Message Font: Serif | Sans-Serif

No. of Recommendations: 3
Hi everyone,

I had a conversation with Travis Hoium yesterday. He's a writer/analyst for Fool.com and has followed solar energy for quite a while. Anyway, he and I discussed the model and he raised what I think were a couple of good points and (important!) he found a significant error in part of the calculation.

First, the error. In rows 18 and 21, columns H - K, where I'm supposed to be putting the total Retained Value each year by adding the previous year's total and the new incremental amount, I wasn't. In G18, for instance, I added 2013's incremental Under Energy Contract RV (G20) to 2012's RV (C18). That was correct. But then I copied that formula over to H18 - K18, which meant that I was using the wrong previous year number. Similar situation for row 21. The correct formula for H18 should have been H20 + G18 (instead of D18, which it was). Before it was \$277 + \$305 (from D18) = \$582. The correct amount should have been \$277 + \$497 (from G18) = \$773.

Changing to the correct formula gives \$773, to \$1,101, \$1,458, and \$1,829 for H18 - K18 and to \$649, \$926, \$1,228, and \$1,543 for H21 - K21. That, of course, raises the RV at end of 2017E to \$3,372 (K25 and K28).

Second, a couple of points Travis made.

First, he argued against increasing the \$\$ / incrmemental MW amount in calculating incremental Est. Nominal Pymts Remaining. Recall that it is the \$\$ / incremental MW installed (row 15) * incremental MW installed (row 4) that leads into the RV calculations further down the spreadsheet (rows 20 and 24).

I started this at \$3 / MW (all dollars in millions) and grew that as the company saved on costs. Travis argued that SolarCity would have to lower its selling prices in order to remain competitive as solar components continue to fall in price. Competition would come from utilities and from homeowners deciding to just install the things themselves. This argument seems reasonable, while a climbing \$\$ / MW implies increasing margins at the company. So, keeping this at \$3 / MW for all years lowers K25 to \$3,038 at this point. Dropping it by 5% a year lowers K25 to \$2,780.

Second, he argued that the 3x multiple I assigned at the end (lower than the 3.5x it was trading for at the time and the 4.24x it's trading at today) didn't make sense. Why should the market assign a multiple like that?

You can value a company however you like, but the valuation needs to make economic sense. RV is future cash flow to the company after expenses, discounted back at 6%, and is what the company would receive over the next 30 years if it stopped growing. It's not FCF (the company would still owe taxes on it) and it doesn't take G&A expenses into account. Companies often trade at P / FCF multiples in the single digits (which anticipates some growth). A more appropriate multiple might be 1.5 x or 2x. At 1.5x, market cap would be \$4.17 B (using the 5% decline to \$\$ / MW version) instead of the \$4.8 B from the earlier version of the model.

So, more conservative assumptions lower the valuation by about 13%. I can live with that.

Cheers,
Jim