Can an auction find the longterm price for financial junk?
The principle behind it is painful but laudable: to reward the banks for their reckless risk taking in order to avoid crashing the real economy, by spending $700bn on financial junk. The issue is can it work? I've been gaming it out in my head and this is how far I have got.
First let's define the dilemma: there is a large amount of bad debt in the system - let's say $2 trillion dollars' worth (Nouriel Roubini's estimate on Newsnight 16 Sept 2008).
If they had to sell a portion of that debt now? Well suppose Bank A says, I've got $2m of bad mortgages I want to get rid of, what will Bank Z pay? Right now Bank Z has no money. It has switched its cellphone to "reject all calls" and is watching a DVD box set of The Wire with a case of Sierra Nevada in its luxury loft apartment with the blinds drawn. Banks B to Y likewise. Nobody is buying....
So the value of Bank A's $2m is, on a "mark to market" basis, is anything from zero to the notional amount the Bank A decides if it has to do its accounts today (the higher the figure, the louder the laughter of impartial market commentators - Ben Bernanke acknowledged this yesterday: "nobody trusts the banks' own valuations").
This price, between zero and "duh" is what Bernanke calls "the firesale price". And he pleads, not unreasonably, that if the banking sector is forced to sell at firesale prices, the whole system will become mired in losses and seize up even further; and that firesale prices, like house repossessions, cause an unnecessary downward spiral.
So instead there is the "hold-to-maturity" price. If I have a 20 year mortgage for $100k, then the "hold to maturity" value of that $100k mortgage is the what it will be worth to Bank A over the next 20 years. That has to be calculated for each mortgage: it's not just the 100k plus the compound interest. The risk of default has to be factored in as well, and the loan to value percentage. Today "hold to maturity" is going to be more than any "mark to market" price estimated by either the bank or the government. But what is the price?
Bernanke's suggestion is to design an auction to find out. The market, according to theory, should determine its true "hold-to-maturity" price. It would be a reverse auction, with Bank A competing with Banks B to Z to unload its bad debts. But, as Forbes reports...
"neither Bernanke nor Treasury Secretary Henry Paulson offered any details of exactly how the auction process would work. Paulson told the committee that it is not possible for he or the senators to design the process, and said it would be left to 'experts'."
OK here is a quote from a dummies' guide to tactics in reverse auctions, aimed at companies participating in procurement bids. Here are some possible tactics in a reverse auction according to the guide:
- To come in second (or third) while keeping the price high,
- To come in second while driving the price down to unprofitable levels for the winner,
- To bid down to a certain price and stop, regardless of winning position and potential profitability,
- To keep the price high even if you would lose the auction at a profitable price,
- To abstain from participation in the auction.
I will summarize this. In a reverse auction many of your options include shafting your rivals by losing the bid. Or shaping the next auction while losing the bid. Not included, but very relevant, is collusion to mess up the entire auction. There is a whole subset of game theory devoted to who wins in reverse auctions and, I am sorry to say, it is not always the buyer. Experience of online reverse auctions suggests any saving to the buyer (ie to Bernanke/Paulson) is usually only 10-20% of the market price.
So problem number one is that you have to assume the banks do not game the auction. May I suggest that, on current form, that is not a 100% safe assumption.
Problem number two is bigger. Is the Fed/Treasury trying to get the best possible deal for the taxpayer in these auctions or not? Is it trying to do what GE or Ford does with its suppliers, by getting them to compete so as to drive their prices down? If so, it may be my stupid brain, trained as it is only in Baroque counterpoint and media law, but it seems to defeat the object.
If the Fed/Treasury are trying to get the best possible deal for the taxpayer, then they should buy at firesale prices. They have already said this will damage the economy systemically, and that the banking industry needs to be able to be rewarded in the bailout process.
So I contend that they will have to design the reverse auction to prevent it achieving the true "hold-to-maturity" value.
Here's why: Suppose now I replace Bank A with a recently acquired part of Lehman Brothers (call it Ex-Lehman), and Bank Z with AIG.
Last week Lehman and AIG were valuing their distressed Alt-A debt at 39 cents on the dollar and 67 cents on the dollar respectively. Suppose they log on to the reverse auction site and say, "hey Bernanke we will sell Paul Mason's $100k mortgage at 39c" (Ex-Lehman) "or 67c" (AIG). Bernanke's gavel hits the wood so fast it makes even the sleepiest Senator wake up:
"Bang! Sold to Ex-Lehman. Here is $39,000 dollars, gimme the documents on that mansion - by the way, AIG, you wanna sell at 39? Going, going?"
For now only AIG has lost, but they are feeling sore. But...
Instead Bernanke waits, gavel poised, using his "For Dummies" guide to reverse auctioneering, until one player undercuts the other. Let's assume the book value of the mortgage really is (39/100) $39,000. With reverse auctioneering, Bernanke should be able to strip 20% off that, says the theory: so now both Ex-Lehman and AIG have to dump my mortgage at, say $35,000.
"Boo, hoo", says one or both. "We're not playing anymore. You're hurting us too hard here. We are gonna stop lending to each other again like we did in August 07 and September 08".
At this point, the rationale falls apart. The whole rationale for rewarding banks with an above-market price is to keep the system working. And it's a laudable rationale. But, call me stoopid, I cannot see how a reverse auction works for this unless it is specifically designed not to depress the price of the toxic debt.
Therefore I propose a reverse auction cannot find the true hold-to-maturity value of the morgage, no matter which expert designs it, because of this moral hazard injected into the design of the auction.
Basically, as Henry Blodget points out (in a post which also contains some realtime commentary on the Paulson/Bernanke testimony), the whole rationale for not forcing a firesale is that they have to involve the banks in sorting out the system. It is a market solution to a market-created problem and the "reverse auction" must have seemed like an elegant solution.
The emerging alternative, in the Dodd Plan and from numerous commentators, both on the re-regulationist and freemarket side of the debate, is for the government to enforce a firesale, enforce a massive loss, then force the banks to raise new capital. This is guaranteed to deliver returns to the taxpayer at the end of the process, just as did RTC and HOLC designed in the 80s and 30s respectively. I am not advocating either course of action. I am simply trying to get my head around them, Of course the problem with the latter course is that one or more banks goes bust and the USA ends up, like with the Nordic banking crisis, owning much of the sector.
If I am wrong, kick my head in on the Comments. If you are a reverse auction expert, feel free to submit your design to me here and I will see if the BBC competition guidelines allow me to offer some kind of Blue Peter Badge.