No. of Recommendations: 0
Hi Will,

Great questions and thanks for taking the time to think through the model and what's happening.

- As I understand, RV doesn't include income taxes, selling&marketing, general&admin and CapEx. Why don't you adjust RV downward to account for those items?

- RV also doesn't include the revenue generated by the purchase&installation of the solarpower system for customers who didn't choose the $0 upfront-cost-though-higher-fixed-lease-payments. Or does it? That extra revenue should increase the fair value.

I didn't include those for one reason. I used a market multiple valuation at the end based solely on RV and compared it to the current market multiple assigned to current RV. Going that route I necessarily had to not include additional expenses like SG&A and income taxes. (CapEx wouldn't be included because RV is not a cash flow, this isn't a DCF model.)

Same for additional revenue from selling the installations outright to the homeowner (your second question).

Basically the model says "Here's what the market is saying the value is based on today's RV. Here's an estimate of what the market will say in 2017 based on a projected 2017 value for RV." All the extra expenses and revenue sources are baked in to the market multiple.

- Would it you deem it too preposterous to try to model the gross expenses?

No, just very, very difficult. I try to keep my models pretty simple because I know how much any assumption can affect the outcome. The fewer assumptions I make, the less chance there is for error (and there's already plenty of chance).

Here's an illustration of that difficulty:

Based on the $1409M value of ENPR, if you assume that they are evenly distributed throughout the next 20 years as an annuity

That's not a valid assumption. A friend of mine was kind enough to give me the exact contract he had just recently signed with SolarCity for the installation on his house. From that it is plain that the output of the panels goes down by some amount each year which means the guaranteed delivered power output goes down. At the same time, his cost per kWh increases by a certain percentage each year.

I know what the increase in price is because it's stated very clearly on the contract and I can calculate an average decline in panel output efficiency. However, that's just one contract. I have no idea if the same changes would apply to all contracts and extrapolating from a single contract to thousands over the course of 4+ years is a very big assumption. How will the technology of the panels improve over time? What if SolarCity applies different price increases for different customers based on how much each customer's regular utility bill has increased in the past? And so on. That's the revenue side.

On the cost side, there are historical prices on panel costs, but I don't know how those will play out in the future nor what the company is actually paying today. Similarly with labor costs, maintenance costs, and so on. What if the company finds that it's maintenance costs come in higher than currently projected, based on more years of data over more system installations? I can't reliably model that.

The biggest lever in the model I built is the growth rate of MW installed. I think I was pretty harsh here, cutting it in half for each of four consecutive years. This covers a multitude of sins, to steal a phrase. Among these are changing gross margins (e.g. increases in panel costs that cannot be passed along to customers), failure to increase revenue by selling systems outright, pushback competition from utilities that cuts into the business, and so on. I have no idea how each of those might affect my numbers, so I lump them all into a single spot and try to be conservative there.

In the follow up post, Travis had pointed out where I might have been too optimistic by letting nominal $$ / MW grow, so I cut further, not letting the company grow it like they've been doing. My friend's contract was $3.60 / MW on a nominal basis (higher than the $3.00 I gave the company). But, the model needs to be more conservative than the contract because using the contract's number as the base case is too big an assumption.

Aside: I took a two-day course on modeling valuation of banks. What the analyst taught was to basically build future versions of the financial statements. The number of assumptions in doing so easily numbered in the dozens as you had to set assumed growth rates for just about everything and had to say whether or not various margins applied on various line items. Frankly, that scares me to death because I know how bad humans are at predicting the future, how many behavioral biases affect those judgments (e.g. overconfidence, anchoring, and recency), and how a mistake in one assumption can have compounding effects by the time you get to the final number, today's value per share.

I want to keep my models as simple as possible while still capturing what I have to think is the essence of how the company works.

Long answer to that question. :-)

Contrast it to the NoRenewal RV value of $364M and you would get something akin to a gross margin of 45%

Not really, as the RV also includes future estimated maintenance costs and payments to the investors.

- Why use a 6% discount rate? How did the company come up with that figure? Why don't you use a higher figure, and how would you go about doing that?

Because I had to. I have no idea what the yearly cash flows are in whatever model they're using (cash inflows vs. outflows) and I can't back engineer it with any reliability. This means I can't model it on a year-by-year basis which would give me the ability to change the discount rate.

I am assuming that this is their cost of capital, but that is a pretty big assumption. On the one hand, it's based on the fact that companies are supposed to use their cost of capital when discounting to find the NPV of projects. On the other hand, there's the temptation to use a lower discount rate to boost NPV, justified by factors X, Y, and Z.

What this implies is that this is the interest rate they're required to pay their investors on what are, essentially, bonds. When they start floating bonds funded by the leases themselves instead of the tax credits (and they're working toward that), I'd expect them to change the discount rate used to calculate RV, assuming the bonds have a higher (or lower) interest rate. If I knew what the interest rates actually were (and they might be available and I just haven't found them), I'd be able to make a judgment and adjust my model to account for any differences.

Good questions, as I wrote, and I hope my answers adequately addressed them.

Print the post  


What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.