New VIC writeup: https://valueinvestorsclub.com/idea/SERITAGE_GROWTH_PROPERTIES/142257
I think this is the 4th one? (3 long, 1 short)
I wrote this write up and feel somewhat compelled to inform people here that this write up received some justified criticism. Non-members of VIC can't see my response so I'll copy both the criticism and response below:
CRITICISM:
We believe you made a simple math error in your Individual Project Valuations that has a significant impact on your calculated fair value of Seritage. As a quick sanity check, the Redmond 2.5 acres sale for $16mm implies the remaining 12.5 acres are worth $80mm. To say this $80mm can be turned into $501mm on an undiscounted basis is aggressive.
You have confused Revenue with NOI. If you assume a 65% NOI margin on Residential, Retail, and Office, the gross value decreases by $155mm, $53mm, and $67mm, respectively. The undiscounted net value decreases from $501mm to $226mm. Making this math correction on just the Redmond property decreases the value of Seritage by $5 per share (or $3.50 with your 30% discount).
Beyond the simple math errors, we identify many further questionable assumptions. For instance, Rent PSF looks very high. You can find modern, new housing on Apartments.com in Redmond for $33 PSF on average as opposed to the $38.50 PSF assumed. We believe Class A Office Space outside of Seattle in Bellevue is renting in the low to mid $30 PSF as compared to the $42 PSF assumed. In addition, your valuation work appears to have excluded financing costs and overhead costs.
I think correcting the math errors would dramatically decrease the fair value for Seritage. Were you to use more realistic assumptions around rent PSF as well, that would further compress the fair value. Based on just the math errors identified above, I believe you have overstated the value of Seritage’s redevelopment opportunities by about $1 billion or $18 per share before discounting.
RESPONSE:
Wow what a glaring and embarrassing error. You're right, I was using revenue multiples and then changed them to cap rates at the last minute and didn't change my input of revenues to be inputs of NOI. Doing so substantially reduces the valuation. I still think the investment might work out because of conservatism built in elsewhere, but it does blow up the approach. After reducing Residential, Retail, and Office revenue by 30%, 30%, and 22.5% respectively to get NOI, and changing the overly conservative 30% PV discount to 20% (a bit more than 3 years at 7%) I get a reduced valuation of $39.80. Again, this approach leaves out what I believe is the substantial upside from further redevelopment of the properties, but on the other hand still doesn't lower the rent inputs or capitalize corporate expense.
A couple of relatively minor pushbacks.
Substantial though it may be, I believe you are over-estimating the magnitude of the "math error". The sum total of the three individual projects in my original valuation was $17.97. Only if these properties were actually worth nothing could they reduce the value estimate by $18 per share. If I were to value these properties how I valued the other properties with no formal projects (the lower of Hilco Dark and Cushman & Wakefield estimates) the result would be $73MM ($1.32 a share), but of course that is too low because the Redmond land alone just got valued at $96MM.
To be fair to me, I wasn't saying that $80MM of land could be turned into $501MM. I was saying $96MM of land and $377MM of construction capital and multiple years of planning and 5 years of construction could be turned into $878MM eventually. Maybe I ought to be locked up but that passed my sanity test.
Yes, I used a high estimate of rent for Redmond, but I also used a high construction cost estimate of $350, because that was the estimate given to me by someone who looked at the actual project specifics, instead of being a general area estimate. It was higher than the other Redmond residential construction estimates I received of $180 (from a 2015 study of the area average), and $251.64 and $291.50 received from other professionals. I figured brand new and mixed use and next to a park would go for a premium. It would seem to me overly punative to reduce the high rent estimate and leave the high residential cost estimate the same. Not sure if you would have any input on a residential cost estimate.
The same story goes for office. We received office rental estimates of $19.64, $25, $30.91, and $42. We received office cost estimates of $165.09, $207.88, and $300. I chose the top estimates for each based on the above logic.
As for not capitalizing and including the corporate expense, you are again correct. I wonder though if I were to take the "General and Administrative" line item from the income statement and capitalize it if I would not be double-counting some of the "soft" project costs which are accounted for in the project cost estimates. I'm all ears on suggestions on how to approach this.
