Conducting Reverse Auctions for the US Treasury
I regularly read “A Dash of Insight,” and greatly appreciate the commentary of Dr. Jeff Miller. What I write here is an effort to encourage what he wrote in this piece advising President-Elect Obama. (I would have my own advice for the President, but there are so many vying for his ear now, that I sigh and say “Let the poor man get on with it. He will be imprisoned for the next four years, and likely find less capability of doing what he wants than he imagined.”)
How would one implement what Dr. Jeff suggests? As a bond manager, I was pretty good at price discovery. I would convene a committee of large holders of the illiquid instruments and ask them what are the largest classes of homogeneous structured securities that no longer have markets now. Once they agree on the classes (probably the AAA portions of the senior-sub structured ABS, RMBS and CMBS deals), the agent for the Treasury picks a subset of the largest deals, and announces how much of each security (say 10% of each tranche) they will offer to buy.
Market participants are then invited to submit binding offers to sell any amount of the securities up to the maximum. The Treasury’s agent could require a minimum amount of bids in order for an auction to be valid (say 2-3x the purchase amount).
One tweak I would put in would be to award the bonds to the winning bidders at the price offered by the bidder with the highest bid not receiving bonds. I used this successfully for years in bond auctions, and though it makes the trader shake his head initially, when I would say, “I’m offering protection against regret in advance, besides, I want aggressive bids.” they would say, “Okay, I get it.”
After the auctions, there would be benchmark prices, yields, and spreads for a wide number of securities, and then the modelers would apply those prices to the mezzanine and maybe the subordinate tranches, which are too small to hold auctions for.
Similar securities might find trading levels as well, but if not, the Treasury could run another set of auctions, and repeat as necessary. Given the most of the securities auctioned are AAA, at worst, the Fed might have an interest in the short-to-intermediate AAA paper.
If the Treasury followed a procedure like this, it could unjam the securitized fixed income markets, and do so at prices where the taxpayer bears modest losses at best. I am not as optimistic as Bill Gross or Warren Buffett on this matter. The point of the auction is to get the sellers to compete against each other, not compete with the government’s agent.
Now, price discovery is a two-edged sword. FInding the market clearing price will make the markets start moving again, but it also might prove that some financial institutions are inverted (negative net worth), if not insolvent (can’t get enough cash to pay all immediate claims). If we are willing to stomach the possible insolvencies that this will reveal, then I am game for Dr. Jeff’s proposal.
And, maybe this will show the need for RTC II, successor to the old Resolution Trust Company. Bad financial institutions need to be conserved/liquidated, so that leverage can be reduced in the financial system of the US.
So, let something like this be tried, but be ready for adverse consequences if the pricing turns out to be worse than anticipated.
November 6th, 200810:40 pm at
Thanks David, both for citing this and for your thoughtful analysis. I plan to write more about specific mechanisms, and your ideas are certainly most helpful.
Jeff