In its Notices of Initiation, the Commission stated that it had provisionally selected a sample of Union producers. The Commission selected the sample on the basis of the largest representative volume of EU sales, taking into account production volume and the geographical location, which could reasonably be investigated within the time available. The sample consisted of six Union producers for modules and three for cells. Both vertically integrated and non-integrated producers were included in the provisional sample. The Commission invited all interested parties to comment on it. All companies that were provisionally selected agreed to be included in the provisional sample.

Comments on the proposed sample were received from several interested parties. They criticised the fact that the names and the location of several Union producers were kept confidential which prevented them from making comments on the selected producers' share with respect to the total production and sales volumes of the Union industry.

The Commission recalled that all sampled Union producers, except for SolarWorld, WARIS Srl (‘Waris’) and Sillia VL (‘Sillia’), requested at initiation stage that their names be kept confidential. The Commission respected those requests, but invited them to confirm their wish to remain anonymous throughout the reviews investigations and to provide good cause for their request. All of the companies concerned, with the exception of one, confirmed their initial request and provided justification for their requests. In particular, the companies indicated that they fear losing business activities in the PRC, and/or supplies of raw materials and components sourced from the PRC. These reasons were found to be warranted. Thus the Commission decided to accept their claim for anonymity and to reject the requests by interested parties to reveal the identity and location of the sampled Union producers. Among the anonymous companies, only Jabil Assembly Poland sp. zoo. (‘Jabil’) agreed to have its name disclosed in the final sample.

One exporting producer argued that the Commission failed to define the Union industry prior to selecting the provisional sample thereof. Therefore, it could not comment whether the latter is representative. In addition, some of the selected Union producers, like SolarWorld, are vertically integrated as they manufacture cells which are mainly used for their own modules production. Consequently, there is a risk that the production of the same final product, modules, has been counted twice.

It follows from the Notices of Initiation that ‘Union industry’ refers to all Union producers of modules and key components (i.e. cells). In addition, the Union industry was already clearly defined in the original investigation. Finally, the captive use of cells production has been deducted when examining the standing and the representativeness of the Union industry. Therefore, this claim was rejected.

Several other interested parties argued that the Commission should not have included in the sample Waris and Sillia since they are small Union producers of modules. As such they have a very specific business model which is not representative. Rather, the Commission should have included big and middle sized companies such as Jabil.

The Commission rejected this argument as it included in the sample a significant number of big and middle sized modules producers. Waris and Sillia were included in the sample in order to ensure a wider geographical representativity.

After the initiation of the proceeding the Commission had to exclude from the sample the company Sillia due to lack of cooperation. As a result, the remaining sampled Union producers accounted for 38,8 % of the total EU sales and 55 % of total Union production of modules. For cells they accounted for 76,6 % of the total EU sales volume and 77 % of the total Union production. Therefore, the modified sample was still considered representative of the Union industry.

In order to decide whether sampling is necessary and, if so, to select a sample, the Commission asked unrelated importers to provide the information specified in the Notices of Initiation.

Two unrelated importers provided the requested information and agreed to be included in the sample. In view of the low number, the Commission decided that sampling was not necessary.

In view of the apparently high number of exporting producers, sampling was envisaged in the Notice of Initiation of an expiry review for the determination of dumping, in accordance with Article 17 of the basic Regulation. To decide whether sampling would be necessary and, if so, to select a sample, the Commission asked all known exporting producers in the PRC to provide the information specified in the Notice of Initiation of an expiry review. The authorities of the country concerned were also consulted.

Eighty-one exporting producers in the country concerned (often groups of several companies) provided the requested information and agreed to be included in the sample. In accordance with Article 17(1) of the basic Regulation, the Commission selected a sample of three groups on the basis of the largest representative volume of exports to the Union which could reasonably be investigated within the time available. All three exported modules to the EU, one exported also cells. In accordance with Article 17(2) of the basic Regulation, all known exporting producers concerned, and the authorities of the country concerned, were consulted on the selection of the sample. No comments were made.

Following disclosure, one interested party argued that the sample of the exporting producers is inappropriate as it differs significantly from the sample of Union producers in terms of the sampled companies' production and production capacity.

Since the PRC is considered to be a non-market economy country, normal value was determined on the basis of the price in a market economy third country, in accordance with Article 2(7)(a) of the basic Regulation. For this purpose the analogue country had to be selected.

India served as analogue country in the original investigation. In the notice of initiation, the Commission informed interested parties that it envisaged the USA and India (as requested by the applicant), as well as Japan, Malaysia, South Korea and Taiwan as potential analogue countries.

The Commission received comments on the choice of the analogue country from six interested parties. All of them supported Taiwan. One supported South Korea, as an alternative. At the same time all of them opposed the USA because of the existing trade defence measures and distortions caused by the domestic subsidisation. Three parties also opposed India because of domestic inefficiencies.

After contacting all known producers of the like product in all potential analogue countries, the Commission secured cooperation from one producer in Taiwan and one in the USA. The cooperating producer in Taiwan produced mainly cells but sold both modules and cells. The majority of modules sold by this producer were manufactured by third parties pursuant to a tolling agreement whereby the third party would receive cells and a tolling fee and supply the modules. Some modules were simply bought from third parties to whom the cooperating producer would be selling cells. All these modules were then sold under the cooperating producer's brand. The cooperating producer in the USA is related to the major Union producer — SolarWorld — and produced both cells and modules, but sold domestically only modules.

Both the Taiwanese and the US markets appear to be competitive, with several domestic producers and significant imports from abroad. However, the solar panels market in the USA is protected by anti-dumping and countervailing duties on the imports from China and anti-dumping duties on the imports from Taiwan. Taiwan does not have any such measures in place.

As the cooperating producer in the USA did not sell cells on the domestic market during the RIP and, unlike the USA, the market in Taiwan is not shielded by trade defence measures, the Commission considered Taiwan as a more appropriate market economy third country.

Following disclosure, two interested parties argued that the Commission chose an inappropriate analogue country producer as it produces smaller quantities of cells than the exporting producers and has their modules produced by third parties under a tolling agreement.

Under Article 2(7) of the basic Regulation, the Commission does not choose analogue country producers. It chooses market economy third countries. Despite the Commission's extensive efforts to secure wide cooperation in all potential analogue countries only one producer in Taiwan cooperated. Furthermore that producer was also the only cooperating producer in the entire investigation that sold both modules and cells. Finally, the cooperating producer in Taiwan operates in a market economy third country, supported as the analogue country for this case by all the parties that commented on this issue, including one of the two parties that raised this claim. This claim was therefore rejected, as there was no alternative, and the choice was, in the light of the circumstances of the case, appropriate.

The information received from the cooperating producer in the analogue country was used as a basis for the determination of the normal value, pursuant to Article 2(7)(a) of the basic Regulation.

The Commission first examined whether the total volume of domestic sales for the analogue country producer was representative, in accordance with Article 2(2) of the basic Regulation. Following this provision, the domestic sales are representative if the total domestic sales volume of that producer and of the like product to independent customers on the domestic market represented at least 5 % of the total export sales volume of each exporting producer of the product concerned to the Union during the review investigation period. In case of representative domestic sales of the like product, the profitable domestic prices were used as normal value. When the domestic sales of the like product were not representative, the normal value was constructed in accordance with Article 2(3) and (6) of the basic Regulation. In the case where the like product was not sold in representative quantities, the Commission furthermore decided not to apply the last sentence of Article 2(2) as the representative sales per company were below 1 %, which is too little to be representative in the meaning of that provision.

• the weighted average selling, general and administrative (‘SG&A’) expenses incurred by the cooperating analogue country producer on domestic sales of the like product, in the ordinary course of trade, during the review investigation period; and

• the weighted average profit realised by the cooperating analogue country producer on domestic sales of the like product, in the ordinary course of trade, during the review investigation period.

For the product types sold on the domestic market, the average SG&A expenses and profit of transactions made in the ordinary course of trade on the domestic market for those types were added. For the product types not sold at all on the domestic market, the weighted average SG&A expenses (between 2 % and 5 %) and profit (between 1,5 % and 6 %) of all transactions made in the ordinary course of trade on the domestic market were added.

Another interested party argued that the Commission failed to provide a breakdown of the volume and representativeness of the production and sales of cells produced by the analogue country producer and modules produced under the tolling arrangement. According to the party, the Commission failed to explain how the cost of production of the modules made under the tolling arrangement was calculated. The same company requested that the Commission provides further details on the calculation of the constructed normal value of both cells and modules. Furthermore, the Commission should provide details on the tolling partner's cost structure and its size to determine whether it could sufficiently use economies of scale and whether the tolling fee to the toller was representative. The Commission should also assess the representativeness of the tolling fee paid by the analogue country producer by comparing it with the tolling fee that Jabil has received for the same service in the Union.

As mentioned above, when construction of the normal value was necessary, the normal value was constructed on the basis of Article 2(3) of the basic Regulation (i.e. cost of production in the country of origin plus SG&A expenses and profit). The alleged cost advantages of Chinese tier 1 companies is not to be taken into consideration in this exercise, as there is no basis for doing so under the basic Regulation. Furthermore, it was not made clear in the relevant submission how the 22 % was quantified by the interested party. The Commission also disagreed that adding SG&A expenses and profit leads to double counting. The price of any raw material being sold or service being provided in a market economy will include part of SG&A expenses and profit of the supplier. Therefore, this claim runs contrary to the way a normal value should be constructed according to Article 2(3) of the basic Regulation.

As mentioned above, the representativeness of domestic sales of the analogue country producer was assessed by the Commission. However, the detailed outcome of this test and the volumes of production and sales cannot be disclosed as they constitute business confidential information of that producer. The costs of production of the modules made under the tolling agreement were composed of the cost of production of the cells used in the modules and the tolling fee. As to further details on the calculation of the constructed normal value of both cells and modules, the Commission noted that this is business confidential information. Furthermore some of it, like cost structure of the tollers is not only confidential but also inaccessible to the Commission as well as the analogue country producer. With regard to this request, it cannot be overstated that the interested party does not question Taiwan as being an appropriate market economy third country. Indeed, in its previous submission, the interested party praised it as such, dismissing USA and India as inappropriate analogue countries. There is no indication, nor argument being put forward by any of the interested parties, that suppliers of a service to an analogue country producer in an appropriate market economy third country are not behaving according to market forces.

These claims were therefore rejected.

The Commission first established the export price on the basis of export prices actually paid or payable by independent customers in the Union or on the basis of resale prices when the product concerned was sold via related importers in the Union.

In order to examine whether export prices to the Union were reliable and given the existence of undertakings, export prices to the Union were analysed in relation to the minimum import price (‘MIP’) of the undertaking. It was in fact necessary to ascertain whether the export price to the Union was set at a certain level mainly because of the MIP established by the undertaking and, therefore, whether it was reliable or not. In this respect, the Commission considered whether, on a weighted average basis at the level of each sampled exporting producer, the export price to the Union was substantially higher than the MIP or not. The Commission also considered how this price related to prices for exports to third countries.

For all sampled exporting producers, the export price to the Union was on average at the level of the MIP. In addition, their export price to the Union was significantly higher than export prices to third countries. Consequently, the export price to the Union was significantly influenced by the undertaking and is therefore unreliable.

Following disclosure, one interested party argued that this finding confirms that the MIP serves as a general price benchmark for the Union market and thus determines the price level of all modules sold in the Union, no matter where the modules were produced.

The Commission did not see this connection. The fact that the sampled exporting producers, because they were subject to the terms of the undertaking, could not sell below the MIP, does not preclude other exporting producers both from the PRC and other third countries from doing so, if economically viable. Furthermore, the MIP can also not serve as a benchmark because it is business-confidential. This claim was therefore rejected.

In the absence of a reliable export price for these Chinese exporting producers, due to the existence of the undertaking in this case, the Commission considered another methodology to establish the export price. Given that the sampled exporting producers were selling solar panels on the world market, the Commission used unit export prices for solar panels sold in the biggest non-EU export markets (Chile, India, Japan and Singapore, depending on the exporting producer) of the sampled exporting producers, where those sales prices did not include trade defence duties. India appeared to have trade defence measures in place, but the Commission was able to use data from the exporting producers, who did not pay those duties, for the reason set out below in recital 60. It then turned out after disclosure that those duties were no longer in force during the RIP (see below recital 86)

If the exporting producers export the product concerned directly to independent customers in the third country, the unit export price was the price actually paid or payable for the product concerned when sold for export to the relevant third country market, in accordance with Article 2(8) of the basic Regulation.

If the exporting producers export the product concerned in the third country through a related company acting as an importer, the unit export price was established on the basis of the price at which the imported product was first resold to independent customers in the relevant third country market, in accordance with Article 2(9) of the basic Regulation. In this case, on the basis of the data supplied by the sampled exporting producers and verified by the European Commission, adjustments to the price were made for all costs incurred between importation and resale, including SG&A expenses (between 0,05 % and 9 %, depending on the data reported by and verified for the company in question), and for profits accruing (between 1 % and 3 %, depending on the data reported by and verified for the company in question). In this case, sales transactions including anti-dumping or countervailing duties were disregarded as they are not a reliable proxy for a price in the absence of measures. Therefore, duty-paid transaction in the United States of America were excluded, as the vast majority of these transactions were done through related importers.

Following disclosure, one interested party argued that, when constructing the export price on the basis of Article 2(9), the Commission wrongly made adjustment for SG&A expenses and profit of related traders in the PRC and in Hong Kong.

The same interested party pointed out that some of the allowances may have been double counted when direct selling expenses were added (as they were already included in the SG&A). The party also pointed out that its level of dumping when compared with its level of undercutting may point to some errors in the calculations.

The Commission accepted these claims. A relevant adjustment on the account of traders in the PRC and in Hong Kong was made under Article 2(10)(i) instead of under Article 2(9) of the basic Regulation. The Commission also adjusted its calculations to avoid double counting of certain allowances. Furthermore, following the comment about dumping and undercutting the Commission discovered a clerical error in its calculations and corrected it. The changes in the calculations have no impact to the conclusion and the proposal. Nevertheless, the party was informed of the changes made after its comments on final disclosure and allowed to comment, if necessary, again.

In order to establish the export price the unit export price established as described above was then multiplied by the quantities sold to the Union during the RIP.

Following disclosure, one interested party questioned the use of third country export prices. According to the party, Chile and Singapore have limited solar installations and represent a minor share of the PRC's total cell and module exports. The same party requested the Commission to provide a breakdown of the export sales to each country in the calculated export price and the weighted average export prices in each of the four export countries considered, namely Chile, India, Japan and Singapore.

As explained above, for each sampled exporting producer the Commission used its biggest export markets. Chile and Singapore do not represent minor shares of the sampled exporting producer(s) in question. Concerning the request for the breakdown of the export sales to each country in the calculated export price and the weighted average export prices in each of the four export countries considered, the Commission notes that this information contains business confidential information. Moreover, this information was obtained from the sampled exporting producers. It was then returned to them with the specific disclosure, so that they could verify whether it was used correctly in the relevant calculation. No comments concerning this issue were received from the sampled exporting producers. The following ranges can be provided in order to illustrate this point: Chile from 12 % to 18 % of the relevant exporting producer(s)' exports; India from 9 % to 15 % of the relevant exporting producer(s)' exports; Japan from 12 % to 22 % of the exporting producer(s)' exports; and Singapore from 40 % to 60 % of the exporting producer(s)' exports.

These claims were therefore rejected.

The Commission compared the normal value with the export prices for solar panels in the sampled exporting producers' biggest non-Union export markets (Chile, India, Japan and Singapore, depending on the sampled exporting producer).

Where justified by the need to ensure a fair comparison, the Commission adjusted the normal value and/or the export price for differences affecting prices and price comparability, in accordance with Article 2(10) of the basic Regulation. Adjustments were made for transport and insurance, (between 0,02 % and 7 %, depending on the data reported by and verified for the company in question) handling, loading and ancillary expenses (between 0 % and 1 %, depending on the data reported by and verified for the company in question), credit cost (between 0,05 % and 0,5 %, depending on the data reported by and verified for the company in question) and bank charges (between 0 % and 0,03 %, depending on the data reported by and verified for the company in question).

For the sampled exporting producers, the Commission compared the weighted average normal value of each type of the like product in the analogue country with the weighted average export price of the corresponding type of the product concerned, in accordance with Article 2(11) and (12) of the basic Regulation.

The dumping margins so established range from 23,5 % to 31,5 %.

As mentioned in recital 48 above, the export price to the Union was significantly influenced by the undertaking and is therefore unreliable. Nevertheless, for the sake of completeness, the Commission compared the weighted average normal value of each type of the like product in the analogue country with the weighted average export price of the corresponding type of the product concerned to the Union. On this basis, dumping margins, expressed as a percentage of the CIF Union frontier price duty unpaid, from 8,9 % to 14,8 % were found in the RIP.

This calculation showed dumping for exporting producers who, during the RIP, were subject to undertakings. It is recalled that the MIP in the undertakings was not based on the dumping margin. Thus the undertakings did not remove the dumping established in the original investigation completely.

Following disclosure, one interested party, who was not a sampled exporting producer, complained about no details of the calculation of the dumping margins were delivered.

The Commission noted that detailed calculations were delivered to and verified by the sampled exporting producers that have provided the data. A disclosure to third parties, that goes beyond the meaningful summary provided in the general disclosure document and in the present Regulation, would violate the applicable provisions on the need to balance business confidentiality and procedural rights.

In 2012, the Union market accounted for up to 60 % of the annual global new installed capacity. Since then its importance has diminished to 14 % in 2015 as indicated in recital 197. Its share is expected to remain important in the future. However, the forecasts for growth of the Union market of annual global new installed capacity are modest in comparison with the rest of the world.

The Chinese producers are still very interested in the Union market. Despite the introduction of the anti-dumping and countervailing measures in 2013, they have maintained a strong position on the Union market. Their share in the modules market dropped from 66 % in 2012 to 41 % in the RIP, whilst their share in the cells market increased from 7 % in 2012 to 16 % in the RIP. This reflects the trend that an increasing number of Union producers limit their activity to the segment of module production, purchasing cells in third countries. They account as Union production of modules. It also reflects the tendency that those exporting producers that also have production facilities in other third countries than the PRC withdraw from the undertaking and are selling to the Union from those third countries.

Furthermore, as demonstrated by the anti-circumvention investigation in 2015, some of the Chinese producers were trying to avoid the measures by circumventing them through Taiwan and Malaysia, the biggest third countries in imports into the Union. Since the imposition of the measures, several undertakings have been withdrawn. The withdrawals were either voluntary or due to breaches or for reasons of impracticability (see for detailed references above footnote 6).

Following disclosure, one interested party argued that, despite the trade defence measures against imports from the PRC, the USA is an attractive market due to incentives and a very high natural sun irradiation level. Furthermore, according to the same party, contrary to the facts set out above, India does not have any trade defence measures against imports from the PRC. Furthermore, India plans to significantly increase its installed capacity.

With regard to the USA, the Commission noted that the party provides no proof that the incentives completely offset the disadvantage posed by the duties to producers from the PRC. Indeed, the Commission did not claim that the US market is unattractive but that the presence of the duties reduced its attractiveness. With regard to India, the proposed duty was not enforced and was allowed to lapse in June 2014. However, even if India is not subject to duties, duties in Canada and the USA are increasing the attractiveness of the Union market, should the measures lapse. It also does not render the Union market unattractive as all the other considerations described above still stand.

The spare capacity of all exporting producers in the PRC cooperating with this investigation is at around 33 %, according to their sampling replies. These companies alone had a spare capacity during the RIP (around 10 GW) capable of satisfying the entire Union market. It is to be noted that the spare capacity of the largest cooperating producers by volume, is much lower, their capacity utilisation ranging from 86 % to 97,8 %.

Following disclosure, one interested party questioned this data arguing that the four largest Chinese manufacturers (Trina, JA Solar, Jinko and Canadian Solar) were operating at full capacity. According to the interested party these four producers account for more than 40 % of the total Chinese exports (about 11,2 GW). Consequently, to reach the 43 % spare capacity in 2016 indicated by the Commission, all smaller producers would have to operate at a mere 20 % capacity utilisation. The interested party requested the Commission to provide details of the calculation of the 43 % average.

The Commission first noted that the spare capacity of all exporting producers in the PRC cooperating with this investigation is estimated at around 33 %, not 43 %. Furthermore the party's calculations outlined above are based on the assumption that JA Solar and Canadian Solar were cooperating with the investigation. These two manufacturers did not cooperate with the investigation. Thus their capacity and capacity utilisation is unknown to the Commission and was not taken into consideration in the calculations above.

The claim was therefore rejected.

Several parties argued that Chinese domestic demand has been increasing in the last few years, reaching 50 % of Chinese solar module production by the first quarter of 2016. China would allegedly have around 20 GW annual new installations per year until 2020. However, as shown above and further developed in recitals 186 to 195, the Chinese exporting producers' excess capacity would be capable of meeting the total global demand in the future, including all rapidly growing markets such as China itself, India, Japan and South America taken together.

Two interested parties questioned the data for the capacity in the PRC and the global consumption used by the Commission. One interested party argued that further capacity will be required to cover the growing demand for solar installations in the PRC and elsewhere in the near future. According to the party the Commission failed to provide any sound arguments that the alleged excess capacity in the PRC would be destined for the Union market.

These claims are addressed in recitals 190-191 below.

In light of the estimated significant spare capacity in the PRC, combined with the attractiveness of the Union market in terms of size and sales price, in particular with regard to the price level of the PRC exports to third countries, and the records of past circumvention practices, the Commission concluded that there is a strong likelihood that the repeal of the anti-dumping measures would result in a significant increase of dumped imports of solar cells and modules from the PRC to the Union.

During the period considered the import volumes of modules from the PRC decreased by 73 %, with a corresponding decrease in market share by 39 %, i.e. from 66 % in 2012 to 41 % during the RIP. However, after the existing measures were imposed in 2013, import volumes of modules decreased by 45 % between 2013 and 2014, while consumption decreased by 31 %.

Regarding cells, the import volumes increased by 65 % during the period considered, which resulted in a much bigger increase in market share, i.e. by 123 % (from 7 % in 2012 to 16 % during the RIP), in the context of a shrinking market. At the same time between 2013 and 2014 the imports of cells increased by 59 %, which resulted in an increase in market share of 10 percentage points. Even though the surge did not continue in the RIP, the level of imports remained much higher during the RIP than in 2012 and 2013.

The Commission established the prices of imports on the basis of Comext and Article 14(6) database.

The average import price from the PRC dropped by 22 % for modules and by 43 % for cells during the period considered. For modules the import price decreased in 2012 and 2013 and then, when the measures entered into force, the price increased by 6,3 % for modules between 2013 and 2014. It slightly decreased again between 2014 and the RIP, i.e. by 1,6 %. Regarding cells, the import price decreased by 43 % during the period considered. It decreased by 30 % between 2012 and 2013 and continued to decrease between 2013 and 2014 when it plummeted further by 19,4 %. However, it slightly increased between 2014 and the RIP, by 1,4 %.

Following disclosure, several parties argued that the export price to the EU should be used to establish undercutting and they calculated that there was no undercutting on this basis. It is true that on the basis of the export prices to the EU there is no undercutting for modules and that the undercutting for cells is very marginal. However, the Commission considered that the lack of undercutting because of compliance with the MIP was not the decisive indicator for the analysis of the current state of the Union industry. As indicated in recitals 170 below, the Commission established that the Union industry continued to suffer injury from the past dumping practices found in the previous investigation and from the circumvention practices found as referred to in recital 4 above and did not have sufficient time to recover.

During the period considered imports from third countries into the Union increased by 30 % for modules. The biggest surge was between 2013 and 2014, when the volumes increased by 48 %. Their market share increased significantly, from 9 % in 2012 to 25 % in the RIP. Again, the biggest change happened between 2013 and 2014 when the market share increased from 13 % to 28 %. Taiwan, Malaysia and Singapore were the biggest exporters after the PRC. It should be noted that imports from Taiwan and Malaysia may have included circumvention (see recital 4 above).

Regarding cells the imports from other countries decreased by 47 % during the period considered. The biggest drop, i.e. by 52 %, occurred between 2013 and 2014, while they slightly increased by 9 % between 2014 and the RIP. This resulted in a decrease in market share from 70 % in 2012 to 51 % in the RIP. The decrease was from 75 % to 48 % between 2013 and 2014 and then slightly increased with three percentage points during the RIP. Regarding cells, Taiwan and Malaysia were the biggest exporters, followed by the PRC and the USA. It should be noted that imports from Taiwan and Malaysia may have included circumvention (see recital 4 above).

The average export prices from third countries for both modules and cells decreased significantly during the period considered, in line with the Chinese and Union prices. For modules they decreased by 21 % and for cells by 45 % during the period considered. Again, those prices may have been influenced by circumvention practices (see recital 4 above).

In accordance with Article 3(5) of the basic Regulation, the Commission examined all economic factors and indices having a bearing on the state of the Union industry during the period considered.

For the injury analysis, the Commission distinguished between macroeconomic and microeconomic injury indicators. The Commission evaluated the macroeconomic indicators on the basis of data obtained from the applicant, cross-checked with the standing replies sent by a number of Union producers at pre-initiation stage and the verified questionnaire replies of the sampled Union producers. The Commission evaluated the microeconomic indicators on the basis of data contained in the questionnaire replies from the sampled Union producers.

The macroeconomic indicators are: production, production capacity, capacity utilisation, sales volume, market share, growth, employment, productivity, magnitude of the dumping margin, and recovery from past dumping.

The microeconomic indicators are: average unit prices, unit cost, labour costs, inventories, profitability, cash flow, investments, return on investments, and ability to raise capital.

Only three cell producers were sampled and two of them are members of EU ProSun. They cooperated in bringing the request, which contained the figures of both producers. Hence, all figures related to microeconomic indicators for cells which can be directly traced to the third company, not member of EU ProSun, are given as ranges to protect the confidentiality of this Union producer who cooperated with the investigation.

In overall terms, the Union industry continued to suffer from injury during the period considered given the short period after the imposition of the original measures and the magnitude of dumping and the level of injury found in the previous investigation. In addition, the circumvention practices found, as indicated in recital 4 above, have also contributed to the continuation of injury. However, from mid-2013 (the provisional measures entered into force on 6 June 2013), and especially during 2014 (the first full year with anti-dumping measures in force) and during the RIP the Union industry started gradually to recover.

Indeed, a number of injury indicators showed positive trends. Regarding modules, the Union industry increased its Union module sales by 2,8 % and, as a result, its market share increased by 9,4 % between 2014 and the RIP. During the same period, the Union industry increased its captive use and Union sales of cells by 6,3 % and maintained the market share of 33 %. In addition, during the same period the Union industry improved its capacity utilisation by 9 percentage points for modules and by 8 percentage points for cells by both increasing production and decreasing existing capacity. It also achieved significant productivity gains, thereby reducing the gap between its sales prices and its average costs of production. Moreover, the previous undercutting from Chinese exports had stopped due to their compliance with the MIP (no undercutting for modules and only marginal undercutting for cells during the RIP). As a result, although the Union industry was still loss making during the RIP, its losses decreased significantly for modules in comparison with 2012 and 2013. However, the losses did not decrease for the Union industry for cells because, as explained in recital 165 above, they were influenced by extraordinary events occurring to two of the sampled producers. By contrast, the third sampled producer had its losses reduced during the RIP and thus followed the same trend as the one observed for modules.

The Union industry also increased its investments both for modules and cells by 41 % and 63 % respectively between 2014 and the RIP.

However, despite the efforts made and all the positive trends that resulted therefrom, the Union industry still did not manage to recover from the past dumping by the Chinese exporters. As already indicated, both cells and modules manufacturers were loss making during the RIP and had negative cash flows and return on investments. In addition, despite the fact that the import volumes of the Chinese exports decreased for modules, their market share was still higher than the one of the Union producers. Regarding cells, the imports increased significantly in the RIP in volume (by 65 %) in comparison with 2012, gaining market share. The Chinese imports of cells exercised indirect pressure also on the module market of captive producers which was thus prevented from growing further. Therefore, Chinese imports continued to enter the Union market with significant volumes and at prices which were below the cost of production of the Union industry.

Following disclosure, several parties claimed that the Commission failed to assess the effects on the Union industry of imports of modules from third countries made in substantial volumes and at prices below the ones of the Chinese exports.

The Commission acknowledged that the impact from module imports from third countries constitutes an important factor for assessing the state of the EU industry, However, such imports were much less important than the Chinese imports — the market share of the former was 25 % (including modules that in reality were Chinese, but fell under circumvention practices), while of the latter was 41 % during the RIP (and, given circumvention, in reality higher). In addition, the market share of the third countries imports decreased by 10 % between 2014 and the RIP, while the share of the Chinese imports increased by 4,9 % during the same period. These two factors show the much stronger impact of the Chinese imports on the Union industry than the imports from third countries. In addition, the prices of modules from third countries were not below the ones of the Chinese exports. As evidenced in Table 4a above, the weighted average price of all imports from third countries was 550 EUR/kW during the RIP, which was higher than the average Chinese export price – 544 EUR/kW. Therefore, these arguments were rejected.

Concerning cells, several parties claimed that the injury has been caused by third countries imports since the profitability decreased during the RIP compared to 2014, which was caused by the decrease of Chinese imports and the simultaneous increase of imports from third countries.

First, as indicated in recital 164, the increase in losses during the RIP for the sampled producers was influenced by extraordinary events which occurred to two of the Union cell producers, while the third (and the biggest) producer had an increase in profitability during the RIP in comparison with 2014. Second, although the imports from China decreased by 3 percentage points between 2014 and the RIP, they increased by 7 percentage points between 2013 and the RIP, despite the entry into force of the measures. Consequently, the imports from China continued to have an important impact on the Union industry. Third, regarding imports from third countries, indeed they increased with 2 percentage points between 2014 and the RIP, but decreased by 24 percentage points between 2013 and the RIP. Therefore, their impact on the Union industry actually decreased in the period after the imposition of the measures. Consequently, the argument that injury has been caused by third country imports was rejected.

Following disclosure, the GOC claimed that some injury indicators improved only in the RIP and not immediately after the imposition of the measures. Hence, there is no clear link between the imposition of the measures and the various positive trends.

The Commission acknowledged that some injury indicators, for example concerning modules the market share, production and capacity utilisation improved only in the RIP and not in 2014. However, given the magnitude of dumping and injury found in the previous investigation, it took some time to reverse the negative trends applicable for the entire industry. This can be explained by the fact that at the time of imposition of the original measures, the Union industry was in the process of consolidation whereby many producers were already in the state of bankruptcy or near bankruptcy, but exited the market only during 2014. This had an important impact on all macro indicators, which also included such companies. It is also worth noting that a number of indicators, such as market share, production, capacity utilisation and production capacity which had negative trend at the level of the entire Union industry, showed positive trends already in 2014 for the sampled Union producers of both modules and cells. Consequently, this claim was rejected.

Several interested parties argued that the Union industry is doing well and has fully recovered from the previous injury. Following disclosure, those parties reiterated these claims. In particular, the figures reported in the financial statements of SolarWorld and Jabil, by far the biggest Union modules producers, allegedly showed that their Union business had been growing in the last few years and they had increased production volumes, capacity, capacity utilisation, export sales and productivity, while the cost of production and stocks have decreased.

The sampled Union producers (including Jabil and SolarWorld) have increased their production volumes, capacity, capacity utilisation, export sales and productivity, while they have decreased their cost of production and stocks in 2014 and in the RIP. However, the claim that the industry has fully recovered from the previous injury contradicts the findings of the investigation which are based on the actual verified data of the sampled Union producers. In particular, many microeconomic indicators are based only on the sales to independent customers in the Union (such as profitability, cash flow, and return on investments). In addition, some of the sampled companies have important production outside the Union, which is not included in the microeconomic indicators. By contrast, the publicly available financial documents focus on all activities of the companies concerned and often provide information on the consolidated accounts of the entire groups. Therefore, conclusions on the economic situation of the Union industry within the meaning of Article 3(5) of the basic Regulation were based not on publicly available financial documents but on the more detailed and verified information regarding the situation in the Union only provided in the investigation. In addition, the conclusions reached on the state of the Union industry were based on data coming from all sampled Union producers and not only from SolarWorld and Jabil. Finally, regarding Jabil, as explained above, this company only assembled modules, but did not assume full contractual responsibility for their sales. This claim was therefore rejected.

Following disclosure, one interested party claimed that the Commission should have taken into account the effects of large-scale investments made by SolarWorld. The latter are said to have negatively affected the company and the industry as a whole, given its important share in the Union industry's output.

Following disclosure, several interested parties questioned the data on capacity in the PRC and the global consumption used by the Commission. They argued that Solar Power Europe was not sufficiently reliable as a source since it takes into account only modules already connected to the grid, while IHS and BNEF provide more accurate outlook as they show the modules purchased for installation.

Following disclosure, one interested party contested the finding that the overcapacity of all types of exporting producers should be taken into account when establishing the spare capacity available in China.

The Commission pointed out that already the capacity available at Tier 1 companies constituted 90 % of the total world consumption for 2015 (estimated at 50,6 GW). In addition, the fact that Tier 2 and Tier 3 companies exported to the EU, irrespective of the small volumes, shows that they are active on the Union market and do not limit their sales only to the Chinese or other markets. Finally, the imports of Tier 2 and Tier 3 companies are expected to increase significantly following the recent withdrawals of the undertaking of predominantly Tier 1 companies. This claim was thus rejected.

Several parties argued that the Union market is no longer attractive for the Chinese producers. They contended that the Chinese production of cells and modules would be rather directed towards the rapidly expanding markets in Asia, such as Japan and India. In addition, Chinese domestic demand has been increasing in the last few years, reaching 50 % of Chinese solar module production by the first quarter of 2016. China would allegedly have around 20 GW installations per year until 2020. Therefore, in the context of an increasing number of solar installations in China, India and other markets in South East Asia, the Chinese solar module production would be primarily destined to satisfy the increasing demand in these markets.

Despite the imposition of the anti-dumping and countervailing measures in 2013, the Chinese exporting producers remain very interested in the Union market which is demonstrated by the fact that they have maintained a strong position on the Union market. As indicated in Section 4.4 above, the imports of modules and cells from China had a market share of 41 % and 16 % respectively during the RIP and have successfully kept (and even increased in the case of cells) their market position in comparison with imports from third countries. The volume and market share of the Chinese imports of modules are much more important than those from third countries; the latter cumulatively accounted for only 25 %. Regarding cells, the market share of third countries accounted for 51 % during the RIP, but this means that they dropped significantly (by 32 %) in comparison with 2013 when they had 75 % market share. In addition, despite the fact that the measures entered into force in 2013, the Chinese imports of cells increased by 77,8 % between 2013 and the RIP. Furthermore, as demonstrated by the anti-circumvention investigation in 2015, some of the Chinese producers were trying to avoid the measures by circumventing them through Taiwan and Malaysia, the biggest third countries in imports.

Following disclosure, several parties contested the finding that the Union market remains an attractive market for the Chinese exporting producers. One of the parties pointed out that the projections for growth of the Union market as a whole should be reduced on the basis of a decline by 18 % of new installations in the Union in the first 9 months of 2016 in comparison with the previous year, thereby leading to a rather pessimistic estimation of 7,1 GW of Union demand in 2016. This would also result into following a pessimistic development in Union demand until 2020. In addition, it contended that the three major Union markets (UK, Germany and France) would lose further attractiveness in the future.

Following disclosure, one party quoted a press release by the Chinese National Energy Administration, in which the latter announced that according to the solar power development plan for the next five years, at least 105 GW of photovoltaic power capacity is planned to be installed by 2020. The party alleged that this would bring further increase in demand in China.

The Commission also analysed whether the Chinese imports would come to the Union at prices lower than the current Union prices should the measures be allowed to lapse.

Almost all exporting producers which sold modules and cells during the RIP from the PRC to the Union had price undertakings and their export prices to the EU were determined by those price undertakings which set a minimum import price. Consequently, such export prices could not be considered a pertinent indicator in order to establish the pricing behaviour of the exporting producers should the measures not be in place.

Therefore, prices to third countries of the sampled exporting producers were used instead. The exports to third countries of the sampled exporting producers were found to undercut the prices of the sampled Union producers on average by 2,2 % for cells, and between 5,6 % and 9,2 % for modules during the RIP. The figures show the average undercutting per exporting producer (the lowest margin among the companies and the highest margin among the companies). For cells there is only one undercutting margin as this is the average undercutting for the sole sampled exporting producer exporting cells to the Union.

One of the parties requested a breakdown of the weighted average price for the four export countries considered (Chile, India, Japan and Singapore) in order to comment on the undercutting findings. For the purposes of the undercutting calculations no account was taken of the weighted average price per third country, but the weighted average price of all four export countries added together, thereby correctly reflecting the quantities and the prices at which those exports have been made. Therefore, this request was rejected.

Following disclosure, one interested party argued that the exporting producers would not have incentives to increase their sales to the Union should the measures be allowed to lapse. The Commission was not convinced by this hypothesis. As demonstrated by the undercutting margins established, the exporting producers could increase their sales volumes to the EU, should the measures be allowed to lapse. Indeed, since their prices in the Union would be lower than the prices of the Union producers, it is reasonable to expect that the Chinese exports would fight for more market share in the Union. This claim was thus rejected.

Therefore, the Commission concluded that the Chinese imports would come to the Union at prices lower than the current Union industry prices and are likely to increase their sales volume and gain market share should the measures be allowed to lapse.

In light of the above, the Commission concluded that there is significant spare capacity in the PRC for both modules and cells. The Union market remains attractive in terms of size and sales price, particularly in comparison with the price level of the PRC exports to third countries, further proven by the records of past circumvention practices. Consequently, the Commission found that there is a strong likelihood that the repeal of the anti-dumping measures would lead to the continuation of dumping resulting in the continuation of injury of the Union industry.

A significant number of Union upstream and downstream companies, either individually or through their associations called on the termination of the measures on the Union interest grounds. They claim that the measures have unforeseen, negative consequences for an overwhelming majority of jobs in the European solar sector. They submitted that the measures cause price increase for solar installations, which has dampening impact on the demand, with all the negative consequences on the downstream and upstream employment. In addition, the wider policy objectives concerning climate change and promotion of renewable energies are also claimed to be negatively affected. They also pointed to changed circumstances since the definitive measures were adopted in December 2013, in particular the move away from support schemes at fixed prices (e.g. feed-in tariffs and feed-in premiums) towards tenders, achievement of grid parity by solar generated power in certain regions of the Union or for certain groups of customers and the ratification by the Union of the UNFCCC Paris Agreement to limit climate change in October 2016. Finally, some parties claimed that there is not enough module manufacturing in the Union to cover the Union demand; that the measures became ineffective due to increase in exports by third countries and that the measures mainly benefitted the exporters in third countries.

Two unrelated importers of modules came forward and provided answers to the questionnaire, which were subsequently verified as set out in recital 28. These unrelated importers also sent several additional submissions setting out their views on why the measures should be terminated. The importers of solar panels suffer from weak profitability. They also provided a list of other importers and wholesalers that went bankrupt or significantly reduced their employment since the registration on Chinese solar modules was imposed in March 2013. Some of these companies pointed to the imposition of measures on Chinese solar modules as the main reason of their insolvency, e.g. Gehrlicher. The unrelated importers claim that the measures increase the price of solar power and depress the demand.

The unrelated importers also submitted that the MIP price undertaking causes disadvantages to them and other downstream companies, active in international markets, such as EPCs (Engineering, Procurement and Construction companies), as it limits their ability to purchase modules from leading Chinese Tier 1 producers. The undertaking does not allow the parallel sales of the product under investigation in the Union and outside of the Union. Therefore, the Chinese exporters cannot deliver modules to unrelated importers who are also active in the markets outside the Union (e.g. in Switzerland or in the US) if they are part of the unrelated importers' wholesale portfolio within the Union. This is claimed to be a significant drawback for these companies' participation as wholesalers and project developers in the globally growing PV market.

The unrelated importers also claimed that the measures in their current form create a significant additional business risk and administrative burden. In their view, additional administrative steps, such as issuance of a certificate by the China Chamber of Commerce for Import and Export of Machinery and Electronic Products (‘CCCME’) and thorough checks by the Union customs authorities extend the overall time from order to delivery from 7 to 11 weeks.

The ban on parallel sales was introduced to avoid compensatory deals that might undermine the undertaking. Additional administrative steps for Chinese imports were introduced to improve the monitoring of the measures and avoid any form of circumvention that might undermine the undertaking.

Following disclosure several parties claimed that the Commission disregarded the interest of the importers. In their view, the measures increased the prices at which solar products could be purchased. Therefore, the importers suffered from decreased demand.

The Commission observed that the key objective of the measures was to re-establish the non-injurious price for the product that was found to be dumped. This logically entails a certain increase in the price of the dumped product. The Chinese dumped products had held a very high market share in the Union before the measures were imposed. Their market share dropped after the measures were imposed. However, as discussed at length in Section 5.3 the Commission found that this price increase caused by restoring the non-injurious price level had only a limited impact on the overall Union demand. Therefore, the Commission concluded that unrelated importers suffered only marginally from the decrease in demand caused by the measures.

More than 140 downstream companies registered as interested parties. The downstream companies are installers of solar panels; EPCs (Engineering, Procurement and Construction), operation & maintenance companies and companies active in project finance. While a majority of companies just registered as interested parties with no follow-up, around 30 companies provided more substantiated submissions opposing the measures. Three EPCs submitted a questionnaire reply. In addition, more than 400 downstream companies from all the Member States were signatories to an open letter calling to terminate the measures.

More than 30 pan-European and national associations representing solar companies sent letters opposing the measures. Among them was the European Association of Electrical Contractors (AIE), claiming to represent the interests of the Union installers at the European level. The most active associations were SPE and SAFE. SAFE is an ad hoc association of 50 German companies whilst SPE claims to be the most representative association of the Solar Power industry in Europe with more than 100 European members, out of which more than 80 support its position on termination of the measures on both solar modules and cells.

EU Pro Sun pointed out that several large solar national associations stayed neutral towards the measures, although some of them are a member of SPE. This is the case in particular for Bundesverband Solarwirtschaft (BSW), Germany (Europe's biggest solar association); British Photovoltaic Association (BPVA); Syndicat des Energies Renouvenables (SER), France; and ANIE Rinnovabili (renewable energy section of Confindustria), Italy. In addition, EU Pro Sun also claims to have support from 150 European installers. However, no installer has openly come forward in support of the measures. The European Trade Union Confederation and IndustriALL European Trade Union sent a joint letter supporting the measures. A German Association of Energy Consumers (Bund der Energieverbraucher) sent a letter in favour of the measures at the level which reflects the cost savings stemming from technological development.

Following disclosure EU ProSun contested the statement that no installer had openly come forward in support of the measures. EU Pro Sun pointed to a letter by 150 installers supporting the measures and a letter signed by two members of the Fachpartnerbeirat der Solar World AG, both sent in October 2016. The Commission noted that the 150 installers asked for anonymity, therefore the Commission continued to consider that they had not openly come forward in support of the measures. The Commission also noted that Fachpartnerbeirat der Solar World AG claimed to represent over 800 installers, but the names of these installers were not provided. The Commission confirmed, however, that the two installers who signed an open latter on behalf of Fachpartnerbeirat der Solar World AG came openly in support of the measures.

The parties opposing the measures on Union interest grounds claimed that downstream companies account for more than 80 % of the employment and value added in the European solar value chain. They pointed to the much higher number of jobs that they create compared to the cell and modules manufacturers in the Union. SPE presented a report prepared by Ernst & Young indicating that the downstream sector employs more than 110 000 people. However, the report did not explain the methodology for the calculation of this high number of employees. Another party indicated that the downstream sector employs around 65 000 people, based on the assumption of roughly 7 FTE/MW/year. However, this party did not substantiate their assumptions, either.

The Commission's investigation also found that most of the jobs involved in the downstream solar sector consist in the installation of modules on residential and commercial rooftops and mounting them on the ground. These jobs usually do not require a significant capital investment specific to solar installation — many tools and machines such as cranes, diggers, drills etc. can be also used for other construction works. While some installers are solely focused on the solar sector and are able to perform very high value added tasks, many installers also perform other jobs in the construction or energy sector or can easily switch to those sectors, without much impact on their revenue. One interested party submitted that many installers were recently switching to the building sector because of low margins in solar installation and increasing margins in the construction of buildings in Germany. Therefore, the survival or economic welfare of many installers is not dependent on the solar sector alone.

The same holds partially true for most of the EPCs that came forward in the investigation. Most of them are also involved in the development of other renewable sources of energy or are large construction companies developing projects in general building sector. The impact of the measures on the revenue and employment of the downstream companies depend on the impact of the measures on demand, which is discussed in the Section 5.3 below.

Following disclosure several parties contested the Commission's consideration that the E&Y report did not explain the methodology for its calculation of employment in the solar downstream sector. Solar Power Europe provided some additional information on the methodology. Even after the additional information was provided, the Commission continued to consider that the methodology was still unclear, in particular neither the report nor the additional information indicated how many people were employed in installation of a representative solar project in each key segment.

The Commission observed that the report by the GWS/DIW/DLR covered the total employment in the whole solar value chain. Consequently, the 31 600 jobs indicated in that report includes also the upstream sector as well as cell and module manufacturing. The German upstream sector is believed to employ several thousand people — Wacker alone claimed to provide around 3 000 jobs. The solar manufacturing equipment and balance of plant equipment manufactures also claimed to employ a few thousand people. These jobs are only partially affected by the sales in the Union as the upstream companies export most of their output outside of the Union.

The Eurobserver report does not make any distinction between upstream, downstream and manufacturing jobs either. In addition, the employment figure in the solar sector in Germany quoted in that report is the same as the one reported in the GWS, DIW, DLR study for 2014. Therefore, the Commission considered that the number of people employed in the downstream sector in the Union that might be allegedly affected by the measures is much less than 120 250 people

Following disclosure several parties disagreed with the Commission's finding that it is easier to switch jobs from the installation of modules to the general construction sector than from manufacturing of modules to other sectors. However, none of the parties provided any precise data what proportion of the jobs in the installation of the panels is carried out by general construction companies (i.e. ground works and landscaping for ground mounted installations and reinforcing the structure of the roof for rooftop installations).

The Commission had already agreed that a significantly larger number of people were employed in the downstream sector than in the manufacturing of modules. It had also observed that many jobs in the downstream sector required specific skills that made them hard to switch to other sectors. It had, however, also concluded that the impact of the measures on demand is the key factor affecting the jobs in the downstream industry. The absolute number of employees and the questions whether it is hard to switch to other sectors or not, is becoming irrelevant in this respect. As discussed extensively in Section 5.3 the Commission found that the measures had only a limited impact on demand for solar module and therefore also a limited impact on the downstream employment.

Upstream operators produce raw materials, such as polysilicon and wafers; manufacturing equipment for cells and modules as well as balance of system components, such as inverters, storage, mounting etc. On the one hand, Solar World, which is also the largest Union wafer manufacturer, supports the measures. In addition, one more Union wafer manufacturer came forward in favour of the measures. On the other hand, eight other Union upstream companies that came forward opposed the measures. However, most of these upstream operators came forward at a later stage or did not provide a questionnaire reply. Only the polysilicon manufacturer — Wacker Chemie AG (‘Wacker’) provided a questionnaire reply and was verified as set out in recital 28 above.

The upstream operators calling to terminate the measures reiterated the argument that the measures depress the demand through increased prices, which negatively affects the whole solar value chain. Several companies claimed that due to depressed demand they suffer from a reduction in turnover, profits, loss of jobs, and inadequate resources to invest in R & D. However, due to the fact that they sent their submissions late and did not reply to the questionnaires, their situation could not be verified. The Commission estimated that the upstream industry may employ several thousand people.

The verified polysilicon producer provides more than 2 000 direct jobs and around 1 000 indirect jobs in the Union. It also has a large direct R & D budget exceeding EUR 17 million related to solar raw material production. Although Wacker's turnover and employment remained stable in the period under investigation, it vehemently opposed the measures claiming that they have a negative impact on the trade relations with the PRC. The PRC is by far the largest producer of solar wafers and cells; therefore the polysilicon manufacturer's turnover and several thousand jobs are dependent on an unrestricted access to the Chinese market, which is declining. Wacker and several other parties claimed that by protecting an inefficient industry — solar cell and module manufacturing, the measures cause serious damage to the industries in which Europe has still a competitive edge.

The association of the German equipment manufacturers (VDMA) sent a letter calling to review the level of the MIP, pointing to the fact that manufacturing costs in the solar sector have been continually decreasing. VDMA indicated that the solar cell and module producers had been following a historic solar learning curve of 21 %. VDMA also claimed that the German photovoltaic equipment manufacturers are the key enabler of such a cost reduction. The equipment manufacturers are estimated to employ several thousand people and are key contributors to R & D in the solar sector.

Several parties claimed that the Commission did not properly weigh the interest of upstream operators against the interest of the Union industry. The Commission reiterated that only one upstream operator — Wacker — had sent a complete reply to the questionnaire and could be verified. The interests of this company were taken into account in the analysis as set out in recital 250. Several other upstream operators came forward only at a very late stage with short submissions. The Commission could not properly verify the impact of the measures on the other upstream operators. In any case, the Commission had found that the measures had only a limited impact on the demand for solar modules and consequently the Union sales and profitability of other upstream operators.

The Commission acknowledged the basic assumption of the unrelated importers, downstream and upstream industries that the lapse of the measures may be beneficial for the turnover and the number of jobs in these industries. It can thus be concluded that prolongation of measures is not in their interest. However, at the same time the Commission found that the impact of the measures and the likely effect of terminating them on these economic operators and new installations in the Union were substantially overstated in most of the submissions by the companies opposing the measures, as analysed in detail in the Section 5.3 below. With respect of the alleged administrative burden arising from the undertaking an interim review on the form of the measures could be initiated to better protect the interests of the unrelated importers and the downstream operators.

Several parties contended that the Commission did not take into account the interest of the consumers. The Commission considered that their interests in lower prices overlap with the interests of downstream users assessed in Section 5.2.3. The Commission therefore did not undertake a separate analysis for consumers.

Following disclosure, SAFE provided a more detailed submission analysing the impact of measures on German electricity consumers. The SAFE study concluded that the removal of the measures could save German electricity consumers approx. EUR 570 million a year under the assumption that modules could be purchased at EUR 0,40 to 0,45/W and that an annual target of 2 500 MW would be achieved. In addition, Wacker contended that the termination of the measures could save consumers EUR 1 billion a year. This figure was calculated by comparing the difference between the MIP and the assumed price of solar modules of EUR 0,42/W in the past and the expected price of EUR 0,32/W in 2017.

The Commission found that the above calculations are based on too simplistic assumptions. First, the MIP was equipped with an adjustment mechanism. Therefore, for most of the period considered the gap between the MIP and an average global module purchase price was not that large. A larger gap between the two opened only in 2016. This gap was closed at least partially by the latest adjustment of the MIP that became effective at the beginning of 2017. In order to avoid a too large difference between the MIP and the global module purchase price in the future and further reduce the impact of the measures on the consumers, the Commission intends to open an interim review on the form and the level of the MIP. It should be noted that it is not possible as part of the present expiry review to amend the level of measures, as this necessitates an interim review.

Second, the prices used by the parties in their calculations relate to multi-silicon modules in the lower quality range. A large proportion of the modules sold in the Union, though, are high-efficiency multi-silicon and mono-silicon modules. Their prices are considerably higher and therefore the gap between the MIP, which applies without distinction to all modules, and the actual purchase price was even smaller.

Therefore, the Commission considered that the measures had only a very limited impact on the finances of the consumers and solar deployment.

Several parties also claimed that the MIP, increasing the price of modules, is responsible for the failure to achieve annual solar deployment targets in Germany. The Commission was not convinced about this mono-causal link. The gap between the MIP and an average selling price was smaller than assumed by the parties, and the measures did not affect significantly the demand. Accordingly, Commission also concluded that the MIP did not have a significant impact on the non-achievement of the solar deployment targets.

Virtually all parties who came forward opposing the measures claimed that the measures increase the price of new solar installations and depress the demand for solar modules, making solar generated energy more expensive to consumers. While some admitted that the decline of solar installations in the Union is also due to the modifications and, in certain Member States, suspension or termination of support schemes, they considered that the measures cause artificially high prices and slow down solar deployment even more. They claimed that recent policy changes such as an introduction of competitive tenders for new solar installations exacerbate the damage caused by higher prices of modules due to the measures in place. In their view, the termination of measures would enable cost reductions for solar power in the Union, benefiting end-consumers, stimulating the demand for solar panels and reducing the cost of clean energy provision as a whole.

• Large utility-scale solar systems (or parks), which are usually mounted on the ground, have the installed capacity exceeding 1 MW, and are usually connected to a high-voltage transmission grid to which they feed the electricity they generate;

• Commercial solar systems (or installations), which are usually mounted on the rooftop of a user's building. The user is a commercial entity (for example a supermarket or a warehouse) and the system is connected to a low-voltage distribution grid. Commercial installations can be used for self-consumption or feeding electricity to the grid;

• Residential solar systems (or installations), which are usually mounted on the rooftops of private houses and have capacity that usually does not exceed 10 kW. Residential installations are generally connected to a low-voltage distribution grid and can be used for self-consumption or feeding electricity to the grid.

The Commission considered that, in the RIP and the preceding years, the demand in all three segments was driven by the support schemes as set out in Section 5.3.2 below. The Commission also considered that, at the end of the RIP and in the year 2016, important changes took place and the demand for the three segments became increasingly driven by separate forces. The demand for large utility scale solar parks is increasingly driven by tenders as set out in Section 5.3.3, and possibly to a very small degree by grid parity. The demand for commercial and residential installations is increasingly driven by the achievement of retail grid parity, both with and absent taxes as set out in Section 5.3.4.

In light of the above arguments, the Commission acknowledged that the solar industry has a steep learning curve and the cost of solar production has been going down. For this reason the undertaking/MIP has been equipped with an adjustment mechanism based on prices quoted by one of the market intelligence agencies, Bloomberg. The prices in a competitive market are supposed to reflect the decreases in the cost of production. Nonetheless, the price index that the Commission has been using changed only marginally. This raises the question — which has not been analysed by the Commission in this investigation — whether the Bloomberg index still adequately reflects the evolution of world market price. As this an expiry review, that cannot change the level or the form of measures, there was also no need to examine that matter further.

The Commission observed that there had been a boom in solar installation demand in the years 2010 to 2013 driven, in certain Member States, by a mismatch between feed-in tariffs set at the level of a fair module price and the overall level of prices driven by unfairly dumped Chinese modules. The decline in demand in 2013 and the following years was an inevitable consequence of an installation boom in the preceding years. The significant decline in solar module consumption/installation had already started in 2012, which coincided with a significant cut in feed-in tariffs in that year in certain Member States. In the years 2014 — RIP, the Union demand was increasing only in the UK, the Member State that had the most attractive support schemes in that time. Therefore, the Commission found that the support schemes were the main driver of demand in the RIP and in the preceding years. It can thus be concluded that the unrelated importers, downstream and upstream industry suffered to a significant extent from the collapse of the Union consumption that is unrelated to the imposition of the measures.

Competitive tenders are one of the market based mechanisms required by these new state aid rules. Under such a mechanism governments put a desired amount of capacity to be installed up for an auction In line with the Union state aid rules, competitive tenders are in principle technology-neutral, but may also be technology-specific, for instance where this is necessary in order to ensure the necessary diversity of energy sources.

The solar parks developers bid for the lowest price they agree to be paid for the energy they are going to inject into the grid over the lifetime of a solar park.

The opponents of the measures claim that under the new price sensitive system of competitive tenders, the MIP will have an increasingly negative impact on solar deployment as it increases the price of modules — the main component of the solar system. According to these parties the tenders for capacity coupled with the termination of measures can yield significant savings. The cheaper the solar systems become, the more appetite the governments may have to build solar installations as the cost savings are going to be directly reflected in the final price of electricity. In addition, decreasing the price of the solar system will make it easier for Member States to achieve the targets for renewable energy deployment.

The parties opposing the measures submitted that in some jurisdictions, in particular in the United Kingdom, solar was bound to compete against other renewable forms of energy, such as on-shore wind. As no trade measures apply to wind energy, solar was unable to compete against wind and very few solar projects went ahead through those auctions. Some parties claimed that Germany also considers introducing technology neutral tenders, and that such neutral tenders exacerbate the disadvantage of higher prices caused by the measures as it makes solar power lose in tenders against other renewable technologies in particular wind.

The parties supporting the measures submitted that the tenders and overall installation targets were introduced exactly for the purpose to enable the governments to control the level of solar deployment and to avoid installation boom and bust cycles that happened before.

The Commission limited its analysis of support schemes for large utility-scale solar systems to Germany, France and the United Kingdom. That choice is justified by the fact that they accounted for roughly 80 % of annual new solar installations in the Union during the RIP. It was thus appropriate to look at the developments in these three Member States as a proxy for the situation in the Union as a whole.

The analysis is based on information submitted by interested parties, information collected by the Commission during the investigation and information submitted by Member States to the Commission for the purpose of State aid control.

Germany and France have already set their targets for the three years to come. France plans 2 annual tender rounds of 500 MW each for solar energy in 2017-2019. That makes for a total annual new installation of 1 000 MW. Germany plans to have tenders for an annual volume for solar energy of 600 MW from 2017 at least until 2020.

The United Kingdom, on the contrary, does not run technology-specific tenders for solar energy. In February 2015 a multi-technology tender took place in which solar modules were in competition with other technologies for the award of contracts for difference (‘CfD’). No new tenders are planned in the United Kingdom for the time being, as the government takes the view that utility-scale on-shore wind and solar energy can compete with other sources of electricity without support schemes.

For Germany and France, the measures cannot have any impact on utility-scale demand for solar modules, as solar energy has ‘reserved’ tenders, and the capacity is fixed. The only difference is a marginally higher price for the end consumers that have to cover the costs of the tender either through taxes or charges.

The Commission first analysed the situation for wholesale grid parity, and then for retail grid parity.

Wholesale grid parity. The opponents of the measures submitted that, if the measures are removed, large solar installations could achieve wholesale grid parity in the sunniest parts of the Union, such as Spain. According to a business plan presented by one of the members of SAFE, they could achieve the LCOE of EUR 3,8 cents/kWh in the Spanish region of Cadiz if they purchased modules at EUR 0,35/W. This party believed that they could get a module price of EUR 0,35/W from Tier 1 Chinese manufactures for a large quantity transaction in the absence of duties, for the project delivered at the beginning of 2017. They claimed that at such low LCOE the solar energy not only achieves grid parity with other conventional sources of energy, but there is no other production method of electricity which could beat it in the very near future. This cost advantage coupled with other assets of South European countries i.e. good grid connections, stable political and economic environment and strong and liquid currency gives them a unique opportunity to become new European leaders in energy production. According to these parties, the measures prevent this from happening.

The Commission observed that the Spanish region of Cadiz has one of the best solar radiations in the Union (the most hours of sun in a year), which maximises the production of energy per module. Given much lesser solar radiation in most parts of the Union, it has to be seen when achieving wholesale grid parity can be replicated in other markets, even though the British government seems to assume that that is soon the case. The Commission also observed that there are large differences in the wholesale price in the different Member States, which means that wholesale grid parity comes at different prices depending on the Member State in question.

Therefore, the Commission took the view that in the near future wholesale grid parity will not be achieved on a wide-spread basis in the Union, even in the absence of the measures.

Following disclosure, several parties disagreed with the Commission's conclusion that wholesale grid parity will not be achieved on a wide-spread basis in the Union in the near future, even if the trade measures expire. These parties pointed to a 6-page report by the Becquerel Institute which examined the potential for wholesale grid parity in eleven Member States. This report was financed by the three parties to the investigation opposing the measures.

Becquerel assumed CAPEX costs at 0,726 EUR/W without differentiating between Member States. BNEF assumed much higher CAPEX and differentiated by Member States: France 0,99 EUR/W; Germany 0,9-0,94 EUR/W, Italy 0,76-0,99 EUR/W and the UK 0,9-0,94 EUR/W. The difference could be explained to some extent by the fact that Becquerel assumed the module prices free of the measures. However, so much lower CAPEX appears unlikely, especially given that, according to BNEF, CAPEX is anyway lower in the Union than in other countries where trade defence measures do not apply, e.g. Turkey 1,04 EUR/W and the United Arab Emirates 1,14 EUR/W.

Finally, the Becquerel report concludes that: ‘PV electricity could become competitive in several of Europe's largest markets by 2019, and in most European countries […] within the five coming years’. This means that most of the Member States, including Germany, are expected to achieve grid parity only when the measures are set to expire in 2019, even in the disputed case that the assumptions and findings of the Becquerel report were proven to be correct by then.

Retail grid parity. The opponents of the measures also submitted that the rooftop solar installations have already achieved retail grid parity, i.e. they became economically viable, even without subsidies, in the Member States where the retail electricity prices are high, such as Germany. One of the parties illustrated the point with an example of a commercial centre. It has a large rooftop area for the panels to be installed and it needs energy during daytime while the main activity goes on. So its energy demand coincides with the peak solar production. In Germany the price they need to pay for each kWh is currently around EUR 20 cents/kWh, while the LCOE of a rooftop solar installation is around EUR 10 cents/kWh. Therefore, installing the panels can yield significant savings on electricity charges, especially to customers who use a lot of energy during the day. This has not been the case in the original investigation when solar-generated electricity always required subsidies to be viable. In the situation when the subsidies are no longer relevant, it is the final module price that influences a customer's decision to install solar system and currently this price is increased by the measures.

The Commission investigated in more detail the market for retail grid parity in Germany, because parties have extensively commented on it during the investigation.

It is necessary to distinguish between installations that are for self-consumption, and installations that feed electricity into the grid.

The Commission concurred with the view that achieving the grid parity by solar energy is a very desirable development as it helps combatting climate change and reducing consumers' electricity bill. It found that removing the measures would reduce the price of solar panels, hence increasing the number of locations at which retail grid parity can be achieved.

For installations feeding into the grid, demand is mainly determined by support schemes, which may continue to be used without carrying out tenders for installations of up to 1 MW. At a given level of support, demand is slightly higher if measures are removed. At the same time, for budgetary reasons, most Member States have significantly reduced support, so that this demand is shrinking irrespective of the measures.

Following an in-depth analysis the Commission found that with regard to the impact of measures on demand, it is important to distinguish different sectors and different types of demand.

Demand from support schemes. Where support schemes are technology-specific, measures do not have an impact on demand. Where support schemes are technology-neutral, measures reduce the chances of solar energy of winning. However, even for those tenders, the Commission observes that solar energy has won a significant market share, which shows that even with the measures, it is capable of competing for utility-scale projects.

Wholesale grid parity. Wholesale grid parity at this stage and in the near future can only be achieved, if at all, at ideal locations, even if prices for solar panels are reduced because measures are lifted. Again, the additional demand to be expected in the absence of the measures is low, and in addition depends on the level of the wholesale price in the Member State in question.

Following disclosure, several parties disagreed with the finding that the measures had only a limited impact on the demand for modules. They contended that investments are driven by the expected rate of return and therefore the lower the price of solar module, the higher the return on investment and the likelihood of realisation.

The Commission recalled that it had always agreed that the price has an impact on demand. Nonetheless, in its view other factors exert currently much more significant impact on demand than a relatively mild increase in the price of modules caused by the MIP. These factors are on the one hand regulatory uncertainty, but on the other hand, the intentional actions by the government to control the level of solar deployment. These actions include setting of overall annual installation targets, introduction of tenders for capacity, and imposition of the taxes on self-consumption.

The Commission also recognised that in some instances, such as technology neutral tenders, the MIP may have a more significant impact on the solar demand. At the same time, the Commission observed that without the MIP removing the effect of dumping solar energy would have an unfair competitive advantage over other renewable sources of energy. The Commission also observed that solar energy was relatively successful in the technology neutral tenders even with the MIP that merely restores the level-playing field.

Therefore, these arguments were rejected and the Commission continued to consider that the measures have only a very limited impact on the solar demand in Europe.

The parties calling to terminate the measures on cells claimed that there is no availability of cells manufactured in the Union on the Union market. There are only few cells manufacturers in the Union and almost all of them use their cells in-house and sell only very small quantities to third parties. The small volumes sold are usually of inferior quality. On the basis of its investigation, the Commission agreed that the cells sold by the Union industry on the Union market satisfy less than 5 % of the demand of the non-vertically integrated Union module manufactures.

The parties further claimed that even if more Union cells were sold on the open market, the Union capacity and production of cells would by far be insufficient to satisfy the total Union demand for cells and even less adequate to satisfy the total demand for modules. The Commission found that cell production in the Union was 1 270 MW in the RIP, which means that it could cover 37 % of the total Union demand for cells, which was estimated at 3 409 MW. The Union cell manufacturing capacity could cover roughly 18 % of the total Union demand for modules — roughly 7 200 MW in the RIP.

Several parties submitted that the measures on cells benefit only one company — Solar World — and put most other non-vertically integrated module manufacturers at a competitive disadvantage in relation to that company. Solar World accounted for more than 70 % of all the cell production in the Union during the RIP. Solar World does not sell any cells on the open market but the ones which it considers not fulfilling its high standards. These parties submitted that the only way for them to purchase cells are the imports from third countries. If the measures on cells are not terminated, they would remain at a competitive disadvantage in relation to the dominant manufacturer of cells in the Union.

As noted above, the Commission found that the Union production of cells accounted for 35 % of the demand of cells in the Union and this ratio increased from 23 % in 2012. This means that a substantial proportion of the Union demand can be covered by Union manufactured cells irrespective if these are captive sales or not. The Commission also recalled that the average cost of cells manufacturing in the Union was above an average Chinese and Taiwanese contract selling price (as reported by PV Insights). This means that non-vertically integrated Union module producers could buy the cells from outside of the Union at similar prices or even lower prices. This conclusion holds even taking into account the fact that the cells produced in the Union by vertically integrated producers are of high efficiency and high quality and hence are more expensive. Therefore, the measures do not confer a competitive advantage to vertically integrated cell and module manufacturers over the non-vertically integrated ones.

This period of overheated demand for cells also partially coincided with the Commission's anti-circumvention investigation of Chinese cells and modules via Malaysia and Taiwan. The registration of cells and modules, effective from the date of the initiation of the anti-circumvention case, i.e. May 2015, caused some additional supply difficulties to the Union module manufactures. The module manufacturers experienced a period of uncertainty as they had not known if their Taiwanese and Malaysian suppliers were genuine producers and would get an exemption from the duties. Eventually, in February 2016, more than 20 Malaysian and Taiwanese cell manufacturers who cooperated in that investigation were found to be genuine producers. The temporary supply uncertainty came to an end removing the supply uncertainty.

The Commission also found that the claim that the measures on cells protect only one company — Solar World — is unfounded. As set out in recital 340, five more cell manufactures expressly support the continuation of the measures on cells. The Commission is aware of more than 10 cells manufacturers in the Union. The key reason why a single producer currently accounts for more than 70 % of all the cells produced in the Union is that many other cell manufactures exited the market as they could not withstand unfair competition from Chinese dumped products. While most of these manufactures went out of the market, Solar World took over one of the largest Union cell manufacturers in 2014. This manufacturer was exiting the market and if the takeover had not happened more than 500 highly skilled workers would have lost their jobs. Solar World claimed that without the measures in place not only it would have been unable to take over the other company and save its employment, but its own cell manufacturing facilities would have been bankrupt by now.

Therefore, the non-vertically integrated Union module assemblers face neither supply shortages nor a competitive disadvantage with regard to the vertically integrated ones. Consequently, the Union consumers of modules are not negatively affected by the fact that the Union production of cells can cover only a relatively small proportion of the modules consumption.

The parties calling to terminate the measures on cells claimed that they impose significant additional business risk, working capital and administrative burden on non-vertically integrated module manufacturers. This would stem from adhering to the complex procedures of the undertaking. Some module manufactures submitted that the situation exacerbated after the anti-circumvention investigation started against Taiwan and Malaysia. The non-vertically integrated manufacturers are subject to strict checks of documents while the goods are customs cleared even if they import from companies that got the exemption. They complained that, in some instances, simple procedural deficiencies in the documentation, such as a lack of a stamp or signature at the right place, caused several weeks of delay. The Chinese cells that are imported under the undertaking are subject to similar rigid customs checks and burdensome administrative procedures. Therefore, these producers submitted that nearly 100 % of the imports of cells in the Union are subject to diligent scrutiny by customs authorities resulting in shipment delays, additional administrative work and increased working capital requirements. This makes them suffer from an additional competitive disadvantage not only compared to the manufactures in the rest of the world, but also to the vertically-integrated leading European module manufacturer.

The Commission recalled that it had taken additional steps to improve the monitoring of the measures and avoid any form of circumvention and cross-compensation that might undermine the undertaking. Other solutions that could effectively protect the Union industry and prevent the circumvention of Chinese modules can be addressed in an interim review. Enhanced checks by the Member States' customs authorities are necessary to effectively protect the Union industry from an influx of products at unfair prices.

Several parties calling to terminate measures on cells claimed that they increase the cost of the key component of solar modules and consequently increase the price of new solar installations, which depresses the demand for solar modules. Some parties submitted that outside of the RIP, in the second half of 2016, the average global contract selling prices fell significantly below the MIP, putting an additional strain on the non-vertically integrated Union module manufacturers.

The Commission found that an average global contract selling price for cells as reported by PVInsights was close to the MIP for most of the time the measures were in place. Therefore the measures did not significantly increase the price of the key component of modules, if the global average contract selling price is used as a benchmark.

The Commission found the global contract selling prices fell significantly in the second half of 2016 as the solar sector went through a boom and bust cycle as set out in recital 356. In the fourth quarter of 2016, the cell prices stabilised and even started going up again, which is predictable for boom and bust cycles. Nonetheless, taking into account the learning curve effect of the solar sector, it is likely that the new longer term price of solar cells will set below the level that prevailed before the recent boom and bust cycle. As noted above in recital 265 the price index, which the Commission has used to adjust the MIP, did not fully reflect such decreases in the cost of solar cells and modules production throughout most of 2016. A better way to reflect the evolution of the solar industry's learning curve in the level of the measures can be examined in an interim review.

The impact of the measures on solar demand in the Union was analysed in-depth in Section 5.3 above. The Commission found that several other factors influence the demand for modules much more than the measures. As the cells are the key component of the modules these findings also hold for the cells.

The MIP followed closely the global prices for cells denominated in the euro from December 2013, when the measures were imposed, until September 2015, when the RIP ended, as reported by PVInsights. The Commission also noted that in 2016, outside of the period considered, the MIP temporarily decoupled from the global purchase prices. Both the quoted article and the data provided indicated solar price developments outside of the period considered. Therefore, the Commission reiterated that the measures on cells had only a very limited impact on the non-integrated module manufactures.