A few numbers to consider:
Corus loan & REO assets sold by FDIC: $4.5B
Corus NPL + REO (est. from 6/30/09 10Q): $3.2B (71% of total)
Price Paid by Starwood consortium: $2.7B (60% of book value)
-- FDIC providing 50% debt financing at very attractive rates (below market).
IStar loan portfolio + REO: $12.0B
IStar NPL + REO: $5.0B (42% of total)
IStar NPL + REO + Watch: $6.2B (52% of total)
If you apply the 60% Corus pricing against the total $12B portfolio, the writedown totals $4.8B vs. SFI's current loss reserves of $1.5B. One might conclude that SFI is underreserved by $3.3B.
However SFI's NPL + REO total, though a high 42% of total loans, is FAR less than the ridiculous 71% posted by Corus. Without having much detailed data for either SFI or Corus loan composition, one can conclude that SFI's portfolio is in materially better shape than Corus' portfolio. So apply the 60% pricing against the $5B of NPL & REO assets and you get a writedown of $2B vs. SFI's $1.5B of reserves today. That's a gap, but not a humongous one (considering we really have no details to dive deeper) and close enough for government work. The entire gap (and maybe more) can be explained by the fact that the FDIC is a distressed seller and SFI is taking steps not to be a distressed seller.
I am surprised at the 60% pricing paid by Starwood (even including the below market financing) for a portfolio of primarily condo loans that are ~70% non-performing and REO. Maybe when/if more details on the loans come out I will change my view, but the FDIC didn't do bad on this sale and this comp is a bullish sign for SFI, IMO.
sry, posted before I was finished.
iStar's portfolio is self liquidating, comprised of (nominally) short maturity loans receivable. It's not as if "liquidating" entailed something extraordinary. The proceeds of liquidation pay off debt. Buying debt at a discount decreases the amount needed, so the fact that "bond purchase needs cash and you need to further liquidate the portfolio" ends up in a circular replay, since what is being liquidated is being used for what it's would be used for if borrowers were paying on time.
What's tough about the math? iStar has positive operating cash flow and doesn't need to sell assets to pay the electric bill. Once the funding obligations are finished it will be simpler.
I have indicated in repeated posts today that the common book value is $15/sh. as of 6/30/09. If you want to adjust it by adding back in the portion of the bond gain that was not taken currently earlier this year ($200MM or $2/sh.), then that can get you to $17/sh. Take out $500MM in additional reserves on the bad assets and $200MM in reserves on the performing assets, that gets you down to $10/sh.
Many SFI watchers make the mistake of taking the total equity reported by SFI ($2B) and dividing by total shares to come up with an erroneous book value for the common equity. You are correct, preferred equity needs to be netted out, but I do not make that mistake myself. Take a look at my earlier posts and you will see $15/sh.
I don't disagree with the difficult math comment, SFI has to work hard to manage sources and uses in a tough environment. So far they've done a good job.
If you want a balance sheet play, you should look for somewhere else. I recently got some QXM. It's the 3rd mobile phone manufacturer in China in terms of market share (although much smaller than the first two manufacturers). It has extremely low debt while its market cap is below its net cash. It has positive cash flow for every quarter. The only and the biggest uncertainty is its financial statements which never file on time.
" So apply the 60% pricing against the $5B of NPL & REO assets and you get a writedown of $2B " I don't follow your logic here. Where is that 60 cents of a dollar for NPL & REO coming from? No write down needed for performing loan???
Also, you have to know that when investor buys Corus portfolio, the existing debt is gone. The new debt is the fdic one which is guaranteed to be lower than the purchase price.
So, let's try to calculate SFI's loan portfolio by using the good loan bad loan strategy suggested by Quad.
Assumption: good loan worth 0.8 of a dollar (this comes from SFI loan portfolio sale, I assume it is selling the performing one).
Corus portfolio has good loan of 1.3 bil, so the price is 1.04 b. That means the investor paid 1.66 b (2.7-1.04) for the bad loan. Bad loan worth: 1.66/3.2 = 51.87% of the original value.
SFI has 6.2 bil bad loans (I use the more conservative number, an asset is on watch list obviously because it's kinda bad if not totally bad.)
The portfolio worth: 6.2 * (0.5187) + 5.8 * (0.8) = 7.86 billion. The CTL worth about 3 bil, same as the book value (correct me if i am wrong on this). So total 10.86 bil. To purchase this asset, you get a "bonus" of 12 billion liability and ongoing corporate overhead. That means SFI pretty much has no value on book, and the only reason of buying this is, the debt can be paid back at a discount.
Hmtsomo, you included loans and CTL's in your SFI asset total, but you left out other assets (Oak Hill, etc.)equal to $392MM and cash of $417MM (at 6/30/09). That's an additional $809MM.
Each investor should use their own estimate of asset value when looking at SFI. I think the 80% estimate you applied to performing loans is too low.
Another way I look at the SFI reserves in light of the Corus deal: SFI has $3B of condo construction/condo inventory exposure out of the $4.6B of NPL's. That portfolio is a reasonable match with the Corus portfolio. Assuming Starwood paid 50% for the NPL/REO assets in Corus, apply the 50% to the $3B and you get $1.5B, SFI's current total reserve. So, they may be underreserved by $500MM, which would be a 31% reserve against the remaining $1.6B of NPL's. I personally wouldn't apply an additional 50% writedown against SFI's $383MM of REO, I think SFI gets it reasonably close to the mark when it values REO.
That leaves the performing loans and what kind of reserve should be used against those assets. Like I said, 20% reserve is too high unless you were going to try and sell those assets today and SFI is not in that position.
All together, if they need another $500MM of reserves that's $5/sh., so $15/sh. current BV - $5/sh. add'l reserve = adjusted book value of $10/sh. Maybe take that down a buck or two for reserves against performing loans, but you should also add in a buck or two for the present value of the lower interest rate in the coming years from the bond exchange offer earlier this year (they could only take part of the gain in the currently period). So that gets me to $10/sh. adjusted BV and the stock trades at ~$3/sh. Worth holding for now to see what happens in the coming few quarters.
iStar has its issues but what is your valuation purporting to do?
If you impute .8 value to "good" loans all US banks would be insolvent. I'm not sure what the theory you are advancing that suggests that good loans in an operating business should be discounted. iStar is not in liquidation.
In the Q2 10K iStar itemized the fair value of its debt, on page 42. Their table indicated over $4 Billion of unrealized fair value in the debt. I'm not sure what it really is, but the illustration suggests about a 35% discount.
SFI has positive cash flow and has for the last several years. Your denigration of corporate overhead as dead weight is dead wrong.
"That means SFI pretty much has no value on book, and the only reason of buying this is, the debt can be paid back at a discount. "
This is an assertion of yours unsupported by your selective analysis. Your asset valuation metrics are subjective and arbitrary and your liability valuation is absent. Your premise of a liquidation value is irrelevant outside of imminent or likely forced liquidation.
It is clear that SFI doesn't warm your heart, but the conclusions you reach appear to have existed prior to the ersatz analysis.
Ops, I forgot the preferred stock for a few hundred million i think. It looks like SFI needs to get 20% discount on its whole debt in order for the common shareholders to get something. If the CTL worth only about 80% on the book, it needs 25% discount.
The difference between iStar and Corus is that iStar has/had their own lending platform, that was hyper active and hyper involved in every single loan that was made. Corus was simply a small regional bank, that through a number of acquisitions grew larger, and yet never lost its community bank background. But given how far afield they were lending, and a totally non-integrated lending platform, they were bound to make huge missteps as they went far out of bounds on their capabilities to understand every single loan they made. iStar has always carried a great deal of granularity to each loan.
The lower interest rate given to Starwood is an attempt trying to increase the amount paid for the assets.
Government is trying to figure out how to get these banks to give up their toxic assets and the only way is to phony up a lower than market financing rate in return for higher payment for the portfolio.
There is so much phony crapola going on right now its impossible to take away anything from Corus and superimpose on SFI.
After all, Starwood has three entities with a ton of cash from the one recent IPO and the private funds raised that would otherwise earn 1%.
I do know from a CitiBank report on STWD that the desired dividend yield for investors is intended to be close to 11%.
Assuming investors agree that SFI has enough reserves and that the asset book values are reasonable marks for current value, then investors are currently valuing the SFI platform at negative $1.2B. I will magnanimously offer shareholders negative $1.0B and top that bid. Shareholders can cut me a check for the negative $1.0B.
If you take out the prospective earnings power, does that mean the platform could be worth negative $1.4B?