Could be raining.
That (apologies to Marty Feldman) is my take on the current iteration of the bailout process.
Actually, it could be a lot worse. The egregious “affordable housing” (AKA let’s dump huge amounts of money in the same ratholes that helped get us into this mess) component of the Democrat proposal is gone. (This is the part that was pushed most aggressively by the usual suspects, Frank & Dodd. What’s next? Willie Sutton to negotiate federal sentencing guidelines for bank robbery? Al Capone to bargain over liquor distribution laws?)
But that is damning with faint praise. As for what’s actually in the proposal? Color me less than enthusiastic.
Everybody involved made a big hullabaloo over the inclusion of warrants in the deal. Hello, finance 101 calling. Those selling troubled assets to the government will take into account the value of the warrants when calculating the price in dollars they will accept for the assets they want to sell. Every dollar increase in their estimate of the value of the warrants they must give up will reduce the reservation price for the cash component of the sale.
But it’s actually worse than that. The warrant component increases the government’s exposure to adverse selection/winner’s curse/lemons problems. The warrant is a claim on the assets (and liabilities) of the pieces of the firm that aren’t sold to the government outright through the auction or whatever process the government uses to buy assets. Indeed, equity, and especially a levered claim on equity like a warrant, is the most information sensitive component of a firm’s capital structure. The seller of the troubled asset has much better information than the government about the value of the unsold portion of the business (which includes not only the value of the assets and liabilities on the balance sheet, but the values of other pieces of the business, such as the franchise value of a retail banking operation.) Thus, if the winner’s curse is bad for the troubled assets, it will be much worse for the warrants.
The CEO of Bankrupt Bank (Motto: “Insolvent? Who? Us?”) will be glad to give Uncle Sucker all the warrants he wants. You want warrants? We got ‘em!
Moreover, the inclusion of warrants will greatly complicate the purchase process. In an auction for a particular MBS or CDO, the government will receive cash bids, and will demand warrants. The cash components of the bids are readily compared, but the warrant components will be extremely difficult to value. (Maybe the government will have to hire investment bankers to do this. Maybe that’s the real point.) This will add an element of arbitrariness and uncertainty to the determination of the winner.
It also strikes at the root of the theory underlying the buy-the-bad-assets-plan. The bids submitted to, and therefore the prices determined in, the auctions will not reflect only the value of the assets qua assets, but they will reflect the value of the warrants. This last component adds noise to the auction prices (and bids) as measures of the true value of the assets. Since the main defensible rationale of the buy the assets approach is that it will increase liquidity in the market for troubled mortgages by generating information about their value and thereby mitigating informational asymmetries, this additional noise impedes the ability of the asset purchases to achieve the stated objective.
UPDATE: My analysis of the warrant feature of the proposal was based on (a) the statement from Pelosi’s office that it “[g]ives taxpayers an ownership stake and profit-making opportunities with participating companies.” My interpretation of this, based on earlier descriptions of this feature, was that any company that sold assets to the Treasury would have to provide warrants. Tapscott’s summary of the “final bill” gives a very different impression: “Mandatory equity interests in total takeover scenario. Proportional equity interest based on percentage of assets sold if deemed appropriate Secretary.”
This is quite different. It means that, absent discretionary intervention by the Treasury secretary, the government will obtain equity only when it takes over an institution. Thus, the auctions for the assets will be clean, and not present the problem of multiple and potentially difficult to value sources of consideration paid for each asset purchase. This isn’t quite so problematic.
Also, the Tapscott summary states that the final bill has no bankruptcy “cramdown” provision as originally proposed. More good news.
UPDATED UPDATE. Seems Tapscott got it wrong. The actual draft bill requires the Treasury to obtain warrants on non-voting common stock (from companies with listed equity) or senior debt (from those without) from any financial institution selling a troubled asset to the Treasury. There is no language limiting the provision to institutions the government takes over. It is unclear from reading the language whether the warrant is required for each transaction in which an institution participates, or whether the warrant is a one-time pay to play provision. Although the act sets out the purpose the warrant provision is intended to achieve, it establishes no specific guidelines. All in all, the warrant provision is impossibly vague, and will no doubt need the Treasury to specify detailed guidelines before it is operative. These specifics could either neuter the provision, or create the problems suggested in my original post. I’ve only scanned the rest of the bill, and “vague” seems the operative word. There are numerous gaps that will have to be filled going forward.
END UPDATED UPDATE.
Another problematic feature of the proposal is the inclusion of pension funds and politically favored financial institutions (e.g., community banks) to sell assets to the government. As I have argued before, one problem with the purchase the assets plan is that it doesn’t necessarily direct liquidity (cash in exchange for illiquid assets) to the institutions that most need it. The expansion of the set of participants only exacerbates this problem, as many of the new participants (a) likely don’t need the liquidity, or (b) don’t pose a systemic risk even if they do.
The only argument I can see for their inclusion is to increase the competitiveness of the auction process. Any such competitive benefit is conjectural at best. I wonder whether these firms would really want to participate in the auctions, knowing that they are likely relatively uninformed. If they don’t participate, though, they won’t get any money, which would alleviate my first concern.
All in all, my verdict on this iteration, as on all the others, is that it seems to put great faith in an untested theoretical conjecture; that the government purchases of some troubled assets will improve liquidity in the market for all troubled assets, thereby increasing their value and facilitating the recapitalization of the institutions that hold them. $700 billion–or whatever–is a helluva lot to pay for an experiment.
As a result, I fear that the epitaph for the bailout may be the same as the one Jefferson Davis penned for the Confederacy: “Died of a Theory.”