Description

I think Suntech equity is worth about $3.50 based on replacement cost. In America the equity would be a bagel, but since their lenders are Chinese banks, it won't be forced to file and will have option value. The solar industry is facing a severe capacity glut that could take a number of years to fill. Being a commodity, I believe all products in the solar supply chain will eventually sell for slightly less than variable cost until supply and demand come back into balance. Also, I believe STP's Q1 accounting was rife with fraud.

The Solar Industry

2008   2009E

Global Module Supply       12GW  17GW
Global Module Demand     6GW    6GW

Demand is about two-thirds commercial, one-third residential. Anecdotally I've been told that last year the market was 50% retrofit, 50% new construction. Commercial construction on the precipice of a huge decline seems to be the elephant in the room that sell-side analysts almost never discuss. With one-third of shipments tied to commercial construction and government subsidies getting cut, I think the industry would be lucky to see shipments flat, but 6GW seems to be the sell-side consensus for 2009. Post-2009, if demand resumes 30% growth, it would take four years for demand to catch up with supply. The essence of my thesis is that the solar industry is still in the denial stage of the fear/greed cycle, and investors haven't yet conceded how bad the oversupply is and that this is a secular bear market for solar product prices. Much like in the early stages of the housing bust, analysts are hoping that shipments will explode in the second half and prices will bottom after only declining for a few quarters. I have no idea what demand will be, but when gross margins are zero, it doesn't really matter.

At current grid power prices solar power is a fantasy. An artist's rendering of the investment scenario looks like this for an end user:

Wholesale Module/kw   2,600
Installation Cost/kw     2,500
Total Module Cost/kw   $5,100

Approx. Retail Grid/kwh   $.04
Hours of Operation/Day  X 6   This is a best case scenario like in southern California
Days/Year                     X 365
Annual Savings              = $88

Return on Installed Cost 1.73%

With at least half the installed cost being labor, it's mathematically impossible for module prices to come down enough to make a solar project economical, unless coal and natural gas prices rise substantially, or maybe if efficiency factors rise to 100% (currently 17% is gold-standard, rising at maybe 1%/year). Put another way, if we wanted to replace 100% of global electricity consumption with solar power, we would have to spend $5/watt X 2 TW X 6 = $60 trillion, 90% of one year's global GDP.

Regardless, solar installations continue mostly because of government subsidies. Germany, for instance, pays solar panel owners about $.35/kWh generated. Apparently many banks will provide project financing to these panel owners, making it a pretty extraordinary levered return. Personally I think this is an irresponsible lending practice - in the event the subsidy is removed the projects will be underwater. The U.S. used to provide a 30% tax credit, now it's an upfront 30% grant. Spain recently slashed their subsidies due to budget pressures. The solar bulls point to China as the next source of subsidies. Their stimulus plan anticipates about 500MW of installations, and bulls believe there will be more to come. It's my gut feel that China doesn't care about the environment, and these one-off subsidies will always be immaterial and largely symbolic. Wasting money on solar projects seems like an affliction of affluence.

Q1 Gimmickery

In Q1 STP reported $315m of revenue, about 10% shy of consensus estimates. However, $100m of shipments were to GSF, an 86%-owned subsidiary of STP (when asked on the conference call who owns the other 14% management said they can't disclose this, which suggests to me that it's the chairman or other affiliate). Through a blatant mockery of accounting rules GSF isn't consolidated and is accounted for as an equity-method investment, despite also having 50% voting power, because management says STP is a ‘limited partner' in GSF. GSF can make capital calls to solicit equity funding from the partners, and STP can veto these capital calls. In Q4 2008 STP contributed $94m to GSF - does anyone think this number is coincidence? Essentially I believe STP moved inventory from one STP warehouse to another.

Additionally, STP marked up inventory despite module prices being down 15% from Q4 2008. When asked how much the write-up was, management stated that they didn't know. Despite sales being down 25% sequentially, gross margin was up in Q1 to 18% from an adjusted 13% in Q4. Either the write-up was huge, or it's possible that the shipments to GSF were transferred at inflated prices.

Despite reporting one of the most fraudulent quarters in the history of capitalism, STP is back to the price it was on the day before reporting just because rumors of a second round of Chinese stimulus purchases were started.

If we include the GSF shipment in inventory (since after all, the projects won't be finished until October or November), STP inventories were up 40% sequentially. Across most suppliers inventories are building dramatically, which bodes poorly for prices when it comes time to purge it. I believe suppliers are building inventory because they're over-producing to keep per unit costs low on the income statement. It's also possible that since many polysilicon contracts are fixed-value take-or-pay, that the polysilicon suppliers are cramming polysilicon down the channel. Either way they're delaying the inevitable.

Below is the total inventory and revenue for the five largest cell/module manufacturers in Q1 '09 versus Q4 '08:

                 Q1 '09   Q4 '08   Seq. % Change
Inventory   1,652     1,309    26%
Revenue    1,348     1,803    -25%

Valuation

I don't think relative valuation is relevant in this case because I think most or all of the sector is overvalued. Note that STP still trades at 7x peak gross profit. I think about valuation on a mid-cycle basis.

Being a commodity I don't see why crystalline module suppliers should trade at much more than replacement cost. Industry experts tell us that current capital costs for a large, low-cost integrated cell and module facility are about $.88/watt, which jives with the roughly $800m STP has spent to build out 1GW of capacity. There's also about $700m of net working capital, so let's say the enterprise is worth $1.5b, less $875m of net debt (pro-forma for the equity deal), the equity's worth $625m, over 176m pro-forma shares gets you to $3.50/share.

Catalyst

Persistent estimate misses