Solar manufacturing stocks are currently at the mercy of the market’s reaction to overcapacity conditions. Second quarter results made palpable that industry is for the second time in last five years in solar material glut. Herman Zhao, CFO of JA Solar(JASO) during the conference call estimated global module capacity at 100GW for all tiers, and within it, 80GW for tier ones. Demand figures compiled by Bloomberg’s New Energy Finance (BNEF) see 2016 to reach 67GW, a figure, also mentioned by Zhao. 2017 BNEF forecast differs from Zhao’s who calls for 71GW and from Trina Solar’s (TSL) forecasting 70GW. Bloomberg’s analysis sees as much as 25% growth for that year, and for 2018 they expect a 23% increase. Using BNEF data, by 2018, all of the current capacity would be sold as the demand would reach 103GW. By 2017, if BNEF is right, all tier one capacity would be sold against the demand of 83GW. Unfortunately, the data is from February, and it appears too optimistic.

23 module manufacturers, also considered by BNEF as tier ones, have combined module capacity of 67GW which is around 99% of this year’s demand. Among them, six are public companies listed in the US, and based on their capacity; they place among top 10.

In the same way, the glut story of 2012/2013 divulged, the second glut will never see top Chinese companies reduce their shipments. The reason is simple. The global glut is a bottom of the industry’s supply chain condition and cannot be reduced by top manufacturers. If they did not expand, someone else would, and they would lose own business opportunities, but glut would not be eliminated. When ASP price drops, only a large-scale organization can lower costs, using volume expansion as a cost reduction. While optics point to severity to be applied equally, the glut creates demand gaps for tier 2 and tier 3 competitors and pushes them out of the market. Furthermore, capacity growth adds efficiency improvement from new equipment and reduces cost as long as an enterprise has normal cycle utilization. Companies capable of capital spending can afford to expand and by doing so, they eliminate competition. The expansion is a natural, a must, activity for the top members of the industry. Therefore, the glut is not miscalculation or an error; it is a low, unattractive period in the holistic cycle of growth, which ends with consolidation (M&A), elimination (bankruptcy) and even more powerful recovery for survivors. Those dynamics are precisely the same for any glut conditions, including one prevailing in oil.

In my estimates, top Chinese companies will expand to replace mainly small competitors who will sell at the loss and ultimately shut down, sometimes by Q4 2017, making the turnaround for the second glut, just a year-long affair. The immediate 12- month impact on big and small companies will come down to specifics of gross margin which is a difference between cost and average selling prices. Lower gross margins will obstruct the markets to see further into a future and reduce the value of the equity until recovery is on the horizon.  What we see financially is companies delivering most impressive results, yet being already sold off this year, in expectations of the glut conditions.

The impact in 2012/2013 caused loss of earnings, even for the top names, but this time, the glut is not trapping the same business formats as it did in 2012. The expected price drop may not be as dramatic, and in some cases, have already happened. The primary change is that Chinese companies branched out into the solar plant development businesses, and most importantly, took on ownership of solar plants to receive revenue and net income from sales of electricity. Trina Solar (TSL), which has a “go private” offer, owns today 1.3GW of solar projects, mostly in China, followed by Jinko Solar’s (JKS) 1.1GW in the same location. Canadian Solar has 472MW globally, and China included, but by the end of this year, it will reach 1.3GW of plants having the majority of it in the US.

Among the six companies listed on the US exchanges, in line with views in my recent articles on SA, the best choices for investment are JinkoSolar(JKS) a gross margin manufacturing leader, which this year has successfully challenged Trina Solar (TSL) for the first place in the volume of MW sold.  Canadian Solar (CSIQ) a Canadian incorporated company, which has a mix of global manufacturing with a majority of the capacity still in China, but with uniquely North American, in addition to also global, solar plant building ambitions. First Solar (FSLR), considered the best company in a sector, and undoubtedly, a money tree among mostly highly leveraged shrubs.

For First Solar, based on its 19% GM guidance for 2016 and achieved average of 25% for the first half, I expect the second half of the year to see 14% GM. The roster of sold, still collecting revenue projects, has 271MW left for this year. This remainder of the PPA contracted and sold capacity is worth $600M, leaving $1.5B from the low end of $3.8B guidance to third-party module sales and new, yet to be announced, solar plants sales. The company has 541MW of projects having PPA agreements with commercial operation dates (COD) in 2016. Those include 250MW Moapa in the US, 150MW in India and 141MW in Chile. The company expects to sell Moapa and California Flats project this year; the last one planned for COD in 2018. There are about 303MW of projects with 2017 COD date, making fewer projects available for sale than ones sold or to be sold in 2016.

I see FSLR have limited potential to raise GM in 2017, beyond 14% range, until  5 Series module takes over at the end of 2017, and despite 24% gross margin for components segment in Q2, which included a third party, and to own projects, module sales. In my opinion, components’ segment GM represents the legacy project margins, a driver which is no longer available in the future forecasts of solar plant sales this year.  

The company has 3.2GW module capacity and switch to “5 Series” module, based on the existing efficiency indicates improved gross margin dynamic. Growing but not yet significant third party module sales (maybe $1B) to the ratio of solar plants in 2017, will not replace gap in revenues, estimated by me to reach only $3B.

The company expects in its corporate plan to reach 6.5GW capacity by 2020. The expansion is dependent on global market conditions, which make FSLR a lot more vulnerable in the bottom-of-the-barrel market of selling to developers than the exclusive niche market of selling solar plants. This market is no longer exclusive and becomes increasingly competitive with likes of Canadian and JinkoSolar entering the bidding process. Also, a 2GW add would cost $800M, in total, move from 3.2GW to 6.5GW could take up to $1.2B in CAPEX. FSLR may not be seen as particularly susceptible to hiccups, most recommendations on SA see FSLR as a buy, but a combination of independent events with internal growing pains could certainly create extra pressure even before, weaker, 2017 guidance.  

Canadian Solar and Jinko sold modules in a range of 16 to 18% GM in Q2 but expect to drop it to 13-15% range for Canadian, and perhaps to 16% for JinkoSolar by 2017, although the transcript of the Jinko’s conference call also mentioned 13%. Neither will reduce the module sales in 2017, as explained in the opening paragraph, but each company will have limited growth for MW volume without an increase in capacity, especially if they continue to build solar plants. Most likely they may hold back on the venture to preserve cash, the order of things which forces module sales, even though GM is low. Canadian, can improve its efficiency with new equipment put online in Thailand this month and register a gross margin gain based on tariff-free sales, but it will not be able to fight ASP any better than Jinko and what is experience, most likely worse. Jinko’s Malaysian plant will also supply 90% shipments to the US removing most of the tariffs during H2. Canadian has a pool of solar plants for sale to add revenues. Under the glut conditions with present auctions of solar/wind projects belonging to a former heavyweight now bankrupt, SunEdison, most likely the market for those is saturated, a conclusion supported by 3-5% drop in gross margin estimates for those assets since Q1.

For all three, due to the ASP, the same volume of third-party module sales estimated for 2016 could see a reduction in revenue by 10% in 2017. I think that drop could be compensated by additional module sales for CSIQ and JKS out of existing capacity, by selling to third parties in 2017 the volume sent to own projects in 2016. If Canadian happened to hold the US portfolio it could also add more of electricity sales, similar can be said for Jinko as plans see the company add solar plants during this year. On a positive side, Canadian would be able to help the cash flow from sales of Japanese assets into J-REIT IPO; I also expect JKS to IPO its power business in Hong-Kong.

In my low-end scenario estimates, I suspect FSLR to trade as low as $30 per share by mid-2017, based on 12 PE and $2.5 EPS setting a tentative bottom. The income statement for 2016 shows $3.90 per share, including equity method sales to own yieldco, but the market is already looking into 2017 with the choppy trading of PE, now around 10. For Canadian Solar without the solar plant sales, I suspect 2016 to deliver just about $1.84 with an average GM of 15% for the second part of the year. The stock trades at a range of $12 or around of 6.5 PE. The share price could drop below $11 with EPS of $1.52 for 2017, using the same PE. Solar plants will add to the bottom line, most likely with sales executed at the end of this year and through 2017. The strongest in this scenario, JinkoSolar, could make North of $3 in 2017. At current PE of 3.5 with my expectations of $4.56 2016 EPS stock trades at $16 per share. Having PE of 6 with the EPS of $3 the price would be at $18.00 suggesting that JinkoSolar selling below the expected bottom.

To be strategic, for the first time in last six years, I am staying away from the manufacturing stocks, unless the bottom price targets will be broken sooner and without a change in the narrative (for better or worse). Perhaps inconsistent with the article, at the first glimpse, I expect stability with yieldcos and substantial rise in their value as buying solar, and wind projects should become even more affordable, offering moderate add-ons to the existing cash available for distribution. I see NRG Yield (NYLD) and Pattern Energy (PEGI) as vehicles to carry on my renewable interest during a period. In particular, NYLD is the entity I see to get that benefit. My choice takes into an account a perception of cushion 8point3 Energy Partners LP (CAFD) could play to GM of own sponsors, something which is less so obvious for a sponsor of NYLD.

SunEdison’s TerraForm yieldco, (TERP) and (GLBL), are also attracting a lot of healthy attention from the industry bidders, and when the dividend is reinstated, and listing requirements are clarified, I expect the market to become further reinvigorated. A new sponsor ownership, along completion of asset sales, would remove the ghost of SunEdison’s failure.

In the similar outcome of the first glut, I expect equity appreciation in solar manufacturers during the recovery stage. I can see at least 50% growth by 2018 in relationship to the bottom prices I expect for CSIQo in 2017. I suspect that JKS can do less but still 30-35%. Those percentiles relate to the growth of the EPS in 2018 versus 2017.

The recovery signals would include the need to add the capacity for 2018. Another, the US solar plant development to come back for 2018/2019 with new projects, and another one yet, all three companies to continue to compete among themselves and win projects globally.

Lastly, I also expect ASP to stabilize by Q4 2017. The cost for internal production is expected to drop to $0.29/watt by the same quarter, as published in the presentation to investors by Canadian this month. Jinko is already looking to drop its blended costs of $0.41 to $0.37 per watt by the second half, a level of internal cost, already reported in Q2. The dynamic of cost, therefore, has a lot fight left. If the ASP of $0.40 per watt is tier 1 selling point in Q4 2017, at the cost of $0.32, gross margin would reregister 20%. A potent level to operate for Jinko and certainly more than optimal for Canadian, the company coming close to it in 2014, its best EPS year on record. However, 20% is only a shadow of past glory for First Solar. Unless the premium identified by management as part of the offering in Q2 adds 5% by the end of 2017, FSLR reflective of past results may need “6 series” and $1.2B capital expansion, delaying the timeline for the market conviction and recovery.