Daqo New Energy Corp. (DQ) Q3 2016 Earnings Report, Transcript and Summary

Good morning and welcome to the Daqo New Energy Third Quarter 2016 Results Conference Call. All participants will be in listen-only mode. [Operator Instructions] After today's presentation, there will be an opportunity to ask questions. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Kevin He, Investor Relations Officer for Daqo New Energy. Please go ahead.

Hello, everyone. I am Kevin He, the Investor Relations of Daqo New Energy. Thank you for joining our conference call today. Daqo New Energy just issued its financial results for the third quarter of 2016, which can be found on our website at www.dqsolar.com. To facilitate today's conference call we also have prepared a PPT presentation for your reference. Today, attending the conference call, we have Dr. Gongda Yao, our Chief Executive Officer; and Mr. Ming Yang, our Chief Financial Officer. The call today will feature an update from Dr. Yao on market and operations, and then Mr. Yang will discuss the company's financial performance for the third quarter of 2016. After that we will open the floor to Q&A from the audience. Before we begin the formal remarks, I would like to remind you that certain statements on today's call, including expected future operational and financial performance and industry growth, are forward-looking statements that are made under the Safe Harbor provision of the U.S. Private Securities Litigation Reform Act of 1995. These statements involve inherent risks and uncertainties. A number of factors could cause actual results to differ materially from those contained in any forward-looking statement. Further information regarding these and other risks is included in the reports or documents we have filed with or furnished to the Securities and Exchange Commission. These statements only reflect our current and the preliminary view as of today and maybe subject to change. Our ability to achieve these projections is subject to risks and uncertainties. All information provided in today's call is as of today and we undertake no duty to update such information except as required under applicable law. Also during the call, we will occasionally reference monetary amounts in U.S. dollar terms. Please keep in mind that our functional currency is the Chinese RMB. We offer these translations into U.S. dollars solely for the convenience for the audience. Without further ado, I now turn the call over to our CEO, Dr. Yao.

Hello, everyone, and thank you for joining our conference call today. During the third quarter of 2016, successfully set new records in terms of high production volume and a cost. I would like to thank our entire team for their hard work and dedication for their excellent results. Our polysilicon production for the third quarter reached a record high of 3,636 metric tons which has surpassed our annual nameplate capacity of 12,150 metric tons. Through technical and process improvements, we also made a significant progress and reduce our production costs even through achieving our lowest cost ever costs structure with $8.66 per kilo, a reduction of 8% from $9.43 per kilo in the second quarter. Cash cost reached at $6.88 per kilo, a reduction of 7% from $7.42 in the second quarter. And at the same time, we were able to improve our production quality silicon and produced more semiconductor grade per second give towards the high efficiency mono-crystal solar segment. In the third quarter of 2016, downstream solar end-market and polysilicon market experienced a significant volatility and due to the slowdown in China end-market demand and installation activities after Chinese assist adjustment and end of during 2016, the market has saw slower level of demand towards end of the third quarter of 2016. Our third quarter ASP was $15.64 per kilo compared to second quarter ASP of $17.24 per kilo. As a result, several polysilicon manufacturers in China partially shut down their capacities due to weak polysilicon pricing. The resulting reduction in polysilicon supply has had to stabilize the market and pave the way for price recovery. Due to the delay in the delivery of critical specialty components, the starting date of annual maintenance of our Xinjiang polysilicon facility was postponed to October 3 from the second half of September as originally planned. We continue -- we conducted more than 500 maintenance items including special projects that are expected to improve our manufacturing efficiency that you have filled. The maintenance was completed successfully and we initiated gradual strata of production on October 24. In early November, production resumed successfully and we reached the full capacity utilization, overall, we estimated that annual maintenance has impacted approximately 800 metric ton to 900 metric tons, our first quarter polysilicon production volume. In the third quarter of 2016, due to the market volatility and declining conditions for the solar wafer segment reduced our wafer production utilization rate. As a result, the company sold 14.4 million pieces of wafer in the third quarter compared to 25 million pieces in the second quarter of 2016. Market condition improved significantly in the middle of October with the increase in customer demand and the marketed wafer pricing in response to improved market conditions, the company ramped up wafer production utilization during that time and reached full production in November. The solar market have begun to recover in early October with recovery in customer demand and strong order maintenance due to the limits of the channel inventories. Polysilicon pricing has recovered particularly well with increasing order from our broadly based customers. By November pricing has recovered back to previous levels, see in July of this year. We continue to see a tighter supply policy environment but in China and a much higher demand and a strong order of our high quality polysilicon than we are able to supply. For the solar wafer segment, wafer pricing also recovered significantly during the same time period with strong order across multiple customers. During early November we are also seeing downstream product pricing for solar cell and solar module going up as a result of strong end market demand which bode us well to stable to rising cost stricken and the wafer pricing for the coming months. Now let me provide outlook for the fourth quarter of 2016. As a result of our annual maintenance, the company expect to sell approximately 2,200 metric ton to 2,300 metric ton polysilicon to external customers during the fourth quarter of 2016. The external sales guidance exclude the shipment of polysilicon to be used internally by our solar wafer facility which utilized polysilicon for its wafer manufacturing operation. Wafer sales volume is expected to approximately 20 million to 21 million pieces for the fourth quarter of 2016. Now I will turn the call to our CFO, Ming Yang, for financial updates.

Thank you, Dr. Yao, and good day, everyone. Thank you for attending our call today. Now, I will provide the financial updates for the third quarter of 2016. Revenues were $54.3 million compared to $71 million in the second quarter of 2016 and $46.6 million in the third quarter of 2015. Revenues from polysilicon sales to external customers were $44.4 million compared to $50.5 million in the second quarter of 2016 and $34.1 million in the third quarter of 2015. External sales volume was 2,838 metric ton in the third quarter compared to 2,931 metric ton in the second quarter of 2016. ASP of polysilicon was $15.64 per kilogram in the third quarter of 2016, compared to $17.24 per kilogram in the second quarter of 2016. The decrease in polysilicon revenues from the second quarter was primarily due to lower polysilicon ASP and lower external sales volume. Revenues from wafer sales were at $9.9 million, compared to $20.5 million in the second quarter and $12.5 million in the third quarter of 2015. Sales volume was 14.4 million pieces, compared to 25 million pieces in the second quarter of 2016. The decreasing wafer revenue from the second quarter of 2016 was primarily due to lower wafer AST and lower sales volume. Gross profit was $20.1 million, compared to $29.4 million in the second quarter of 2016 and $8.6 million in the third quarter of 2015. Non-GAAP gross profits which excludes costs related to the non-operational polysilicon operation in Chongqing was approximately $21.6 million, compared to $31.2 million in the second quarter of 2016 and $10.9 million in the third quarter of 2015. Gross margin was 37.1%, compared to 41.4% in the second quarter of 2016 and 18.4% in the third quarter of 2015. In the third quarter of 2016, total cost related to the non-operational Chongqing polysilicon plant including depreciation were $1.5 million, compared to $1.8 million in the second quarter of 2016 and $2.3 million in the third quarter of 2015. Excluding such cost, the non-GAAP gross margin was approximately 39.9%, compared to 43.9% in the second quarter of 2016 and 23.4% in the third quarter of 2015. Selling, general, and administrative expenses were $4.9 million, compared to $3.7 million in the second quarter of 2016 and $2.9 million in the third quarter of 2015. The increase in selling general and administrative expenses as compared to the second quarter of 2016 was primarily due to higher relocation and moving expenses related to the company's relocation of its idle polysilicon manufacturing equipment from Congqing to Xinjiang, which were $1.7 million during the third quarter of 2016 compared to $0.6 million during the second quarter of 2016, as sure as higher non-cash share-based compensation expenses. We conducted the majority of our equipment relocation during the third quarter and we expect those cost to normalize in the fourth quarter. Research and development expenses were approximately $1 million, compared to $0.1 million in the second quarter of 2016 and $0.1 million in the third quarter of 2015. During the quarter, we conducted R&D and technology upgrade projects to the rich silicon crystal growth and off-cash recovery, which we expect to both for the improved product quality and reduced production cost over the coming months. Other operating income was $2.2 million, compared to $0.6 million in the second quarter of 2016 and $1.1 million in the third quarter of 2015. Other operating income mainly consists of unrestricted cash incentive that the company received from local government authorities, the amount of which varied from period to period. Income from operations was $16.4 million, compared to $26.1 million in the second quarter of 2016 and $6.7 million in the third quarter of 2015. Operating margin was 30.3% compared to 36.8% in the second quarter of 2016 and 14.3% in the third quarter of 2015. Interest expenses were $3.1 million, compared to $3.5 million in the second quarter of 2016 and $3.1 million in the third quarter of 2015. EBITDA was $25 million, compared to $34.7 million in the second quarter of 2016 and $15 million in the third quarter of 2015. EBITDA margin was 46%, compared to 48.9% in the second quarter of 2016 and 32.1% in the third quarter of 2015. Net income attributable to Daqo New Energy shareholders was $11.2 million, compared to $19.8 million in the second quarter of 2016 and $3.1 million in the third quarter of 2015. Earnings per basic ADS were $1.07, compared to $1.90 in the second quarter of 2016 and $0.29 in the third quarter of 2015. As of September 30, 2016, the company had $29.2 million in cash and cash equivalents and restricted cash, compared to $42.9 million as of June 30, 2016 and $68.7 million until September 30, 2016. As of September 30, 2016, accounts receivable balance was $4.6 million, compared to $10.1 million as of June 30, 2016 and $15.4 million as of September 30, 2015. As of September 30, 2016, the notes receivable balance was $17 million, compared to $14.8 million as of June 30, 2016 and $16.5 million as of September 30, 2015. As of September 30, 2016, total bank borrowings were $227.6 million, of which $129 million were long-term borrowings compared to total borrowings of $227.9 million including $118.4 million long-term borrowings as of June 30, 2016; and total borrowings of $259.1 million including $144 million long-term borrowings as of September 30, 2015. As of September 30, 2016, notes payable balance was $14.4 million, compared to $26.1 million as of June 30, 2016 and $32.2 million as of September 30, 2015. For the nine months ended September 30, 2016, net cash provided by operating activities was $70.9 million, compared to $65.6 million in the same period of 2015. For the nine months ended September 30, 2016, net cash used to investing activities was $51 million, compared to $82.7 million in the same period of 2015. For the nine months ended September 30, 2016, net cash used in financing activities was $12.3 million, compared to net cash provided by financing activities of $38.1 million in the same period of 2015. And that concludes the official part of our presentation. Now let's have the Q&A session.

Thank you. We will now begin the question-and-answer session. [Operator Instructions] Today's first question comes from Philip Shen of ROTH Capital Partners. Please go ahead.

Good morning, and good evening to you. Great job on your cost structure. It seems like in terms of maintenance that your polysilicon facilities -- typically from what I recall has about five to 15 days of maintenance, but it looks like this quarter or this time around, it seems like it may have been as much as 20 days, given the volume impact of 800 and 900 metric tons. Could you help us understand why maintenance may be longer or may have been longer this time around versus in prior years?

Yes. We have two things -- one is the maintenance content, the first time for hydrochlorination facility have been running for about around 15-18 months since installed about 15 months ago. That's the one thing. Maintenance, even looking just exactly the same, compared to what used to be last the maintenance before 6,000 metric ton facility, this time it's a 12,000 metric ton, first in the new facility for hydrochlorination. In addition to that, we also did all the preparation for the Phase 3A, which is additional 6,000 metric ton. According to original schedule, we will fully ramp up by middle of next year and we're trying to do our best to try to put it in. So this maintenance period time, all that did is foundation for that purpose. You're right; originally, we're try to manage it for 6,000 metric ton time. Our maintenance will never be more than two weeks, like 15 days. This time, it will be the longer. The situation is the sizes are getting bigger and that because maintenance also shifted from -- typically we are doing the April. Right now I'm doing the September so the accountant of maintenance is much more. Actually, items are much more. We mentioned briefly in the presentation and we said about 500 items have been maintained for that period of time. We also want to reduce that because this is the first time since we're running 12,000 metric ton with hydrochlorination. I'm pretty sure next time we would do a little bit earlier, so probably around Summer time in around the third quarter and that we're trying to manage it to 15 days to 20 days period. It should be less next time. This is first time again. I don't know if I answered your question, but that's roughly the idea.

You definitely answered the question and gave us lots of color on that. Can you help us understand how much more you may have spent on maintenance in Q3 versus -- did you spend what you had expected to spend and if you can give a rough number as to what it was?

Yes. For financial side, a major impact is we lost the metric ton during the shutdown time. We said it's around 800-900 metric ton polysilicon impact. The financial money side, not much compared with the timing we lost. It's obviously more important at this moment. I'm going to put a color on that. We also installed some new facilities to further improve our quality. We're expecting our quality of the polysilicon will be improved after this is fully stabilized, running. We'll probably see results in the Q4-Q1 next whole year, actually, after some installations, some technical improvement on this maintenance. So it's a little more complicated compared with the history of all the other maintenance we've done. So same time as I've said, we also learned a lot and we're trying to do -- even next time we're expecting the maintenance -- company's similarly complicated because that were our plan to getting bigger and bigger, but we have the confidence. Next time, we can reduce the time compared with this one.

Great. Shifting to your cost structure. You guys, before your maintenance activity achieved really attractive levels, given that success, can you update us on what you expect your cash cost and all-in cost to be for 2017? Previously, you talked about a range between $9 and $9.50 per kilogram for the year, but you're already below that level, so what is the updated range we should expect for 2017?

So for 2017, we have confidence with the result of third quarter we can achieve the cost of $8.50 or below. That's no problem. I still know that Q4 cost were starting to get higher because we have maintenance, down time for October, but at 2017, we have confidence that we can achieve $8.50 below.

So the full production on our new capacity. We think we could hit that target. Yes.

Also we have a lot of rooms to improve that because of volume and also the new equipment we put in at a new -- those quality improvement facility would be starting and running right now.

So with better quality polysilicon as well.

Okay. Shifting to the wafer side of the business. You had reduced utilization or at least [indiscernible] shipments in Q3 given the slowdown in general. Can you share what the margin was on wafers in Q3 and then what you expect for margins in Q4?

Margins for Q3 for the wafer business was in the low single-digit percentage. It's below 5% and then for Q4, based on the current pricing for polysilicon, we're doing product transfer based on market pricing for poly and for wafer. It's around 5% or so, slightly [ph].

Okay, good. Ming, you guys often have a great view on what's going on in the market place. You're at the beginning of the value chain. Can you talk to us about fundamentally what happened with the dropping demand in Q3 and then is the Q4 recovery, simply just a return of normalized demand or can you just talk about in general the market conditions that you see for Q4 and then in Q1 and then talk about what you expect ASPs to be like for polysilicon in those periods as well? Thanks.

Let me start and then Ming can add on. The Q3 as you know, even first half of the Q3 is very successful before. At the time we announced the Q2 results, the polysilicon price is very solid and the second half and the polysilicon become weaker. Although the wafer pricing seems like had weakening, actually ahead of polysilicon. So the wafer business, normally like Ming mentioned, Q3, Q4 is more challenged actually because Q4, when the pricing recurring, wafer pricing becoming slow than polysilicon also. There's lagging ahead of the decline and the lagging and recurring. So the wafer business is much challenging than polysilicon. On the polysilicon side, we have high confidence because the current pricing we see in China is already $16.50. At currency price, the pricing for polysilicon is already high than average of our Q3 number. We're expecting this price and probably we're stable. Maybe we review that, if this price is probably were stabilized through the remaining of this year as we have confidence to live with that. Because our major customers, we are in the middle of processing to do some contracts for 2017 and then even 2018 and demanding and ordering for remaining of this year is very strong. We don't see any problem. As we mentioned in the presentation also, we checked some pricing in China for sales in module, also study recurring. Polysilicon pricing recovery is much faster than anything else in the volume trend. We're in the right business and also in the relative shortage, also we get a channel check that imports from some other region like Taiwan has the crackdown due to the efforts of Chinese government for -- this is long standing problem. We usually see like [indiscernible] import more than 1,000 metric tons, sometimes couple thousand metric tons, but very recently go down to 100 or those levels. It shows that effect of the imports has been down and the internal demand in pick-up, that's why driving the polysilicon price is back to $16 to $17 level, which is relatively high, compared even with our Q3 average.

Okay, great. Go ahead, Ming.

I think you remember in terms of demand, we had a very strong rush demand for downstream solar products like selling modules all the way through the end of June to meet China FIT cut-off data. And really that the market, particularly for China had a more or less break in terms of lease activities right after that time, so we did see a slowdown of orders for China and that actually had an impact in terms of global demand because as pricing came down -- and actually a lot of what we think are orders for the second half of the year actually, customers would delay their orders until maybe September or even later, just to see where pricing would stabilize, and we did see pricing bottom in the second half of September and that's when really for the entire value chain, the demand was really weak at that time. But I think what's important for the polysilicon sector is that when pricing went down, we did see a lot of plants shut down. We've seen China and globally and even in some of the very large manufacturers momentarily reduced the utilization to maybe half of their production volume. We actually didn't see any inventory buildup. There's actually still very low levels of inventory in the market for polysilicon. And then when the orders came back really in the early October and came back really strongly in the second half of October, that we saw a very significant V-shaped price recovery. Polysilicon pricing we are today talking about, we're looking at $16 or higher polysilicon cost. We're really back to where pricing was in the early Q3, late Q2 time period. I think this is the self-correcting mechanism in the polysilicon sector. And by even supporting that, I think is the way we are just in the recent weeks, maybe two weeks module pricing also is going up in terms of talking with our customers. That is really supporting solo wafer and polysilicon price recovery. That's what we're seeing right now.

That's really interesting. The module pricing might be going up. We haven't seen that yet and haven't heard that yet, but we'll certainly keep an eye out for that. Sorry, I cut you off. Did you guys have something else you wanted to say?

Yes. One is the China market really came back pretty strongly and also we think that the Indian market also is really starting up because the Indian installation is really in the Q4 and then Q1 time frame, normally due to their seasonality. So we're seeing a lot of orders for India, for Chinese modules as well, or Chinese solo products in general.

Okay. I'll ask one more, but it's a two-part question and then I'll jump back in the queue. Can you update us on the timing of the Phase 3B expansion? The potential timing there, and how does the reduction of China's five-year plan target from 150 to 110 gigawatts or more impact your decision to expand potentially your capacity to 25,000 metric tons? And then how did the Trump win potentially impact that sinking? And then shifting to another point on that. To what degree would you consider not expanding and then focusing your efforts to -- what could you do with that cash flow generation in other words in 2017? Would you consider a share repurchase, or debt reduction, or special dividend? And then finally if there's an update on the Third Board, that would be helpful. Thanks.

Okay. The position first, Phase 3B is dependent on several things. One is the macro economy for the solar industry like you said. The five years plan seems the total installation remaining is reduction of forecast for last few years. That's the one consideration. That's to give you overall picture of the demanding. By China, demanding is the most important country for the installation, so we'll see that. Probably we're awaiting for probably Q1 of next year. We see the China market, what's going on, especially for industry response to governments. As you know, Chinese government's planning and actually happening, there's a gap. Sometimes could be crazy. We will see. That's the first effect. Second effect is as you know, even the demanding may be modest or even reduced, but in polysilicon industry, we believe there a lot of capacity is not a factor because if future, the pricing maintaining and the $14, $15, $16 range. So we still believe we are very competitive if we can continue to do sell cost. The sum capacity were shifting from high cost area to the low cost companies. We believe we still have some chance if we were favored as to expanding because we're expanding in the very low cost to future. Even based on whatever we're talking about, $8.50, if we add a more capacity, we were half the economy and maybe the potential for reduced sale cost improvement. That's the positive side. But overall, as you also mentioned, macro situation is as China, we can see probably clearly because China is representing our downstream of 80% of manufacture capacity in worldwide, so we were able to see what's going on for how the industry responds to Chinese government's planning forecast. We will not make a decision probably most likely in this year. We probably will look at in the first quarter. Now, if we, despite not going to expand for Phase 3B, most likely priority like you mentioned the several things, the one thing is obviously the debt percentage of the company's sale is relatively higher. Although it's not very high within the industry, still, we want to be lower. Probably that's the first priority and the other option is that we also could consider that reduced debt probably the first priority compared with others.

Great. And what's the update on the Third Board if there is one?

The Third Board, we think that China gets a lot of the intention for investors to invest in the company. There are several reasons. One is a very attractive value for the company because we achieved the low cost and very stable, very committed to the operation excellence and also very stable expansion so far, also significantly, last year we have 100% capacity increasing. In current turnaround we have additional 50% increasing capacity results at that percentage is huge increasing. That's amazing. Cash flow for the company for next few years, if we built to not have an addition of projects, that would be significantly improved for last year, even before that. So in very good position. Secondly, the Xinjiang location is very favorable for government to encourage to develop the industry and also attract the investment. Those things were investor point of view. Now as you said, this is linked to our expansion project. Right now, we have a lot of meetings with those investors. Probably the company will make a decision by Q1 of 2017 regarding Phase 3B. And those are things that connected each, so we were considered as a whole.

Great. I'll pass it on. Thank you very much.

Thank you.

Thank you, Phil.

[Operator Instructions] Our next question today comes from Sheng Zhong of Morgan Stanley. Please go ahead.

Thanks for taking my question. Just a few follow-ups. On the polysilicon business, you mentioned there you are shifting your supply to more mono-wafer makers. Last quarter your percentage is around $0.15. Any update on this?

Yes. Q3, actually, we increased a number of customers for mono-wafer application. We shipped around the 20%. We have a plan in the 2017, we will probably increase to 30% at least for next year. Partial of the project, maintenance project with that in October is related to this wafer quality improvement. We've done that. We needed some time to check the results. Probably we'll have need additional one-man [ph] running. So by the end of this month, probably we're able to see and we have confidence we can achieve that in more than 30% of polysilicon. We should be able to utilize it by mono-wafer customers in future.

Yes, that's great. So another follow-up is on the wafer cost -- actually with -- we know that the third quarter could be a low-demand quarter, so that your assess volume is not so high, but in the fourth quarter, it looks like it's still not in fully transition. Can you please add more color on where the demand of the market is now on the wafer side?

Yes. Fourth quarter, the demand on the wafer is pretty good, actually. The reason we are shifting the guidance is lower than the Q2 is because we restart our manufacture. Because we reduced our utilization until middle of October, we're starting. So first quarter, we only count down like two and-a-half months' production because at one moment, the wafer pricing going to unbelievably lower end is not rational pricing. We are now participating at this kind of pushing the pricing down. You stop production wafers, we do some [indiscernible] instead in the end of Q3. So we're restarting the manufacture actually by second half of October. That's what we're counting for, the total shipment of wafer for Q4 would be reduced compared with Q2. But our shipment as you can tell, our shipment in the Q4 are away from that numbers. They are higher than Q3.

Yes. We see -- actually I am trying to -- I would say more about the demand side because actually in the industry, I think there is a lot of entry on I think most of value chain including poly, wafer's sale in the module and I think from your balance sheet, I think your -- your level also higher than second quarter. Can you maybe share with us what your inventory for poly and wafer at end of the quarter?

Right now, the inventory for poly was probably two to three times more than what we would like it to be for that wafer and then what is a normal low level that we were used to having. Now that the inventory level has returned to for example, the level that we were seeing in the Q2 of 2016. They had completely normalized and same for wafer.

Okay.

I think in the second half of September, the volume move in the entire market was very low, actually [indiscernible]. You only represent maybe one half weeks of inventory, something like that.

The half weeks for poly?

Yes, into the finished goods. Somewhere…

In the Q3, you see that inventory of a poly is such higher. There are several reasons. One of the reason is because in October, we know that we were shut down. Actually it's about three weeks for maintenance and during that time, we need to continue to provide some poly to our key customers and so we keep the shipment. It's not sale immediately to some wafers because as you know, our poly were divided by several categories, especially for basically at least of two as a mono-wafer versus multi-wafers. We kept some inventory especially for mono-wafer customers. They are taking the maturity of schedule based on by that time because that time, remember, the price is going unreasonably lower, so people don't want to buy until the hope like October price in the lower. So we keep that material for those customers. As a result, actually in October, we sell those materials at prices higher.

Higher price because prices went up.

So it's not like we cannot sell. If we want to sell a raw price to multi-wafer customers, I think we can sell into September. That's the situation I tell you. It's not like we cannot sell it to our demand. As you know, demand and the pricing is linked. If you sell at very low price, you always can sell it. But we didn't choose in that way from our point of view. And secondly, at some point is the mono-wafer, we need the shipping according to schedule. They took a certain amount every few weeks. That's what we're doing also. It's really a bit more complicated than one product like mono-wafer away from customers you can ship to anywhere any time. So the inventory control is a little bit easier. Right now, it's actually complicated. But if once we stabilized with the percentage for mono-wafer application using our total polysilicon production, I think that that will be stabilized to the future.

That's very helpful. Thank you.

Thank you.

And our next question today comes from Wyan Yang [ph] of CICC. Please go ahead.

Hi. Thank you for taking my questions, I just have two questions. Why is the product cost -- I see cost reduction in 3Q; and I want to know how much is due to exchange rate and R&D depreciation? How much is due to maybe power tariff cost or what is the reason behind the cost reduction for 3Q?

Okay, very good question. If you just focus the full year, the three quarters from Q1 to Q3, we did -- Ming did a basic calculation around 10% reduction for cost in dollar basis -- in U.S. dollar basis. If you come to RMB, it's about 8%, so roughly like only 2% for currency basis. Now I'm saying is, as September 30 -- after September 30, the exchange rates changed, so we expect in Q4, more -- maybe the U.S. dollar based across impact or more -- obviously we don't know that until December but in the three quarters, majority cost structure reduction is based on RMB based reduction, it's about 8% which is very significant because during that time our hydro transition process is getting more stable and they are efficient to get the maximum rise [ph]. So that's not surprising, so it's from our point of view. If you look at that, even Q3 we put a number there -- majority is because the volume is increasing, we achieved more than we have our nameplate capacity. So the depreciation cost is ahead of cash cost reduction, it's like 8% to 7%, so that's the basic. To answer your question the majority 2% is from reductions from the exchange rate change and the 8% is the structure -- cost structure investment for RMB basis.

So just a follow-up Dr. Yao's comment, so for example in terms of electricity usage, so a power turf rate didn't change but in terms of utilization of electricity -- amount just be per kilogram, that's reduced by more than 7% in terms of kilowatt hour that's been used. So I think that's a significant saving that we achieved to manage the production cost for the electricity usage. And we also reduced used the amount that we used for silicon metal as well.

Very impressive cost reduction and I just want to make sure that in 3Q you produced more than you assumed and I want to know whether you increased your inventory to have a better delivery in 4Q?

Two things; one, so you're talking about produce and the shipment is because I think the shipment is -- we still have internal use of polysilicon. So...

It increased a lot.

Inventory or talking about?

No, I mean that in tough times, if I divided internal use by the wafer, you would be like maintain a stable level but in fact you will number -- the gap become much larger because you lower your wafer shipment which means that you lower your internal use but make sure you produce more than you consume.

Right. So from previous question mentioned by -- also we answered that, the inventory in the Q3 is particularly higher than Q2. And also I'm answering you, the shipment is not only the inventory, it's also the factors on internal usage. So internal usage of cost inventory by end of Q3 and plus the shipment should be equal to production, so that's my answer. Yes, you're right, the inventory by Q3 is slightly higher and some is because the market condition, we don't want to sell in the very low price because in a very short period time the prices hit the very low position, so we believe that's the very temporary and which is true and proving we are right. So we sold all this inventory in October during our shut down for maintenance time and right now our maintenance is becoming normal -- the inventory I mean, inventory has become normal.

Okay. So the inventory is policy turn or wafer?

That time it's both and the wafer price as you know, that high test is included VAT text about $6 -- RMB6 and RMB6.5, and the low point is below RMB4. So basically a lot of people sell under the cost, so we are not competing with those suppliers. So we kept some wafers and actually we sell in October time when the price is much higher than in September. So again, we choose the right thing to do and we are not competing irrationally with other competitors to follow low price bidding for the contract. So -- so you will see the norm with we expecting the Q4 inventory was normal.

Yes. So if we see a normal inventory that means in 4Q which means that we should see higher delivery in terms of poly or wafer because actually your volume in 3Q and inventories you saw in October and you thought the inventory should prevent to maybe 2Q. So we should spend more in 4Q, that does it mean?

Yes, you're right. So relatively if we want to back inventory for wafer and the parts we can normal -- we're shipping more than we produced in Q4. Yes, you're right.

Okay. So which part do you sell more polysilicon or wafer; you sell more wafer or you sell more polysilicon?

Both. Okay, so it's -- off the polysilicon volume, it's much bigger compared with wafer manufacturer capacity-wise but both will sell more because our -- if you want to achieve low inventory, we have to sell more materials than we produced in Q4. Yes, both for wafer and polysilicon, that's our target for Q4, our planning.

Okay. So under this conditions, so would see guide up of 2,000 to 2,300 polysilicon [ph] under this condition because actually we inventory more, we have more inventories in the quarter so the shipment in 4Q should be higher than production?

I mean the range of our shipments, right?

Yes.

Because we're counting for example, about 2,000 -- because we still include our internal policy usage for wafer manufacturing.

And then Q4 is impacted by our annual maintenance. So we have…

I know, I know. So let me -- in other words, so how much in terms of policy can -- how much you have surplus inventory at September? How much in third-party used store?

With third-party to draw 200 metric tons of polysilicon inventory. [Cross talks].

Okay. Thank you.

Around four days, less than four days production. Because we only count in last few days as -- maybe we cannot acreage to like -- minimum would be two days because we cannot always shipping out quick enough. But we try to maintain minimum inventory by Q4.

Okay, thank you. That's all my questions.

This concludes our question-and-answer session. I would like to turn the call back over to the management team for any closing remarks.

Thank you again for joining the call today. Should you have any further query, please feel free to contact us. Thank you. Bye, bye.