There has been several discussions of the disappearance of Graham stocks in the current market. However, such opportunities still appear if we do some research and go beyond just screening.
What is a Graham stock? It can mean anything to anyone. In this case, I use it as a company that is very close to being a net cash: a company with more in cash and marketable securities than its enterprise value.
The company is Gramercy Capital (GKK), a lender and operator of commercial real estate assets (CRE). Gramercy was formed in 2004 and it was the structured finance arm of SL Green, the New York REIT, with SL Green still holding around 10% of the common shares. It is a real estate investment trust (REIT) and it has two divisions: Gramercy Finance is the owner of the equity portion of three CDOs and Gramercy Realty owns offices and banks branches all across the country.
The contours of the situation are so interesting that Gramercy Capital would have been my pick for the Ira Sohn Conference if it was not for the $1 billion plus market cap limit (damn). I think Michael Price, Seth Klarman and Bill Ackman would have appreciated a cash flow positive Graham stock with a dividend catalyst. Well, can someone relay this write-up to them?
Despite the superficial complexity the thesis is simple: a margin of safety based squarely on ring-fenced liquid assets at the holding company. Let's start with a rhetorical question: would you be willing to buy a $110 million market cap company with the following balance sheet
That simple and safe balance sheet is Gramercy Corporate's balance sheet isolated from its two divisions.
Gramercy Corporate by itself is not a bargain but is cheap and it has two more cash flow positive divisions (Gramercy Finance and Gramercy Realty). However, with so much cash, fantastic CRE opportunities, and a discount to tangible book value I think Gramercy Corporate would still be a proposition that could trigger the curiosity of a value investor. Well, it triggered mine and obviously the great thing was: there were more assets and all non-recourse!
These type of situation raises some typical questions:
- Quality of the jockey: Roger Cozzi is an iStar Financial and Goldman Sachs alumni and SL Green still has a friendly relationship and stock participation. However, the most important point is that the track record is quite good despite appearances. We will review that track record in detail.
- Corporate SG&A: it has a quite high annual run rate of $28 million but most of it covers the operations of the two cash flow positive divisions. If any of these divisions stopped producing, payroll most probably will be cut. In 2006, when they only had Gramercy Finance, SG&A was $16 million and headcount was less than 15 people.
- Cash: with the 10K not yet filled, because of the uncertain Gramercy Realty division future, one important cash event is not included in the most recent financial statements. The selling in December of several New York leasehold agreements to SL Green netting $89 million in unrestricted cash. We will review the effect on total cash later on.
- Real estate valuation: it consists of two properties acquired at distressed prices: 58% discount over the November 2009 stabilized appraised value.
- Securities valuation: acquired in 2010 at a 50% discount to par. They are obscure securities – my type of asset – that are so relevant for Gramercy's upside that we will discuss its potential in detail.
- Preferred equity: the only liability is the $106 million preferred equity that has the benefit that is long term financing. It includes arrears since the dividend was suspended ten quarters ago.
- Size of the opportunity: Big! Double digit unlevered returns still exist, with the whispered securities of two paragraphs ago being the most close and simple.
Well, we will get to tackle in more detail those issues in due time, but first we will review Gramercy's two divisions: Gramercy Finance and Gramercy Realty to get a handle of that upside.
Gramercy Finance is the structured finance division that started Gramercy Capital with the original mission of becoming a CRE lender by taking advantage of non-recourse, long term, non-mark-to-market, low cost CDO bonds. Gramercy owns the most junior tranches – the equity portion – and is the collateral manager of the three CDOs (2005, 2006, 2007 named by the year of their creation).
One thing I want to emphasize. All these CDOs are non-recourse to Corporate even if you do not believe squat about my expectations of recovery and profit. Non-recourse means that the lenders cannot go for more than the CDO assets to recover their principal. If there was still a deficiency, their only legal recourse is to sue the sponsor or manager for violating guarantees or not complying with the CDO covenants; but after close reading of CDO transactions I could not find anything indicating that Gramercy had violated either one of those. Non-recourse CDOs are not only the usual standard practice but also Gramercy heavily emphasizes this point in the 10Ks, corporate communications and the CDOs offering memoranda.
CDOs carry a deserving stigma from the abuses in the residential market and managers chasing yield. However, only a couple of CRE CDO REITs collapsed in the process (ie: Anthracite that invested in lowly rated CMBS and Crystal River that was heavy into residential mortgage backed securities RMBS) but most survived (Capital Trust CT, RAIT Financial RAS, Newcastle Investments NCT) and even recovered (Resource Capital RSO, Northstar Realty NRF, Arbor Realty ABR, Redwood Trust RWT). So, I just prefer to see what is under the hood and reach my own conclusions.
Under that hood you will find that Gramercy's CRE CDOs are some of the most simple of CDOs. CDO 2005 and CDO 2006 have a reasonable probability of recovery and their assets are basically 80% CRE loans (including some mezzanine loans) and 20% CMBS (AAA when they were issued) and they have very little in exotic securities.
How CDOs Work
I do not want to take too much space explaining how CDOs work, especially when others have done it so well. The key is to understand how overcollaterization tests work (OC test or par test) and the implications of the reinvestment period. I am embedding some links but probably the most simple jargon-free explanation is this post:
CDOs are like subsidiaries. There are many variations of CDOs (cashflow, arbitrage, synthetic), I will only talk about the ones GKK has. They start out by the sponsor (GKK) putting up $100 million of equity (RR: it has been more, it has been less), much like how you would start a company. Then they borrow $900 million at L+0.5% (RR: issuing bonds called CDO bonds or notes). They can't just turn around and take the $900 million out of the CDOs and into GKK corporate's pocket. That is why you have restricted cash. Cash either stays in the CDO company structure, or they gets paid according to seniority (which GKK owns the most junior ones).
GKK as manager and equity tranche owner can decide what assets they want to invest in to earn a return. Of course we want the assets (called collateral of the CDO bonds) to generate a higher interest than our liabilities. No matter what GKK does with the $1 billion, the senior tranches will get their interest first, be it all in cash or REOs, CMBS, whole loans . The interest will be paid from the CDO's assets (loans, CMBS, others).
GKK gets to keep any interest income over and above what is owed to the senior tranche holders (RR: what I call excess interest cash flow), as long as the CDOs pass an overcollateralization test. Which is a test to ensure there is enough collateral (assets) in the CDOs to pay off the senior tranche principal. (RR: the equivalent of a leverage ratio, if the test fails senior tranches will get paid their principal first and on down until it cures)
CDOs typically have reinvestment periods of 3-5 years after their birth, during which time principal payments from the loans we make can be rolled into new investments. The alternative is to pay down the senior tranche. This is like you being a landlord, when you receive rent, you first pay your mortgage, then you can decide to keep the rest or make additional payment on your principal. Of course GKK would want to keep more interest-earning assets in the CDO.
When the reinvestment period ends, all principal payment received will be used to pay down debt. As debt goes down, so does the CDO's assets, therefore decreasing the total interest income and distribution to GKK. This is what we mean by "runoff". Eventually, when all loans are paid back, GKK will get their $100 million back as well, barring any discount payoff/default.
All of Gramercy's three CDOs pay fees to Gramercy for being the senior collateral manager ($3.4 million per year) but the real upside comes when a CDO passes its overcollaterization test (OC or par test) and distributes its excess interest cash flow, after paying interest to all the more senior notes, to the equity tranches. If it does not pass the test, it redirects that cash flow to reduce its senior CDO bonds to recover on the OC test and can start cash flowing to corporate again.
How to value a CDO
Some ways to value a CDO would be using a multiple of its net interest income or excess interest cash flow but this could be misleading if the chosen multiple is too high. The problem is that CDOs, after the end of its reinvestment period, are in a loooong runoff where principal payments of the CDO assets go to reduce reduce bonds. This process can last 6 to 8 years slowly reducing liabilities along the way and putting some pressure on cash flows over time. After the bonds are paid back the equity owner keeps the remaining collateral
Therefore, for valuation purposes I prefer the more conservative equity collateral with one caveat: the CDO must have a reasonable chance of recovery of that equity collateral. If not, because the loans will turn bad or the cash flow is not enough to reduce its liabilities, it is better to consider it a zero.
Equity collateral is the equivalent of book value. It represents the residual portion for equityholders if all assets are sold at par and the proceeds are used to pay the liabilities at par. In an environment where the liabilities (aka CDO bonds) are trading at 50-60% of par (keep note of that) while the assets (aka CMBS and loans) are trading close to par, I think it is a conservative estimate.
So let's see this story of turnaround and profit to see if I can persuade the readers. At least CDO 2006 offers a story of profit with no turnaround needed.
CDO 2006 is the crown jewel of Gramercy Finance. As of the first quarter 2011 it had $85 million in equity collateral with a reasonable expectation of payment of interests and an equity recovery at the end of the CDO's life.
Despite taking some big hits – like a very junior Stuveysant mezzanine loan- it has proved resilient and has almost recovered to its original collateral levels. It is well above the OC test trigger of 105.15%, the level where it would have to redirect excess interest cash flow to reduce bonds instead of paying it to the equity portion, and CDO 2006 was consistently above that level during the whole crisis. CDO 2006 has been a cash flowing machine.
It actually improved the interest cash flow received by the equity portion after paying the more senior CDO bonds interest (excess interest cash flow) during the crisis. It also used principal maturities to reinvest in better loans and conservatively increased untapped restricted cash (restricted because it has to be invested in eligible securities for the CDO) in a period where it could instead have closed CRE whole loans at 6%+ interest rates.
To put that in perspective, CDO 2006 generated so much cash flow to Corporate that it more than paid for all Corporate SG&A (of both the Realty and Finance divisions) PLUS the preferred shares dividend (in arrear for the moment). And we still have $196 million in unrestricted cash, $3.5 in annual collateral manager fees, CDO 2005 in road to recovery (with an estimated $35 million in excess interest cash flow when/if it cures) and a Realty division that, while it lasts, it has $60 million in free cash flow (that most probably will be lost).
However, as I mentioned before, for valuation purpose I prefer to use the more conservative $85 million equity collateral despite the very high $37 million excess interest cash flow (43% return over the equity collateral) and the $1.5 million in annual management fees of CDO 2006.
I will leave CDO 2006 at that for the moment but there is a lot more to talk about. It has a couple of distressed assets (Las Vegas Hilton Hotel, Jemal Properties in Washington, AFR Portfolio) but also much was reinvested after the crisis in very high quality loans while it has today in cash $141 million of its $950 million assets. With so much cash and the reinvestment period ending in August, CDO 2006 has another shot for a big improvement of its test by buying more assets at a discount to par/collateral value.
I would love to discuss more these issues if the idea sparks some interest in the comment section or emails.
CDO 2005 is a CDO that is failing its OC test and its reinvestment period ended. Therefore, it is redirecting the excess interest cash flow to pay senior CDO bonds until the OC test improves and cures over time. That is what has been happening the last few quarters. The latest count was 115% in April 2011 so it is getting very close to passing its test.
I do not want to overpromise, but it seems like CDO 2005 is on track to heal which would redirect its excess interest cash flow to Corporate at a current rate of around $35 million per year. The equity collateral is a quite high $104 million. This is a CDO from the times where sponsors of the equity tranche were investment partners not just side gamblers.
There are some potential setbacks (Coyote Land in San Jose, California is one loan that I worry about) but also Gramercy has some options despite the end of the reinvestment period. The most important of the options is to repurchase CDO bonds at a discount with Corporate cash and retire them. This reduces the CDO bonds and improves the OC test. Even more, the discounts are still big (40-50%) so it is very good use of unrestricted cash.
I mention some obscure securities bought in 2010 when we reviewed Gramercy Corporate balance sheet at the very beginning. Those were repurchases of CDO bonds and that is where Gramercy is focusing its use of unrestricted cash. Here is a good explanation of the incentives behind these repurchases.
After reviewing the previous CDOs one gets the impression that the investment underwriting for the CDOs was not perfect but that Gramercy has a reasonable chance of recovery from both CDO 2005 and CDO 2006 Gramercy. Even after taking some heavily publicized hits (loan loses from Stuveysant, Cupertino Mall, Suncal/Lehman) the OC tests have not collapsed.
However, I think it is fair to ask how CDOs die and never recover. Or as an old sage would say: I want to know where I am going to die so I don't go there. CDO 2007 is a good example of a busted CDO.
The problem with CDO 2007 was that it never got off the ground.
- Bought CMBSs too early: When the CDO was created, management moved fast to buy a large percentage of its portfolio into what at the time were liquid AAA CMBS trading at some discounts to par. Gramercy was too early in the crisis, with CMBS delinquency rates only recently stabilizing and above 10%.
- Thin Equity Cushion: besides the bad investment timing, Gramercy did not put too much equity in this CDO. It was more of a "heads I win, tails I do not lose much" investment.
- Fixed Rate Structure: And finally, CDO 2007 was structured as a fixed-rate instrument with hedges for the few variable rate assets. So it could not benefit from low interest rates like CDO 2005 and 2006 did (that are "libor plus" instruments) by using libor floors and other tools that can protect revenues while interest expenses kept going down.
The compounding effect was that CDO 2007 from the beginning did not generate enough cash to even pay the interest on the CDO bonds. It also invested in Stuveysant in a very junior mezzanine that got clobbered. So do not even ask if the redirected excess interest cash flow will in time reduce its liabilities and cure its OC test. It cannot. There is no excess cash flow. It is a dead parrot.
So the key point for a CDO aspiring to have a second life after failing an OC test, is cash flow, cash flow, cash flow to reduce fast the CDO bonds and cure the OC test. In consequence, CDO 2007 does not have an opportunity for recovery but CDO 2005 and CDO 2006 do.
There is only one redeeming factor for CDO 2007: because it was created in 2007, late in the last credit boom, it is still in its reinvestment period and has very light covenants on the type of instruments and concentrations it can invest. Because it is failing its OC excess interest cash flow is redirected to reduce CDO bonds but since it is still in its reinvestment period it can still reinvest principal maturities. Therefore, if Gramercy needs funding for some creative financial engineering, CDO 2007 can be at the center of it.
For example, CDO 2007 can buy CDO bonds from the other CDOs (those obscure securities again) that are trading at a discount and providing such nice yield. Another example is if Corporate wants to negotiate keeping some Gramercy Realty properties in exchange for the deed-in-lieu foreclosure. CDO 2007 could provide a mortgage for those properties transforming its CDO restricted cash into Corporate unrestricted cash. The possibilities are endless with a good jockey managing those proceeds.
Just a brief accounting note. We have discussed how CDOs 2005 and 2006 have positive net interest margin, with CDO 2006 cash flowing to corporate, and that the CDOs are non recourse to corporate assets. Surprisingly, given that the absolute worst case scenario should still leave some recovery for Gramercy, Gramercy Finance has a large NEGATIVE book value.
To understand more this distortion, here is some data from the most recent 10Q that compares the value of the assets and liabilities of the CDOs at carrying value versus fair value (market value).
The reason for the large CDO fair value discount has been the general run on "toxic" securities and the difficulty on deciding which were fine. That run was even more extreme for CDO bonds because of their opacity.
Granted, Gramercy is not going to be able to buy all its CDO bonds at 50% discounts: it does not have that much liquidity and just the announcement of their intent would cause a price increase. Though at the same time the negative carrying value is not realistic either: the securities are cash flowing and are non-recourse.
If you want to find more about GAAP impact on CDO accounting, a good place to start is REIT Wrecks. Next, we are going to review the mystery: the uncertain outcome of the valuable Gramercy Realty division
Disclosure: Long GKK