Before first quarter 2011 earnings season began for US-listed Chinese solar companies, LDK Solar (LDK) became the first to warn of potential shipment shortfalls. Although LDK did not cite reasons for its lower revenue forecast, it was widely believed by industry observers that policy uncertainties in Italy since March played a major role.
Earlier this week, Trina Solar (TSL) also lowered its shipment forecast, linking the shortfall to Italian uncertainties. A day later, Yingli Green Energy (YGE) joined the pre-announcement party by lowering its own shipment expectations for the first quarter due to project delays in Italy as a result of policy uncertainties.
These three cases may be more company-specific, linked to their respective Italian exposure, since not all companies among the same group experienced the same softness in their shipment order flows. Similar to peers who have recently commented on full year guidance, LDK also reaffirmed its prior 2011 guidance issued in its latest earnings report despite revisions made to its first quarter outlook.
Most notable in LDK’s pre-announcement is the lower guidance for revenues and module shipments. Other quarterly metrics given by the company more or less fall in line with prior guidance. For revenues, LDK now expects Q1 2011 revenues to range between $745-755m, down from $800-850m. Module shipments were also noticeably lowered from 120-140mw to 109-114mw. Based on lower module average selling prices (“asp”) indicated in the company’s investor day conference call, roughly half of the $75m mean revenue shortfall was due to LDK’s module business. Given the company’s potentially higher Italian exposure due to the purchase of Solar Green Technology S.p.A., a system integrator based in Italy, LDK’s module shortfall may simply be mirroring module producing rivals TSL and YGE.
However, LDK revised upward gross margin expectations from 27-29% to 30-31%. Since LDK is not yet a fully integrated module producer, it must rely on either external cell procurement or cell tolling agreements in order to satisfy the cell requirements in its module production. Modules produced in this manner yield much lower gross margin. As a result, module shipments based on externally produced cells are high revenue but low margin in nature. Shipment shortfalls in this product mix would thus disproportionately impact revenues but raise blended gross margin. This may be exactly the situation witnessed at LDK, based on its guidance.
In addition, because there are so many verticals within LDK’s business structure, a number of factors may also lower reported revenues but not corresponding gross profit. For example, the company may opt to sell a lower percentage of its polysilicon production and instead consume it internally. Lost polysilicon sales volume would negatively impact revenues, but since it is consumed internally, higher ratios of lower cost internally produced polysilicon would lower blended silicon costs and thus raise the gross profit levels for LDK’s wafer and module sales. The end result is lower revenues but the same level of gross profit, or in other words, higher gross margin.
Another example may involve LDK accounting for its module component costs differently. Instead of selling wafers and purchasing cells, the company could simply use cell tolling arrangements. Internally produced wafers would be contracted for cell conversion in this example. As a result, LDK would lose revenues due to lost wafer sales, but the gross profit potential remains constant since it is transferred into higher gross profit levels for module sales. As with the polysilicon example detailed above, cell tolling would lower reported revenues but keep gross profit levels constant and thus raise consolidated gross margin.
These are only two of many different avenues LDK could account for its revenues. Since LDK not only operates in every vertical of module production but can also potentially generate revenue in each vertical, actual reported revenues can range widely. However, gross profit should remain fairly constant, dependent on the company’s capacity at each vertical and the current pricing spreads between each vertical. This was the point I tried to convey in less detail in my last LDK earnings forecast:
There may be changes in LDK’s actual product mix, but the overall absolute gross profit potential should not deviate too far from this estimate since it reflects the company’s earnings power potential per unit of installed capacity.
As a result, LDK’s revenue shortfall may appear alarming at first glance, but in fact the company’s revised guidance did not materially lower its implied gross profit potential. Using the mid-point of LDK’s prior $800-850m revenue and 27-29% gross margin guidance, the implied gross profit is $231m. Using the mid-point of LDK’s revised forecast of $745-755m in revenues and 30-31% gross margin, the new implied gross profit is $229m. Assuming operating and non-operating costs remained constant, the corresponding earnings per share (“EPS”) should deviate by less than 1% from LDK’s original guidance.
However, I have revised my prior estimate for LDK’s Q1 2011 earnings. Although lower module shipment expectations do not lower my gross profit estimate significantly, I have revised my wafer shipment assumptions, which appear too aggressive now.
Assuming LDK’s revised guidance on overall shipments is not conservative, it does appear the company experienced flat or even declining sequential wafer production and/or sales during the first quarter. Starting in 2011, the company changed its per-piece wafer wattage assumption from 3.8/watt to 4.1/watt. As a result, LDK’s corresponding capacity should increase accordingly. Wafer capacity starting 2011 should increase by over 7% without any changes to the actual installed physical capacity. Thus, if LDK produced the same volume of wafers on a per piece basis, the actual reported megawatt volume should increase by over 7%.
However, LDK’s total shipment guidance, which would involve internally produced wafers, is not much higher than Q4 2010 levels on a megawatt basis. This implies the actual per-piece wafer production level is flat or lower sequentially. Given wafer pricing was still strong in the first quarter, demand should have remained strong enough for LDK to operate at full capacity. It would appear, as already reported by some peers, that LDK may have experienced some production downtime around the Chinese New Year. I may have underestimated the potential impact of lower operating days during the first quarter of this year.
Below are my revised estimates for LDK’s first quarter 2011 earnings. Although much of the rationale remains unchanged from my prior forecast, a major revision has been made to overall wafer shipment expectations. As usual, these estimates are more a reflection of the company’s core operational earnings and do not include any unannounced gains or charges.
LDK Q1 2011 Revised Earnings Estimate:
Core Wafer: 600mw @ 0.84/watt = $504m
OEM Wafer: 35mw @ 0.48/watt = $17m
Module: 112mw @ 1.65/watt = $185m
Polysilicon: 800mt @ 65/kg = $52m
Total Revenues: $758m
Cost of Goods:
Core Wafer: 600mw @ 0.54/watt = $324m
OEM Wafer: 35mw @ 0.30/watt = $11m
Module: 64mw @ 1.58/watt = $101m + 48mw @ 1.12/watt = $54m, $155m total
Polysilicon: 800mt @ 40/kg = $32m
Total COGS: $522m
Gross Profit: $236m
Gross Margin: 31.1%
Operating Expenses: $45m
Operating Profit: $191m
Net Interest Expenses: $25m
Forex Gain: $5m
Government Subsidies/Other: $2m
Net Income: $145m
Diluted Share Count: 143m
While the actual foreign exchange translation is always difficult to estimate, my $5m gain is based on the assumption LDK did not materially change its currency or hedging exposure relative to Q4 2010 levels. LDK historically did little hedging; thus, with the euro appreciation vs. the USD, the company should experience exchange gains as a result of its euro assets, notably accounts receivables in euros. In addition, the company should experience gains as the rmb gained vs. the USD, notably from USD-denominated debt such as its convertible debt outstanding.
Additionally, as noted in its recently released annual report, LDK last December repurchased the 15% stake in the polysilicon plant it sold in late 2009. As a result, there should not be any minority payments made during the first quarter related to this formerly divested stake. However, I did not include any potential charges which may have resulted from the repurchase transaction.
Disclosure: I am long LDK, TSL, YGE.