One of the earliest pieces at this blog was What is Liquidity?, followed by What is Liquidity? (Part II). I’ve written a bunch of pieces on liquidity (after doing a Google search and being surprised at the result), largely because people, even sophisticated investors and unsophisticated politicians and regulators misunderstand it. Let’s start with one very simple premise:
Many markets are not supposed to be liquid.
- Small markets are illiquid because they are small. Big sophisticated players can’t play there without overwhelming the market, making volatility high.
- Securitization takes illiquid small loans and transforms them into a bigger security(if it were left as a passthrough), which then gets tranched into smaller illiquid securities which are more difficult to analyze. Any analysis begins with analyzing the underlying loan collateral, and then the risks of cashflow timing and default. There is an investment of time and effort that must go into each analysis of each unique security, and is it worth it when the available amount to invest in is small?
- Buy-and-hold investors dominate some markets, so the amount available for sale is a small portion of the total outstanding.
- Some assets are opaque, where the entity is private, and does not publish regular financial statements. An example would be lending to a subsidiary of a corporation without a guarantee from the parent company. They would never let and important subsidiary go under, right?
- The value of other assets can be contingent on lawsuits or other exogenous events such as natural disasters and credit defaults. As the degree of uncertainty about the present value of free cash flows rises, the liquidity of the security falls.
When is a securitization most liquid? On day one. Big firms do their due diligence, and put in orders for the various tranches, and then they receive their security allocations. For most of the small tranches, that’s the last time they trade. They are buy-and-hold securities by design, meant to be held by institutions that have the balance sheet capacity to buy-and-hold.
When are most securitizations issued? During the boom phase of the market. During that time, liquidity is ample, and many financial firms believe that the ability to buy-and-hold is large. Thus thin slices of a securitization get gobbled down during boom times.
As an aside, I remember talking to a lady at a CMBS conference in 2000 who was the CMBS manager for Principal Financial. She commented that they always bought as much of the AA, single-A and BBB tranches that they could when they liked the deals, because the yield over the AAA tranches was “free yield.” Losses would never be that great. Privately, I asked her how the securitizations would fare if we had another era like 1989-92 in the commercial property markets. She said that the market was too rational to have that happen again. I kept buying AAA securities; I could not see the reason for giving up liquidity and safety for 10, 20, or 40 basis points, respectively.
Typically, only the big AAA tranches have any liquidity. Small slices of securitizations (whether credit-sensitive or not) trade by appointment even in the boom times. In the bust times, they are not only not liquid, they are permafrost. In boom times, who wants to waste analytical time on an old deal when there are a lot of new deals coming to market with a lot more information and transparency?
So, how do managers keep track of these securities as they age? Typically, they don’t track them individually. There are pricing grids or formulas constructed by the investment banks, and other third-party pricing services. During the boom phase, tight spread relationships show good prices, and an illusion of liquidity. Liquidity follows quality in the long run, but in the short run, the willingness of investors to take additional credit risk supports the prices calculated by the formulas. The formulas price the market as a whole.
But what of the bust phase, where time horizons are trimmed, balance sheets are mismatched, and there is considerable uncertainty over the timing and likelihood of cash flows? All of a sudden those pricing grids and formulas seem wrong. They have to be based on transactional data. There are few new deals, and few trades in the secondary market. Those trades dominate pricing, and are they too high, too low, or just right? Most people think the trades are too low, because they are driven by parties needing liquidity or tax losses.
Then the assets get marked too low? Well, not necessarily. SFAS 157 is more flexible than most give it credit for, if the auditors don’t become “last trade” Nazis, or if managements don’t give into them. More often than not, financial firms with a bunch of illiquid level 3 assets act as if they eating elephants. How do you eat an elephant? One bite at a time. They write it down to 80, because that’s what they can afford to do. The model provides the backing and filling. Next year they plan on writing it down to 60, and hopefully it doesn’t become an obvious default before then. Of course, this is all subject to limits on income, and needed writedowns on other assets. I have seen this firsthand with a number of banks.
So, relative to where the banks or other financials have them marked, the market clearing price may be significantly below where they are currently marked, even though that market clearing price might be above what the pricing formulas suggest.
The US Treasury Proposal
The basics of the recent US Treasury proposal is this:
- Banks and other financial institutions gather up loans and bonds that they want to sell.
- Qualified bidders receive information on and bid for these assets.
- High bid wins, subject to the price being high enough for the seller.
- The government lends anywhere from 50-84% of the purchase price, depending on the quality and class of assets purchased. (I am assuming that 1:1 leverage is the minimum. 6:1 leverage is definitely the maximum.) The assets collateralize the debt.
- The FDIC backs the debt issued to acquire the assets, there is a maximum 10 year term, extendable at the option of the Treasury.
- The US Treasury and the winning private investor put in equal amounts, 7-25% each, to complete the funding through equity.
- The assets are managed by the buyers, who can sell as they wish.
- If the deal goes well, the winning private investors receive cash flows in excess of their financing costs, and/or sell the asset for a higher price. The government wins along with the private investor, and maybe a bit more, if the warrants (ill-defined at present) kick in.
- If the deal goes badly, the winning private investors receive cash flows in lower than their financing costs, and/or sell the asset for a lower price. The government may lose more than the private investor if the assets are not adequate to pay off the debt.
I suspect that once we get a TLGP [Treasury Liquidity Guaranty Program] yield curve extending past 3 years, that spreads on the TLGP debt will exceed 1% over Treasuries on the long end. Why? The spreads are in the 50-150 basis point region now for TLGP borrowers at 3 years, and if it were regarded to be as solid as the US Treasury, the spread would just be a small one for illiquidity. (Note: the guarantee is “full faith and credit” of the US Government, but it is not widely trusted. Personally, I would hold TLGP debt in lieu of short Treasuries and Agencies — if one doesn’t trust the TLGP guarantee, one shouldn’t trust a Treasury note — the guarantees are the same.)
One thing I am unclear on with respect to the financing on asset disposition: does the TLGP bondholder get his money back then and there when an asset is sold? If so, the cashflow uncertainty will push the TLGP spread over Treasuries higher.
Thinking About it as an Asset Manager
There are a number of things to consider:
- Sweet financing rates — 1-2% over Treasuries. Maybe a little higher with the TLGP fees to pay. Not bad.
- Auction? Does the winner suffer the winner’s curse? Some might not play if there are too many bidders — the odds of being wrong go up with the number of bidders.
- What sorts of assets will be auctioned? [Originally rated AAA Residential and Commercial MBS] How good are the models there versus competitors? Where have the models failed in the past?
- There will certainly be positive carry (interest margins) on these transactions initially, but what will eventual losses be?
The asset managers would have to consider that they are a new buyer in what is a thin market. The leverage that the FDIC will provide will have a tendency to make some of the bidders overpay, because they will factor some of the positive carry into the bid price.
I personally have seen this in other thin market situations. Thin markets take patience and delicate handling; I stick to my levels and wait for the market to see it my way. I give one broker the trade, and let him beat the bushes. If nothing comes, nothing comes.
But when a new buyer comes into a thin market waving money, pricing terms change dramatically after a few trades get done. He can only pick off a few ignorant owners initially, and then the rest raise their prices, because the new buyer is there. He then becomes a part of the market ecosystem, with a position that is hard to liquidate in any short order.
Thinking About it as a Bank
More to consider:
- What to sell?
- What is marked lower than what the bank thinks the market is, or at least not much higher?
- Where does the bank know more about a given set of assets than any bidder, but looks innocuous enough to be presumed to be a generic risk?
- Loss tolerances — where to set reservation prices?
- Does participating in the program amount to an admission of weakness? What happens to the stock price?
Management might conclude that they are better off holding on, and just keep eating tasty elephant. Price discovery from the auctions might force them to write up or down securities, subject to the defense that prices from the auctions are one-off, and not realistic relative to the long term value. Also, there is option value in holding on to the assets; the bank management might as well play for time, realizing that the worst they can be is insolvent. Better to delay and keep the paychecks coming in.
Thinking about it as the Government and as Taxpayers
Still more to consider:
- Will the action process lead to overpriced assets, and we take losses? Still, the banks will be better off.
- Will any significant amount of assets be offered, or will this be another dud program? Quite possibly a dud.
- Will the program expand to take down rasty crud like CDOs, or lower rated RMBSand CMBS? Possibly, and the banks might look more kindly on that idea.
- Will the taxpayers be happy if some asset managers make a lot of money? Probably, because then the government and taxpayers win.
This program is not a magic bullet. There is no guarantee that assets will be offered, or that bids for illiquid assets will be good guides to price discovery. There is no guarantee that investors and the government might not get hosed. Personally, I don’t think the banks will offer many assets, so the program could be a dud. But this has some chance of success in my opinion, and so is worth a try. If they follow my advice from my article Conducting Reverse Auctions for the US Treasury, I think the odds of success would go up, but this is one murky situation where anything could happen. Just don’t the markets to magically reliquefy because a new well-heeled buyer shows up.