACME's stock had dropped in half; ZED's stock had doubled in price. A rational government would sell the ZED asset for LESS than it was worth on the open market, because for every $1 the government lost on the sale of the asset it had originally received from ZED, it would gain $2.50 in ZED stock (125% of the loss measured by the original price of the stock, which has now doubled).

When stock prices have climbed since acquiring the troubled assets, the worse the government does in selling troubled assets -- ie, the lower the prices it accepts from buyers — the more money it makes. This perverse incentive renders the scheme unworkable if you want actors in the Dodd scheme to maximize returns.

ANOTHER DRAFTING ERROR

The bill also contains this bizarre definition:

As used in this subsection, the term "contingent share" means any equity security traded on a national securities exchange.

Again, this makes no sense. Contingent shares of bank stocks are not generally traded on exchanges such as the NYSE or AMEX and will not be traded on exchanges under the bill.

From Crony Capitalism to Crony Community Organizing: "Profit" Loophole in Bailout Bill Doesn't Require Net Profits.

Much of the blogosphere is up in arms because of the provision in Senator Dodd's financial bailout bill that might funnel profits from the bailout plan to ACORN Housing (related to the disreputable activist group ACORN), and other more reputable service organizations.

I have read Dodd's proposed statute and in some respects, it is far worse than has been reported. Senator Dodd has placed a loophole in the bill that is explicitly designed to siphon off tens or hundreds of billions of dollars to the Housing Trust Fund and the Capital Magnet Fund even if there are no net profits in the $700 billion venture.

Here is the provision that has already been widely noted:

d) TRANSFER OF A PERCENTAGE OF PROFITS.-

(1) DEPOSITS.-Not less than 20 percent of any profit realized on the sale of each troubled asset purchased under this Act shall be deposited as provided in paragraph (2).

(2) USE OF DEPOSITS.-Of the amount referred to in paragraph (1)-

(A) 65 percent shall be deposited into the Housing Trust Fund established under section 1338 of the Federal Housing Enterprises Regulatory Reform Act of 1992 (12 U.S.C. 4568); and

(B) 35 percent shall be deposited into the Capital Magnet Fund established under section 1339 of that Act (12 U.S.C. 4569).

(3) REMAINDER DEPOSITED IN THE TREASURY.-All amounts remaining after payments under paragraph (1) shall be paid into the General Fund of the Treasury for reduction of the public debt.

The biggest problem here is that the 20% is not taken from net profits, but rather from any profit in the sale of each and every individual troubled asset.

For example, assume that the new Agency buys three troubled assets for $1 million each. One is sold for $2 million, while the other two are sold for $300,000. Thus, $3 million in investments are sold for $2.6 million, representing a $400,000 loss.

But Senator Dodd's bill does not provide for losses to offset gains: "Not less than 20 percent of any profit realized on the sale of each troubled asset" must be given to the two housing funds, so $200,000 of the $1 million profit on the one asset that made a profit must be siphoned off to the housing funds, despite the $400,000 net loss on the three deals taken together.

As an analogy, imagine a regular trader of stocks who takes lots of hedged positions and had net losses of 25% this year, but couldn't offset his gains with his losses, instead having to pay 15% income taxes only on his gains.

How much might be siphoned off under the Dodd bill? It all depends on how long the new credit Agency is in force, how often it turns over its portfolio, and how variable its returns are.

If the agency is in force for 4 years and turns over its portfolio every two months, then it would generate about $15 trillion in sales overall (650 billion x 6 x 4 = 15.6 trillion).

Let's assume that $7 trillion of sales generate a profit of $2 trillion and $8 trillion of sales generate a loss of $2.1 trillion, leaving a net loss of about $100 billion.

With a net loss, one might think that nothing would be funneled to the housing funds for service organizations, but that is not what the statute says or means. One looks only at the sales generating gains to determine the size of the payments to the housing funds. With $2 trillion in profits and $2.1 trillion in losses, the housing funds nonetheless get $400 billion dollars in "profits." (This is over 40% of a typical year's US total federal income tax receipts.) And that is the result if only 20% of "profits" are skimmed; the statute puts no upper limit on the skimming, so long as they come from profits (not net profits). Theoretically, the new Agency could potentially siphon off $2 trillion to the two housing funds, more than its $700 billion portfolio limit.

400 billion dollars may be a high estimate for the housing fund payments, but if they turn out to be only a tenth as large ($40 billion), they would still be huge. To reduce this massive skimming required by the Dodd statute, the new government Agency would have the incentive to engage in fewer transactions and do less to create a public market for troubled assets, thus significantly undercutting the chance that the bailout will work.

I was mildly in favor of the bailout until I read Dodd's proposed statute. The way that the statute is drafted is so tricky and its definition of profit is so unsophisticated and nonsensical that the statute smells more of graft than of an honest attempt to solve the financial crisis. We are moving from failed "crony capitalism" to failed "crony community organizing."

Other posts will deal with other provisions in the Dodd bill and whether ACORN Housing will actually apply for any funding.