In the Notice of Initiation, the Commission invited interested parties to contact it in order to participate in the investigation. In addition, the Commission informed the CCCME, known exporting producers in the PRC and the PRC authorities of the investigations and invited them to participate.

Interested parties had an opportunity to comment on the initiation of the investigation and to request a hearing with the Commission and/or the Hearing Officer in trade proceedings.

On 19 July 2017, the Commission disclosed to all interested parties the essential facts and considerations of the investigation and invited them to comment within 14 days. The Commission received replies, within the deadline, from 20 interested parties, namely the association of the Union producers, seven Union producers, two associations of users, four upstream and downstream interested parties in the Union, four Chinese exporting producers, the CCCME and the Government of the PRC. Subsequently, the Commission sent an additional disclosure document to all interested parties and invited them to comment. This redisclosure was limited to only two elements of the methodology for establishing the MIP and a provision regarding the entry into force of this regulation.

All the interested parties who replied to the request for information considered that a variable duty in the form of a MIP (‘variable duty MIP’) is a more appropriate form of measures than the previous ad valorem duty coupled with the price undertaking (‘undertaking MIP’). In particular the interested parties considered that a variable duty MIP will be more transparent, predictable and enforceable. The interested parties considered that a variable duty MIP would reduce the administrative burden and costs for the importers. Some of the interested parties encouraged the Commission to ensure that the new form of the measures does not impose significant restrictions on the Union companies in terms of their business deals with manufacturers worldwide. In their view, these restrictions resulted in significant risks, liabilities, expensive due diligence and delays for the Union importers. The same parties also claimed that the existing cap on volumes of imports contained in the undertaking should be removed as it added further administrative burden and did not serve any purpose as the imports were anyway significantly below it.

The Commission also agreed with the interested parties that the variable duty MIP should not be accompanied by a list of additional restrictions and caps. Indeed, the exports have always been well below the annual level. It will be up to the Union customs authorities to verify if the companies involved did not enter into any cross compensation agreements and other arrangements circumventing the MIP.

The distinction between mono-crystalline and multi-crystalline products also fits into the reasoning brought forward in the expiry review investigations to strike an appropriate balance between competing interests. On the one hand, it will better protect the Union industry, which is increasingly focusing on manufacturing high-end mono-crystalline products for the rooftop sector. On the other hand, such a distinction will serve better the interests of unrelated importers and engineering, procurement and construction companies (EPCs) active in the utility-scale sector, which need access to cheap commodity type multi-crystalline modules to be able to compete with other renewable energy sources in technology neutral tenders.

Mono-crystalline and multi-crystalline cells can be easily distinguished by customs authorities. Multi-crystalline cells are made of multi-crystalline silicon (multi-Si) consisting of small crystals. Mono-crystalline cells are made of mono-crystalline silicon (mono-Si), a continuous crystal. Mono- and multi-crystalline cells are never combined in one device, therefore there are no modules that are made with both mono- and multi-crystalline cells. Multi-crystalline modules are made exclusively of multi-crystalline cells; and mono-crystalline modules are made exclusively of mono-crystalline cells. Mono-crystalline products have higher efficiency of converting sunlight into electrical current, which results in a higher output per area of space. Mono-crystalline products can be identified from multi-crystalline products by physical inspection. The multi-crystalline cell is perfectly rectangular. A mono-crystalline cell, by contrast, has its four corners cut off.

Therefore, the Commission considered that there should be separate MIPs for mono-crystalline and multi-crystalline cells and modules and each of the four product types should have its own TARIC code.

In the expiry review investigations the Commission became aware that throughout most of 2016 the undertaking MIP adjustment mechanism did not follow global price decreases, and hence no longer reflected the non-injurious price, as established in the original investigation.

Therefore the Commission investigated whether there was another benchmark, which would better reflect the non-injurious price level as established in the original investigation and global cost and price decreases.

Therefore, the Commission concluded that using the solar industry learning rates to adapt the MIP would introduce considerable uncertainties, which would render any precise predictions on price developments impossible. Therefore, the Commission decided to rely on another benchmark, which is based on more recent, transparent and reliable data.

Most of the interested parties claimed that the new adjustment mechanism should be based on the price quotes by Taiwanese market intelligence agency PV Insights. Only Solar World, the largest European producer, considered PV Insights unreliable. PV Insights was also considered to be the most widely used by the interested parties. Several parties pointed out that the prices quoted by PV Insights and its price development trends were in line with the prices and trends quoted by another index trusted by the industry, i.e. Energy Trend PV (run by another market intelligence also based in Taiwan). By contrast, the prices quoted by the index used at present, i.e. Bloomberg database were subject to much more volatility and around December 2015 Bloomberg spot prices index started following a different trend from PV Insights and Energy Trend PV. The Bloomberg database is based on voluntary price quote submissions, which means that it has captured only a very small part of the market.

The Commission requested the ITRPV to provide more information on the solar prices that they used to calculate the solar industry learning rate. The ITRPV provided the price data and indicated that they currently use two sources — PV Insights and Energy Trend PV. Before the end of 2016, ITRPV had used a wider basket of prices, including Bloomberg spot prices index. Given that PV Insights is one of the two sources used by the ITRPV and that the prices quoted by PV Insights and Energy Trend PV have been broadly in line with each other, the level and evolution of prices that the ITRPV used to calculate the learning rate have been closely in line with the data reported by PV Insights, especially since the end of 2016.

The Union cell and module makers and their association also claimed that PV Insights is currently under investigation by the Taiwan Fair Trade Commission, Taiwan's national competition authority, following a complaint by the Taiwanese solar manufactures association. This investigation was opened on the basis of the claims that PV Insights index was dominated or even manipulated by Chinese price quotes and that the index's price level is below the cost of production in Taiwan. For them, relying on PV Insights would thus not be appropriate.

The Commission recalled that the downstream and upstream industries had considered PV Insights the most reliable index in their daily business. PV Insights was also one of the key price benchmarks for the ITRPV reports when estimating the solar industry learning rate. Until now the upstream and downstream industry did not cast any doubt on the reliability of the PV Insights index. Finally, the Taiwanese authorities have not yet made definitive findings on the allegations. The Commission will monitor the developments of the Taiwan Fair Trade Commission and will consider any action necessary in view of potential findings. At this stage, the Commission therefore rejected this claim.

The upstream and downstream European interested parties as well the Government of the PRC and the CCCME took issue with several aspects of the new MIP decrease.

First, these parties considered that the starting MIP was too high. They claimed that for mono-products the new variable duty MIP would be even higher than the previous undertaking MIP. Setting the variable duty MIP at an even higher level would, in their view, be in contradiction with the findings of the review investigation that the undertaking MIP was set too high relative to the non-injurious price. Consequently it would need to be brought down to the appropriate level. Some parties also claimed that the MIP applicable in the first quarter of 2017 was not an appropriate starting point for the variable duty MIP as the Commission had itself found that it was out of line with the global price developments.

The Commission took account of these comments and developed a new quarterly gradual decrease of the variable duty MIP. As the undertaking MIP was frozen from the second quarter of 2017 onwards, the Commission frontloaded the starting point of the gradual decrease. This starting point is set at the level of the frozen undertaking MIP decreased by the value of two quarterly adjustments that should have occurred while it was frozen, namely the second and third quarter of 2017.

Second, several parties considered that the ending variable duty MIP, i.e. the one applicable at the expiry of the measures in September 2018, was also too high. They claimed that, according to the PV Insights price quotes that were available following disclosure, the global solar prices already decreased. The Commission accepted the proposition that the latest available data should be used as the most appropriate proxy for the ending quarter. Therefore, it set the final variable duty MIP at the level of prices in the latest quarter available i.e. the second quarter of 2017.

These parties also claimed that the Commission's forecast that the decrease in solar prices would slow down was unfounded. However, the analysis of the PV Insights long-term price curve suggests that solar prices are cyclical — historically solar prices were dropping aggressively throughout several quarters and then stabilised or even slightly increased thereafter. During the current cycle the prices of modules have been continuously falling for a relatively long period of time, i.e. since the fourth quarter of 2015. At the same time, the cell prices that used to follow a similar trend have already stabilised or even slightly increased. The fact that the prices for the main raw material, i.e. cells, stabilised after a particularly long period of decreasing prices reinforces the Commission's forecast that the module prices would eventually stabilise too. Therefore, the claim was rejected.

The non-integrated module makers also claimed that the variable duty MIP indicated in the disclosure document decreased much faster for modules than for cells, which would affect disproportionately their profit margins. The Commission pointed out that such a difference in the slope of the decrease is an unavoidable consequence of the fact that the undertaking MIP for cells was much closer to global market prices than the undertaking MIP for modules. In addition, following disclosure, the Commission decreased the variable duty starting MIP, therefore the variable duty MIP will no longer be above the undertaking MIP for mono-cells.

Following redisclosure, upstream and downstream companies as well as their associations and the CCCME repeated their view that the MIP was too high even if it was adjusted further downwards, which some of them welcomed. On the other hand, the Union producers and their association reiterated that the MIP was too low and that it did not reflect the non-injurious price; that the MIP decreased disproportionally faster for modules than for cells and that PV Insights was not a reliable benchmark.

The Commission observed that none of these parties brought forward new arguments on the two new elements disclosed (frontloading of the gradual decrease and use of the most recent quarterly data). Rather, they repeated their general approach on the MIP that they had already outlined post disclosure, adapted to the new levels of the quarterly MIPs. Therefore, the Commission considered that it had already addressed the essence of these claims after disclosure.

Several parties also claimed that the period for comments was too short. The Commission considered that one working day was sufficient for the parties to comment, given that the disclosure was limited to only two elements of the methodology for establishing the MIP and a provision regarding the entry into force of this regulation. Therefore, the Commission rejected this claim.