PV-Tech recently announced that its most read blog post of 2011 was an article from IMS Research entitled ‘PV module costs and prices – what is really happening?’, written almost 18 months ago in August 2010. The popularity of this post clearly says something about what the PV industry was talking about throughout all of last year, and still will be talking about in 2012.
Looking back now at the article I thought it would be interesting to revisit some of the predictions made versus what happened in reality, as well as considering what the future holds 12 months on.
Why didn’t the PV industry see 2011’s price collapse coming and why did it hurt so much?
I’m sure no one will disagree that making predictions about the PV industry is a challenge. Whilst every effort can be made to include every possible outcome or variable, an unpredictable political decision, or better-than-expected weather conditions (e.g. Germany at the end of 2011) can make a big difference. But it looks like we got this one right, when we predicted back in 2010 “that installations in EMEA will decline by 80% Q-o-Q in Q1’11. This, coupled with capacities that have been ramped to their maximum in order to serve the strong demand of H2’10, is forecast to quickly reverse the current imbalance between supply and demand, and average PV module prices are forecast to commence their downward trend once again.”
As early as the middle of 2010, IMS Research (and we weren’t the only ones I’m sure) were predicting that 2011 would see oversupply, and ultimately trouble, for suppliers; but still the capacity expansions continued and production lines continued to churn out more modules causing inventory levels to balloon. The industry was still in expansion mode after spending all of 2010 trying to keep capacity up with demand.
Throughout 2011, IMS Research predicted that demand would grow by at least 20%, though few agreed with us. As late as September 2011, PV component suppliers were still saying that “there is no way installations will go past 20GW – it’s going to be more like 16GW”. Our latest analysis and checks show that in actual fact installations exceeded even our predictions and more than 26GW of PV was installed in 2011 (note we count installations, not connections). So why then did 2011 hurt suppliers so much? Most industries would have been delighted at the prospect of 30% underlying demand growth, so why couldn’t they ‘see’ that demand and why did PV manufacturers have such a terrible 2011?
There are two key reasons why.
Firstly, a 30% increase in demand is a somewhat modest increase in comparison to the capacity expansions executed by the majority of suppliers, who looked to double their capacities (and production) during the year. Many suppliers were still able to grow their shipments in 2011, but not by enough to stop their newly installed production equipment from remaining expensively underutilised or their stock levels from growing rapidly. Whilst demand for modules was growing, in comparison to 2010’s triple-digit growth, it felt like it wasn’t.
Secondly, although full-year megawatt shipments and installations grew considerably, much of the demand came towards the end of the year. In fact, nearly 40% of installations happened in the last quarter. What this meant was that the relatively weak demand in the first half of the year led to high channel inventory and collapsing prices as many suppliers sold at a loss, holding out for the good times to return. The fiercely competitive pricing seen throughout 2011 meant that industry revenues actually declined by around 15% for PV modules. More modules were made, more modules were shipped, and more modules were installed; but the suppliers of them made less money.
Costs are down– but not as much as prices
Last year will certainly be remembered for being the year that prices were, for the most part, in free fall. Much has been written about the spectacular decline of ASPs throughout 2011, but the simple facts are that demand in Europe stopped abruptly in early 2011 caused largely by the cancellation of Italy’s FiT, but capacity expansions continued uninterrupted. The result was that whilst modules could easily be sold at almost any price in 2010, the market suddenly became very competitive. With gigawatts of modules flooding the market, and very little to differentiate one from the other, unsurprisingly competition naturally came down to price. However, many were surprised at just how low prices went. Crystalline prices at the end of 2011 were a massive 45% lower than they had been at the end of 2010, exceeding even the most aggressive forecast for price reductions.
It doesn’t take an economist to see that to significantly reduce prices (and survive), you will need to significantly reduce your costs as well, and this was certainly not the case. No supplier was able to reduce their cost structures at the same rate as their prices and margins throughout the industry have certainly felt the consequences. Average gross-margins for crystalline PV module suppliers have fallen into the single-digits, having been in the high twenties a year ago.
For crystalline module manufacturers, the biggest difficulty in reducing costs lies in the comparatively stubborn pricing of polysilicon. Whilst much has been reported of polysilicon prices declining rapidly and reaching record lows, particularly in the second half of 2011, the reality is that the silicon being offered at these low prices is from lower-tier suppliers and sold on the spot market. As the majority of PV-grade silicon is supplied under long-term contracts, fluctuations in spot pricing have only a small effect on the actual average price that suppliers are buying polysilicon at.
How will suppliers become profitable again?
With incentive levels likely to be pared back considerably in 2012, module suppliers’ cost structures remain at the mercy of stubborn long-term polysilicon prices, and hopes for future cost reduction are understandably pinned poly prices falling. This is likely to happen in 2012, especially given that the capacity expansions of Tier-1 suppliers (originally put in motion several years ago) that are due to come online during the year. Tier-1 capacity for polysilicon is predicted to reach close to 300,000 MT in 2012 – enough to serve over 40GW of installations; yet installations are forecast to be broadly flat at 26-28 GW. With Tier-1 polysilicon capacity alone enough to serve the entire market, and suppliers like GCL claiming costs are reaching close to US$20/kg, things are likely to get a lot more competitive for the big polysilicon players. This will create some much needed breathing room back to downstream manufacturers’ cost structures.
With impending polysilicon price drops likely, many suppliers have begun accepting penalty charges for cancelling long-term supply contracts in order to purchase polysilicon, wafers and cells on the spot market instead, or renegotiate new contracts.
Looking back at 2011, there are certain similarities that can be drawn to 2009 (changes in government subsidies causing a sharp slowdown in demand, leading to over supply, falling prices and consequently a strong end to the year and many being surprised that installations grew), but there is one clear difference – it will not be followed by another year of massive demand like 2010 and 2012 will undoubtedly be a lot tougher for suppliers.