Why Canadian Solar is the Best Solar Stock

by: Jack Yetiv posted on: February 04, 2008 | about stocks: AKNS / CSIQ / FSLR / JASO / LDK / SOLF / SPWR / STP / TSL / WFR / YGE    

In a recent article, I compared Canadian Solar (CSIQ) to First Solar (FSLR), Sunpower (SPWR) and Suntech Power (STP). I concluded that among these solars, CSIQ offers a far better risk-reward ratio than the other three stocks. In response to the article, a couple of posters asked if I would expand my comparison to other solar stocks, and I said I would. Here is the outcome of that analysis.

Before presenting the comparison, I want to discuss my assumptions, biases and methodology:

1) My biggest “bias” is that all things being equal, a lower PE (or, more accurately, PEG) among comparable companies wins this contest. Obviously, since things are not “all equal,” some adjustments must be made, but nevertheless, PEG is the primary driver of this comparison. If that is not a belief that you hold, this article won’t have much meaning to you.

2) This comparison is based more on facts that on projections, with one exception. I have calculated PEs against projected 2008 earnings obtained from either company guidance or the midpoint of analyst estimates. This was essential because in the solar space, a fast-growing-company’s trailing PE is not a meaningful number. I further thought it was reasonable to use forward (2008) PE because we are already into 2008, and therefore, visibility should be reasonable, at least for most of this year. Next, somewhat lowering the inherent error in using projections is the fact that I took the midpoint of analysts’ estimates, which should increase the reliability of the estimate. Next, even if analysts’ estimates turn out to be overoptimistic, that error is likely to impact all these companies fairly similarly, thus not altering the conclusion reached here. Finally, even a 20% change in the earnings estimate does not really alter the conclusion here, further reinforcing this analysis.

3) In my analysis, I put very little—if any—value in projections for 2009, or, perish the thought, 2010 or beyond. To me, given the range of possibilities on both macroeconomic (What will we have? Deep recession? Mild one? Soft landing? Business as usual?) and microeconomic (Will there be an oversupply of solar panels in 2009? Will global incentives increase, stay level or decrease?) fronts, NOBODY can reliably predict what the earnings will be for these companies in the 2009 calendar year (which basically ends two years from now).

It is enough of a stretch to guess where earnings will be in the fourth quarter of this year. Indeed, to highlight the unreliability of 2009 “estimates” (read “speculative guesses,”) one only needs look at the range of estimates for 2008 for some of these companies. For example, one analyst thinks Trina Solar (TSL) will earn $1.65 in 2008, while another believes TSL will earn $3.49 — more than double the lower value. Other companies also have a large spread, although I don’t think any company’s range of analyst estimates is as wide as TSL’s.

4) I have not included any information regarding analyst upgrades, downgrades, target prices, or recommendations, for several reasons. First, this is meant to be an objective analysis—and as can be seen from the spread in analyst estimates, the amount of subjectivity in analyst opinions is quite large. Second, I have for months disagreed with most analyst upgrades of several of the sexy names in the solar space, pushing their valuations—before 2007 was even over—to 50X 2009 EPS (I am, of course, thinking of FSLR, but a few other companies have been similarly hyped, without just cause, in my opinion—see my other article comparing CSIQ to these companies). To me, using 50X multipliers against EPS two years down the line is witchcraft, not financial analysis. Third, analysts’ opinions tend to be even worse at inflection points in the economy—and regardless of whether one believes we’re in a bear market or not, I think everyone agrees we are at an inflection point. Therefore, I provide these numbers as a springboard for further discussion and analysis by visitors to this site.

5) This is not meant to be a commentary on the solar space in general because that would take an article all by itself. Briefly, though, I believe that solar will do very well going forward, and that selected companies in this space will do even better. I believe that demand for PV panels will exceed supply over the next several years because I believe grid parity will be reached in stages and that the initial stages will come sooner than most people think and generate demand that is probably not yet appreciated. (I do believe ASPs will come down, but margin compression will not be severe because feedstock costs will drop and because the fabs will become more efficient.) In addition, I personally believe that the true cost of conventional power (“dirty electricity,” I like to call it) will start being recognized more and more, which will give renewable technologies (and especially solar) a continuing financial boost. Finally, I also believe that in the next year or two, total incentives around the world will increase collectively, and that incentives in the US will be renewed, if not expanded. Finally, I believe that within 5 years, incentives will largely be unnecessary because many locales will have already reached grid parity by then, which will, of course, constitute its own incentive.However, this article concludes that some of the companies in the solar space are such compelling values that even if they were not in a fast-growing space (that I believe solar will be), they would still constitute a bargain.

Having reviewed my assumptions, I want to present the spreadsheet I created to compare 11 solar companies. Due to space limitations, I only show a few columns of information.

The first column shows market cap, ranging from $200 million for AKNS to $16.5 billion for WFR. The second column has two pieces of information—the first number shows 2007 sales followed by earnings in 2007 (or, more accurately, estimated sales and earnings since the 4th quarter hasn’t yet been reported by most companies). The third and fourth columns show 2008 estimated sales and earnings, respectively. The fifth column shows closing price on 1-25-08, and the 6th and 7th columns calculate forward P/E and forward P/S (1-25 price against 2008 numbers). Although I initially input these companies in random order, I then sorted them in order of PE.

Although I present AKNS on this spreadsheet because people have asked me to analyze it, this company does not belong in a comparison of companies based on PE because it is not expected to have any earnings in 2008. This is not meant to be a negative comment against AKNS—since I have not carefully evaluated the company nor its prospects, I cannot really comment on it. I would say, however, that because it isn’t expected to have earnings this year, and because even its sales are quite small (about $35 million in 2007), it is a much more speculative play than the top-rated companies on this spreadsheet.

As is obvious from this spreadsheet, Canadian Solar (CSIQ) stands alone, in several respects.

First, it clearly has the best PE of the group. At 11, CSIQ’s PE is ridiculously low for a company that will nearly triple its revenues (from $290 million in 2007 to $700 million in 2008), and grow its earnings 6-fold from 26 cents to $1.63.

Second, CSIQ’s revenues growth rate just discussed is also the fastest of this whole group (JASO’s expected revenue growth rate is very close, as is its income growth rate, but JASO’s PE, at 29, is nearly triple that of CSIQ).

Finally, CSIQ’s market cap is actually less than its 2008 expected sales—it is the only company trading at a forward P/S ratio under unity (in fact, substantially under unity against 2008 income, and about at unity against its expected 4Q07 revenue, annualized).

TSL also has an attractive forward PE, but the variability in 2008 earnings estimates ($1.65 to $3.49) gives me pause. If 2008 income ends up being closer to the low end of the range, the PE climbs to 15 or 20. And TSL’s P/S ratio (at 1.3) is almost double CSIQ’s. Since TSL does not offer any obvious advantages over CSIQ, and has a somewhat higher, and less reliable, PE ratio, it seems CSIQ is the better play.

If someone is aware of some significant advantage to TSL, or someone can make a good argument that 2008 earnings will be closer to $3.49/sh, it might be the better play given the closeness of their PEs.WFR comes next, with a PE of 17 and a rather daunting PS of 6.7 (almost 10 times CSIQ’s PS).

As far as these metrics are concerned, WFR does not offer any advantages, and is indeed, at a significant disadvantage, but WFR is, in some respects, the leader in this space (size, length of time in business), and one can legitimately argue that the risk with WFR would be substantially less than with CSIQ. I have not adequately studied WFR to give a valid opinion on this issue, but I feel it would be worthy of further comment from participants in this site.

The rest of the companies have PEs of around 20 and up, and PSs of 1.3 and up, and even if these companies have certain advantages, they would have to be substantial advantages for me to pay double (or more) for them than for CSIQ. If readers of this article can offer documentable reasons why the rest of the companies are worth more—even though that value is not reflected in either their sales or earnings growth rate from 2007 to 2008—I would be very interested in knowing those reasons.

But most of the justifications I have read are based on speculation, much of it going out to 2009 and beyond, which to me is no basis for investing. Obviously, the folks who bought FSLR when it was trading at $283 believed the sales pitches that this company was worth 50X 2009 EPS, but that was no more true than the pitches during the dot-com era that companies with minimal earnings and great hopes were worth PEs in the hundreds.But if belief or hope is one’s basis for evaluating a stock, then, as I stated at the beginning of this article, the fact-based, PE-driven approach I took in this article would not be of interest to that person.

Finally, when I identify a very fast-growing company with low PE, I always wonder why the market is pricing it so low, so I look at other metrics—e.g. price-to-book and debt-to-equity ratios. Both of these metrics are also quite positive for CSIQ—with the price-to-book ratio at about 4, and LT debt-to-equity ratio of about 8%, with total debt less than 50% of equity. Both metrics are likely to be even more improved this quarter when earnings are expected to be higher than any other quarter.

In summary, with a PE of 11, PS under unity, sales that are expected to grow from $290 million in 2007 to $700 million in 2008, and a decent balance sheet, CSIQ seems to me to be the most compelling value in the solar space.

Disclosure: The author owns a long position in CSIQ but has no position (long or short) in the other stocks mentioned in this article.

This article has 39 comments! Add yours below...

Jack Yetiv

About this author:
Become a Contributor Submit an Article

This article has 39 comments:

  •  
    MSA
    Feb 04 11:36 AM
    This analysis, bravely so, is 100% dependent on future revenue and earnings estimates. For CSIQ, it's history of revenue and earnings growth is poor relative to its estimates. Based on its "underperformance... in the past, its low to negative manufacturing margins, and increasing wafer costs, the probability of CSIQ continuting to underperform, in comparison to its overly optimistic estimates, is more likely than CSIQ suddenly becoming a competetive, high margin force, in a landscape of much better capitalized and lower cost companies that have proven manufacturing margins and better control of raw material sourcing.

    Simply put, CSIQ is a run of the mill "wanna-be" solar player, that is not capitalized sufficiently to expand rapidly and maintain sourcing of polysilicon at competetive costs CSIQ is likely to continue be a second tier solar play due to its higher cost/low margin manufacturing. Its most beneficial outcome, long term, would be to be acquired by a larger, more capitalized player in the industry. And in a sector that will see consolidation due to excessive and disparate capacity plans, and the need to reach grid parity through cost reductions, the major players will be valued primarily on their ability to lower costs, with the second tier players being squeezed out of the picture due to the inability to compete at large scale. With the prime success driver being low cost at large scale, and world markets likely to be very "selective" of who gets rewarded for cost control and real growth, CSIQ looks to be on the outside looking in.

    When viewing analyses it is important to ask yourself if what is being said is true, then why is the market valuing CSIQ as a low quality player. And the reason is if we "cherry-pick"... our estimates and numbers, we can make a forward looking case that is pretty darn rosy; but when the whole picture is viewed, that being past performance + market trepidation + future hopes, a more complete picture emerges, that is not quite so rosy.

    Good luck to all.
    AckerAnalytics
  •  
    Proud
    Canadian
    Feb 04 07:51 PM
    Here in Canada, you should bet in RIMM, instead of CFIQ, believe me !
  •  
    Jack Yetiv
    Feb 04 09:33 PM
    A few responses to AckerAnalytics:

    1) I acknowledged the potential "error" in using forward PE, but that error applies to all of thes companies. Surely, in coming up with their 2008 estimates--which I averaged in order to minimize the erro--the analysts are aware of what you are saying above, and yet they are still collectively estimating $1.63.

    2) Here are CSIQ's revenues in the first 3 quarters of 2007 (all in millions): 17, 60, and 97. Projections for the fourth quarter are about 110-120 million, which seems believable given the first three quarters. CSIQ has guided to 650-750 million in 2008, and analyst Adam Hinckley thinks this guidance is conservative. These numbers certainly seem to dispute your argument that CSIQ has failed to execute.

    3) AS to its supposed lack of capitalization--CSIQ's LT debt is only 8% of equity, and it appears well enough capitalized to go from $290 million in sales in 2007 to $650 million (low end of guidance for 2008). Which other solar company is going to grow faster in 2008, and what is your basis for saying so?

    4) As to silicon sourcing, here is a direct quote from the third quarter earnings report: "The company believes that it has contractually secured 90% of its silicon or cell requirements to support module production of 200-220 MW in 2008. The company continues to evaluate new technologies, including the use of metallurgical silicon (UMG) products, which, if successful, would have the potential to increase total shipments by 30-40 MW in 2008." 35 MW should translate to an extra $120 million in sales.

    5) Finally, you don't say which company you prefer, and why. That would be truly much more helpful than simply criticizing CSIQ without reference to any data we can look at.

    Jack Yetiv
  •  
    MSA
    Feb 05 11:18 AM
    The revenue and earnings estimates for CSIQ are nearly 4x the projected growth rate for the solar sector as a whole. With that, CSIQ must "gain" or collect market share from the big boys to the tune of 4 x their rate of growth. And this is to be done on margins that to this point are negative to marginally postive. So if you believe that CSIQ, which is an "assembler" of solar modules - CSIQ does not make solar cells, they source the cells from the likes of STP, YGE, SPWR, etc. and assemble/integrate the cells into solar panels and stand-alone products - much like a contract fab assembles integrated circuits, you may wish to ask yourself how is CSIQ going to gain all that market share and increase its dismal margins, while it must source solar cells from outside the company.

    Further, when CSIQ says "The company believes that it has contractually secured 90% of its silicon or cell requirements to support module production of 200-220 MW in 2008", it is important to understand what they are saying. "Believes" is a very uncertain word when speaking of contracts. What is in those contracts that makes them say "we believe" as opposed to "we have contractually secured....". And then, to go on and cite an unproven technology/means of reducing polysilicon usage as their "back-up" plan, the whole claim of sufficient polysilicon sources sounds very shaky.

    Moreover, if you read the latest earnings report, almost all of their sales in the North America region were for sales of portions of their polysilicon supply. So what they are doing is, because of their low module assembly margins, they have used the rise in polysilicon costs to sell polysilicon inventory to boost revenue and earnings in lieu of making solar module products. Clever means of revenue/earnings recognition I might add, but I was of the thought that CSIQ is a solar module manufacturer, not a polysilicon trader.

    Regarding analyst's estimates, it is very risky to take any analyst's estimates at face value, and when an analyst has numbers out on a second tier player that are 4 x the industry average, those numbers need more than three quarters of operating history to have a foundation for such optimistic forward projections. I would love to see Adam Hinckley's product breakdown projections for 2008, to see how he, or any one else could arrive at revenue growth of 2.4 x 2007, and earnings growth of 6.3 x 2007. These numbers are so out of line with the industry, there must be something that CSIQ is doing that no one else in the industry is doing, or perhaps, the numbers are just not justifiable. Maybe CSIQ is selling bizillions of their solar car battery charges?; or they have a secret stash of thousands of tonnes of polysilicon that they will sell to boost sales and earnings. Short of that, I just don't agree with the excessive growth projections, and frankly, I value my opinion as much as any "paid analysts" and more when I see things that are just darn wacky.

    Good luck to all,
    AckerAnalytics
  •  
    Faraz Yusufi
    Feb 05 05:09 PM
    comment for Jack Yetiv: Jack what you have here is a good objective analysis based on what you could get from the balance sheet, it does make you wonder though if something as speculative as the solar energy sector warrants more of Philip Fishers 'Scuttlebutt' approach, like how about digging deeper in to the quality/commitment of the management team and the over all sentiment about CSIQ amongst its competitors / suppliers etc.

    AckerAnalytics: you still haven't named any other stock that you think is worthy of a second look from investors, with out that, your comment here seems more like a rant from someone who has a huge short position in CSIQ, you seem to have done your homework on solar so why don't you come out clean and just tell us if there is any player that looks attractive at this time or is it that you think its all just a house of cards.

  •  
    ecoman12453
    Feb 05 06:25 PM
    I agree that CSIQ is cheaper than TSL on a P/E valuation basis, but TSL deserves a premium multiple to CSIQ because they are going to be perhaps the most vertically integrated solar maker, producing modules fro polysilicon all internally controlled.

    Smaller less integrated companies run the risk of being killed in a ASP pricing war or not having enough poly silicon
  •  
    Dr.H.S.Sarka
    r
    Feb 06 01:09 AM
    We are interested for marketing the canadian solar in India. If you are interested Kindly contact me through my email: harisadhansarkar@yahoo...
    Thanks,
    Dr.Sarkar.
  •  
    Dr.H.S.Sarka
    r
    Feb 06 01:41 AM
    o.k., thanks.
  •  
    David White
    Feb 06 09:26 AM
    The avg. earnings estimate for SOLF is $1.30 (TDAmeritrade) for 2008. This would give it a 2008 earnings P/E of 12.5. A lot different than your figure. Also LDK has a PEG of 0.24 (Yahoo Finance). It is hard to beat that. I think some of your data are faulty. I do tend to agree that CSIQ is a good stock though.
  •  
    zawy
    Feb 06 10:25 AM
    The core technology at the 5-year level for FSLR is the best of all these. They will keep a $1/Wp profit for 3 to 5 years while their price to utilities reduces to nearly 1/2 current price which is already 1/2 the price per Wp of the others you have listed. The 5-year year yoy growth of FSLR is 167% per year and and 180% per year for the past 2 years. Their "scheduled production" (underestimate) is 1 Billion watts to utilities in 2010. $1.5 B Wp sold in 2010 is reasonable. At a current market cap of $15 B (error in your sheet), that's a PE of 10. By listing FSLR last, it shows your methodolgy is pretty blind, as you explained at the top.
  •  
    zawy
    Feb 06 10:40 AM
    Sorry, I was posting from memory and there's a missing divide by 4 in my post. PE in 2009 is 68 based on 5-year historical growth which is very close to guidance. PE less than 10 time frame is in 2012 assuming dollar does not increase against the Euro by more than 50%. Current yoy growth is 167%, not 180%.
  •  
    Plasticman
    Feb 06 11:51 AM
    I would love to invest in the "right" solar company. I say "right" because I believe there are going to be major winners and losers depending on the cost of their technology. Right now the company with the lowest cost technology, Nanosolar is not public.

    Now to the problem. I can't find a source for accurate numbers. For instance if I pull up quotes on E-trade for CSIQ, and SOLF I see EPS of $-0.44 and $0.3252 respectively. However if I go to the earnings screen it shows quarterly numbers that add up to $-0.42 and $2.47 respectively. Both numbers for CSIQ are very different than the article above, and the quarterly numbers for SOLF don't add up to the trailing 12. If I go to Yahoo Finance for SOLF it shows EPS of $.36 and shows the same quarterly numbers as E-trade (in analyst estimates screen) for the 1st, 2nd, and 3rd quarters. So what is going on?
  •  
    Plasticman
    Feb 06 11:58 AM
    What happened to Conergy (ceyhf) today? They are down over 20% with no news. Nanosolar had announced a strategic arrangement with them in 2006 and I thought they might be a good way to take advantage of nanosolar's technology with a public company.
  •  
    zawy
    Feb 06 01:48 PM
    SOLF was cited by the non-male on Fast Money to be the worst stock to put your money in. This Chinese company had a mysterious fantastic increase in profit and they will not disclose how much they are selling their cells for. She seemed to be saying it's a very questionable company. My impression is that no one knows what their profit or sales are.
  •  
    zawy
    Feb 06 02:07 PM
    Here's the fast money link:
    www.cnbc.com/id/227348.../

    She says she's talked about negatively repeatedly. Things like odd measuring sticks, nondisclosure, and the CEO owning half the stock.

    I really wish seeking alpha would fix the awful programming: they have independent URLs or something for each image and they're loading seperately, even though it's the same image. Their pages, especially if there are any comments with the 3 little images next to the name, time, and date, cause really spooky things. One article had to load 300 freaking images, in addition to all the fetch and reply requests needed for each one, simply because people posted a lot of replies....but there are only needed to be 30 or so independent images with only one fetch and reply each to do the same thing. They must have paid a lot of money for someone to do such a bad job. Normal programmers can't do this bad if they tried.
  •  
    Plasticman
    Feb 06 03:26 PM
    I'm not saying that I recommend SOLF, because I agree, they have had some weird dealings. In fact, I have shorted them in the past. They are just an example of numbers that don't add up for some reason.

    I am a big believe that in the long run the CIGS companies are going to be the big winners, and the silicon based companies will implode because they will be priced out of the market. There is a new article out by the Motley Fool today on the subject. The only thing that they state that I disagree with is that CIGS is automatically less efficient. The Nanosolar technology seems to be at par with silicon. Here is a link www.fool.com/investing...
  •  
    Plasticman
    Feb 06 05:04 PM
    I was not suggesting that SOLF was a buy. In fact, I have shorted it in the past for many of the same reasons that Finnerman suggested. I was just using them as an example of earnings information out there that doesn't make sense. Maybe something is getting screwed up in the ADR translation.

    Personally I like the solar companies that use CIGS technology. I believe that eventually they will take over the market, and the silicon based companies will implode because they won't be able to compete. Today there was a good article written in the Motley Fool that does a good job of comparing the technology. Here is a link to it. www.fool.com/investing.... The only aspect that I believe that they got wrong is the statement about CIGS having a lower efficiency. While that is true for most of the companies, Nanosolar states an efficiency in the 20% range, which is right there with the silicon manufacturers.
  •  
    Plasticman
    Feb 06 05:05 PM
    Sorry about the double post. The first one didn't show up when I checked later, so I rewrote it.
  •  
    Jack Yetiv
    Feb 06 07:25 PM
    OK, I see I have my work cut out for me, responding to some of the posts above. Regardless of whether I agree or disagree with you, I will not be disagreeable about it and I VERY much appreciate this interchange.

    Let me begin with a general, but very important comment: Most of the posts critical of my article are simply pointing out two weaknessses I pointed out extensively in my own article--namely, (1) that it is based on 2008 earnings and sales estimates, and (2) that is is a STARTING point for analysis, not an end point. I challenged everyone to come up with a better alternative, and DOCUMENT why it was better.

    NOBODY has done so. I think that says a lot.

    Some people (primarily Acker) have pointed possible problems with his anticipation of CSIQ's ability to execute at a rate that he concedes to be faster than any other company in this space. I share his concern, but (1) he has not referred me to any data to support his doubts, (2) I think a 2008 FPE of 11 already prices in that doubt, and (3) don't other companies in this space also have their issues? I have read fewer negative issues (collectively) on CSIQ than almost any other company in this space.

    Acker also fails to explain the amazing ramp of revenues in the first 3 quarters of 2007 of $17 million, $60 million, and $97 million. If the last quarter comes in at the range guided by the company ($110-120 million), does that 4-quarter performance suggest poor execution to ANYONE? Not to me, that's for sure.

    Also, for a company that has performed so well in 2007, is it THAT hard to believe that they can grow sales by $20 million per quarter (about 15%/qtr) in 2008 to $135, 155, 175 and 195 million? If they do that, they get to $655 million in revenues in 2008--at the low end of their guidance.

    Acker is also incorrect in some of his facts. His biggest error is when he says CSIQ does not make cells--they simply "assemble" modules. Quote from 3rd quarter report: "Our second 25MW solar cell MANUFACTURING line is now operating at full production capacity. In addition, we have completed the installation of our third and fourth lines and expect to bring our total INTERNAL SOLAR CELL MANUFACTURING CAPACITY TO 100 MW STARTING NEXT MONTH." (Emphasis mine).

    For perspective, 100 MW at todays ASP's equates to about $350 million in revenue. So far from Acker's assertion that CSIQ does not produce cells, effective TWO MONTH AGO, they were producing THEIR OWN CELLS at an annual rate of $350 million. Not so hard to believe they can hit $650 million in total sales in 2008.

    Acker also gets excited because some of CSIQ's revenues in the third quarter were for "silicon material sales." What he forgot to mention is that those sales amount to $3.8 million, while the other $93.6 million in sales came from module sales. I can't get worked up about 4% of CSIQ's sales.

    As to CSIQ's use of the word "believe" it has secured 90% of its silicon, I have looked at other company's reports and that sort of word is common. I think it is a CYA word because things can happen and they don't want to be nailed for stating as "fact" something that is more like avery realistic expectation. For example, let's say they have a signed contract with a new supplier, and have even made a payment, but that supplier defaults. They are entitled to say they "believe" they have that supply, but things go wrong. In any case, even 70% would be good--90% is incredible compared to almost any other company.

    Next, Acker criticizes all analysts and says his projections are as good as any of theirs. Cool. I hereby ask Acker to provide his estimate for sales and earnings for CSIQ in 2008, and I will average his value with all the rest of the analysts. He says he would love to see Adam Hinckley's analysis. It might have been better to look at it before criticizing it, but in any case, that analysis is available and I would encourage Acker to get it, analyze it, and THEN tell us what is wrong with it.

    My next post will continue where this one has left off.

    Jack Yetiv
  •  
    Jack Yetiv
    Feb 06 09:05 PM
    One final comment to Acker. His comment that CSIQ cites a potential technology (UMG) as their "back-up" plan is a misstatement of what I posted and what the company said in their last quarterlt report. The UMG technology--if it pans out--will ADD to their sales, not serve as a "back-up." It's a bonus, not part of the guidance. If it comes in fully, they should approach $800 million in sales. If it does not, they' come at the midrange of their $650-$750 million guidance.

    In summary, without being discourteous to Acker, I tend to agree with Mr. Yusufi who believes that Acker perhaps holds a short position in CSIQ. Much of Acker's "information"... is wildly inaccurate, even on critical issues such as the fact that he says CSIQ does not manufacture solar cells, whereas in reality, the company has manufacturing capacity of 100 MW. If he has "missed" this critical fact, he is either poorly informed or is not objective.

    Mr. Yusufi believes that quality of management and "sentiment" should be taken into account. I agree, but did not do so in my article for several reasons. First, I wanted my article to be objective, and these two issues are as subjective as things get. Second, I don't know enough about the quality of mgt at the 11 companies I compared to even make a semi-credible argument on the issue, and when I don't know about something, I simply avoid it. And third, "sentiment" is a changeable thing--and a central theme in my article was that I believed that "sentiment" should soon be changing, against the likes of FSLR and in favor of the low-PE stocks in this sector, if they continue to perform.

    Solarman likes TSL because it is "integrated,"... but so is CSIQ for at least 50% of their panel production, and I suspect this percent will increase going forward. Having said that, I stated in my article that TSL is indeed worthy of further analysis, and that one area I would want to get comfortable with is why there is such a disaprity between high and low EPS expectations for TSL in 2008.

    Solarman also says "smaller" companies may get killed. As you can see from the table in my article, I did not really include any "small" companies on the list. Expected 2008 revenues range from $600 million for SOLF to $1.5 billion for STP, with both TSL and CSIQ at $700 million. If the absolute biggest manufacturer of solar panels does $1.5 billion in sales in 2008, $700 is not "small," in my view. If Exxon has sales of $300 billion per year, another oil company that does $150 billion per year would hardly be considered "small."

    Dr. Sarkar is asking me about marketing for CSIQ. I cannot help since I have asbolutely no relationship to CSIQ (except owing its shares), but he can go to the website and contact the company directly.

    David White says I made a mistake on expected EPS for SOLF. He is correct. He says Yahoo shows $1.30, Schwab shows me $1.23. So at today's closing price of $15.35, we're looking at a PE of 12, rather than the incorrect PE of 19 in my spreadsheet. But we all acknowledge that SOLF has a lot of questions about it--and still has a slightly higher PE than CSIQ, a much higher PS (almost double) and a 2007-to-2008 expected growth rate that is LESS than CSIQ's. Given these facts, SOLF would only be a better play if its PE was less than CSIQ's. Since it is not, to me CSIQ is clearly the better play.

    Mr. White also believes LDK has a better PEG than CSIQ. That is vastly incorrect. LDK's sales are expected to less-than-double in 2008, while CSIQ's are expected to double-and-a-half, so CSIQ's growth is faster. In addition, LDK's PE is about 20, versus 11 for CSIQ. Thus, CSIQ's PEG is actually about 3 times BETTER than LDK's.

    Zawy says FSLR's growth rate y-o-y is 167%. That is incorrect--the correct number is 67%. CSIQ's expected growth rate from 2007 to 2008 (700/290) is 241%. FSLR isn't even in the ballpark in terms of growth rate, gentlemen (and ladies). Let me also correct another commonly-held misconception--namely, that FSLR is a "big" solar company whereas CSIQ and the rest are small. But the spreadsheet in my article shows otherwise--FSLR's expected revenues in 2008 are $794 million, whereas the low-end of CSIQ's guidance for 2008 is $650. Hardly a difference to get excited about. If you compare expected EPS in 2008 between CSIQ and FSLR, it's not all that different eithyer ($2.01 for FSLR, $1.63 for CSIQ).

    Zawy also says FSLR's PE against 2009 (NOT 2008) expected EPS is 68--which is still more than SIX times CSIQ's PE for 2008 (not 2009). That does not sound like a bargain to me. Indeed, the next time FSLR goes to $190, I'm gonna buy puts. I haven't done so before, but I am getting extremely tempted to do so.

    Finally, Plasticman also says my EPS numbers for 2007 appear "wrong" to him. They are not. Here is the reason for the discrepancy: As I explained in the article, I quoted 2007 EXPECTED EPS, NOT trailing 4 quarters. The EPS he finds by looking on Yahoo and elsewhere almost always shows trailing 12 months, hence the difference. Basically, I included 4Q07 in my 2007 EPS numbers, but his sources did not because most companies have not yet reported fourth quarter.

    Thanks, Jack Yetiv

  •  
    robert.bohre
    r
    Feb 06 09:54 PM
    Nice article Jack, and way to back up your analysis. I'll be looking into CSIQ. Thanks.
  •  
    zawy
    Feb 07 09:44 AM
    The best-fit curve for FSLR growth for the past 5 years is 167% yoy. Multiply 2.67 times 2002 production and keep doing that until you reach 2007 and even into 2009 guidance, and it comes out right. 167% yoy. I did an article on it for seeking alpha. CSIQ has yoy revenue growth of 189%, best fit curve 2003 to 2008, and 220% for 2006 to 2008. But looking at their fact sheet, their gross profit from revenue has gone from 40% to 17% to 5% in the past 3 years. They hope for it to return to 14%. FSLR guidance is that their profit per revenue will stay close to 50% thanks to no competition with others who make this type of cell. CSIQ has revenue of $3.8 per total watt and FSLR has $2.40. Let's be optimistic and say CSIQ is returning to 14% profit per revenue, or $0.53 profit per watt verses FSLR's $1.3 profit per watt. So CSIQ has to produce and/or install and/or assemble 2.5 times more to match FSLR in profit. If CSIQ stays at 5% profit, it has to produce 7 times more. So it's not fair to compare just watt production.

    If world demand is really three times lower than this year's production (according to another seeking alpha article). In excess supply, only the cheapest cells will be sold, killing everyone's profit except for the one (or two) company's that already have the cheapest cell. They will be able to retain their profit margin better. So it's not risk-free to invest in this Chinese company that has a Canadian name.

    On the big plus side for CSIQ, if it acheives the 14% margin and equals FSLR watt production very soon as expected, then having a market capitalization 26 times less with profit/watt only 2.5 times less means it would have to be 26/2.5 = about 10 times higher to match the price of FSLR. At only 5% margin, it still has increase a factor of 3 to match FSLR price.

    I can't believe anyone's still discussing CIGS when there's not enough indium in the world's minable crust to meet LCD needs in the next 5 years, let alone solar cells. I thought that pig was porked already.
  •  
    zawy
    Feb 07 10:15 AM
    OK, to correct some bad english: if supplies are going to be three higher than demand this year thanks to all these solar cell companies ramping up production, FSLR is in a good position (ignoring current high price) because it has the cheapest cells and largest profit margin. It sales could stay high, profit stay the same, while the sales and profit of everyone else plummets turns into massive quarterly losses. The 5% margin of CSIQ isn't reassuring in this scenario. We depend heavily on CSIQ's guidance on what they think they can sale and at what price, if we are to ignore the seeking alpha article that discusses too much supply in 2008.
  •  
    Jack Yetiv
    Feb 07 10:35 AM
    To Zawy: I agree with your analysis above, but really, your conclusion contradicts the first 2/3rds of your post. What you miss in the first part of your post is number of shares of each company. The point is, Who cares which company has more profit--whether 3X, 7X, or 100X? If Exxon makes $40 billion profit in a year, and has 40 billion shares, its earns $1/sh. If another company only makes 1/40th as much profit--only $1 billion--but only has 100 million shares outstanding, it makes $10/sh EVEN THOUGH IT MADE a tiny profit compared to Exxon.

    The conclusion to your post basically is another way to compare PE's, which is how you get to the conclusion that if CSIQ makes the money that the analysts (and the company) believes, then it worth potentially 10X more than FSLR.

    I agree. Your can reach the same conclusion by comparing FSLR's forward PE of 85 (see my table) versus CSIQ's FPE of 11--about an 8-fold difference.

    Two last comments: The best-fit curve analysis you've made is correct, but alas (with all due respect), it is useless. Stock buyers do not buy stock based on PREVIOUS growth rates--they assign premium to stocks for FORWARD growth rates. You acknowledge that FSLR's forward growth rate is less than CSIQ's (assuming they both perform as expected), which makes sense because it is harder to double bigger sales than it is to double smaller sales.

    Second, I AGREE (as I stated in my article) that FSLR--because it is the recognized low-cost producer, because people think its sexy ("sentiment"... because it has executed longer than CSIQ, etc--DESERVES A PREMIUM.

    What I vehemently DISAGREE about is that it deserves a 10-FOLD premium to CSIQ. It simply does not under any reasonable investing principles. Two-fold, maybe. Three-fold--that's a stretch.

    But 10-fold? That's simply ridiculous.

    Zawy, do you hold a position in FSLR, and if so, what is your average cost and what are your plans for holding?

    Jack Yetiv
  •  
    Plasticman
    Feb 07 10:49 AM
    Jack, I understand your statement about trailing twelve vs. 2007 calendar earnings. According to Yahoo Finance and Etrade, the consensus estimates are for -$0.06 2007 and $1.43 2008. For the first three quarters of 2007, CSIQ has total earnings of -$0.23, so they would need to earn $0.49 in the fourth quarter to get to your number. The consensus estimate that I see for the fourth quarter is $0.18 with a low of $.15 and a high of $.22.

    I am not trying to criticize your methodology, I just want to find a source that gives me (and you) credible numbers.

    Zawy, I have not heard any discussion on Indium. Please show me a link. Obviously Popular Science missed that when they made Nanosolar Top Innovation of the Year 3 months ago. I don't claim to be an expert, but I am an engineer and can weed through the hype easily. I agree with Jack in that this discussion is helping to bring out the facts, and separate them from the hype. It is the whole issue of not being able to get to the facts that is so frustrating for me. SOLF is a good example of that.
  •  
    MSA
    Feb 07 11:13 AM
    Hi Jack: I appreciate your well thought out and comprehensive responses. When I have more time I will address the solar industry as a whole, and the companys that I see as the "winners". Suffice to say, while your analysis is "correct" in terms of crunching the numbers, when you leave out any sense of objective analysis of the management, performance history vs, estimates, market supply and demand drivers, and technological and capital (scale) competetive factors, then the analysis, is thin at best, and implies that all we need to do is crunch numbers (projected numbers, no less) on stocks and the results will be as good as gold. I do not buy in to that, which is why I criticized your analysis. BTW - I am not short CSIQ, but I am long a couple of other solar stocks for reasons that I will go into later (LDK, and EMKR).

    Good luck to all,
    AckerAnalytics
  •  
    Plasticman
    Feb 07 11:25 AM
    I will stick with CIGS. My research on idium points to a mining and refining short term issue, not an ultimate source issue. Right now a considerable source of idium is thrown away by the major ore miners. Now that the price has risen about 8X they are now looking into ways to refine it from their waste streams instead of throwing it away. Also new mines are being developed. The Indium Company issued a press release talking about how the shortage was short term not long term, right after the Wall Street Journal reported on the shortage possibly causing a problem with the manufacture of flat panel displays. Indium is about 4X more plentiful than silver in the earth's crust. It is used in miniscule amounts in FPDs and Solar Cells.
  •  
    zawy
    Feb 07 03:09 PM
    Jack, neither of my post were meant to promote or detract any stock. It's more like my doodling and musing over your article and wondering if FSLR needs to be so low and if CSIQ needs to be so high. I have 1.5% of my net worth in FSLR and 5% in solar. I recently divested myself from 10% net worth in oil and gold. Thanks to your article motivating me to look into CSIQ, i am considering switching to CSIQ. The CEO feb 2007 speech on the web site makes me like management as well as can be. Engineering, motivation, ethics, and long-term plan seem to be all in place as solid as a rock. Only new technology and disinterest in green can displace them.

    ITO is using 84% of the world production. Poduction is 510 tonnes/year, leaving 80 tonnes for non-ITO. Being very generous to CIGS, ITO indium recovery could improve to keep up with bigger TVs eventhough demand from that sector is increasing 50% per year. There's already a shortage. Solar cells can't outbid big TVs, but can pay more for non-ITO indium. 80 tonnes means only 2 GW indium per year can be produced, enough for $2B in profits in 2010 for all CIGS companies. But the indium stops there. Recovery can improve only about 30% and CIGS at 1.5 um can theoretically get down to 20 tonnes per GW, so the indium might last until 2012 to make 8GW and $8B for the sector before hitting the indium wall. Indium is 100 ppm in zinc and you can't get it anywhere else. Indium accounts for 5% of a zinc producer's gross and maybe 15% of its profit. So if the price goes up another factor of 10 (which is supportable by both CIGS and ITO ...$0.10 per watt just for the indium), I guess you could flood the market with zinc just get to the indium and extend it even further. The theoretical max reserve base of indium with 100% recovery is 16,000 tonnes, theoretically enough for 1 TW, or $1 trillion in nansolar profits. But 5% of that disappeared into LCDs last year, given the so-so recovery. In 2010, 10% of the reserve base could go into LCDs, someday coming out for recycling.

    Take your pick:

    1) LCDs have to stop using so much, or
    2) zinc plummets on indium going up a factor of 10, or
    3) CIGS is idled by 2012 if not 2010.
  •  
    Plasticman
    Feb 07 08:39 PM
    In the 1970s they said we were 10 years away from depleting our oil reserves. That was for a commodity that many people cared about. In the past, very few people cared about indium. Now a lot do. Indium does not only come from zinc mining, but that supply has been plentiful enough in the past to take care of the negligible demand. Now that it is becoming a valuable product people will start looking for it, and voila, they will find it. Here is an excerpt on the subject from electronicdisplay.com. They are mainly talking about how it will affect flat panels, but they do mention solar.

    "Indium Corp. of America (Utica, NY), in an apparent restoration endeavor against the WSJ article, published a press release that says it’s confident of the sustained indium-metal supply. The company insists the supply imbalance occurred because of a time lag between emerging demand and the presently available supply and that “...higher prices will draw forward additional supplies which will alleviate any scarcity.” " - I expect that a company that only deals with indium would know more than either you or me. Plus they benefit from scarcity talk since that drives the price higher, yet they are talking it down.

    Here is an excerpt from displaydaily.com
    "Solar cells are growing in importance and have a distinct similarity to FPDs. For example, the recently announced Sharp Gen 10 fab for $3.2B will be used not only for LCD panels but for manufacturing solar cells as well. In one year, the fab is expected to be able to produce enough solar cells to produce 1000MW of power indefinitely. This is serious power, the equivalent of a medium-sized nuclear power plant. Is Sharp likely to close this fab due to a lack of indium? Not likely!

    Insight Media sees three ways of easing the coming indium shortage. First, the existing zinc and lead miners are likely to recognize the economic value of indium and make more of an effort to refine it from their ores. Second, dedicated indium mines could be developed. While $533M may not be a big market to an established multinational mining company, it is a lot in terms of a startup.


    Finally, substitute transparent electrodes could be developed that reduce the demand for Indium...

    Increased indium supplies and the availability of substitutes make it unlikely the LCD industry will ever need to turn away customers due to a lack of indium. That Sharp Gen 10 fab, with its 2,850 by 3,050 mm substrates, will probably find a way to operate, even with indium at $1000 (or more) per KG."
  •  
    zawy
    Feb 07 11:26 PM
    Purified indium already costs something like $2,000 per kg.

    Although i agree with all your facts, i don't think they detract very much from my concerns. A likely scenario is that production goes up a little (eventhough it went down in 2007), LCD manufacturers find "good-enough"... alternatives, and everybody pays a bit more. But in only 3 years there has to be massive changes in the indium world before CIGS can continue to grow.

    I would be interested to know where sharp plans to get 50 tonnes of indium for 1 GW. That's 10% of the world supply. The LCD world in 2 yeyears will not allow that much to be sold for non-ITO uses. They'll need about one gram for a $800 TV? Whereas $800 of CIGS cells will need 5 to 20 grams. Solar cells might be important, but on a square foot basis, people think a big TV is more important, and will pay accordingly.
  •  
    Plasticman
    Feb 08 09:58 AM
    Final post on the subject. According to pm-connect.com the spot price for indium today (2/8/08) is $425 - $575/kg. Here is a link to a paper by the Indium Corporation that was updated on 9/07 www.indium.com/solar/i....
    It goes into great detail of how they believe that supply will meet future demand. They do have a graph showing demand exceeding supply and using up reserves starting in 2009, but they follow it up with a comment that that scenario doesn't have new production and greater usage efficiencies factored in that will come from all the improvements that they state in the paper. They basically believe that as long as they are given the proper notice (2 years), they can meet any demand. Another chart shows that prices peaked in 2005 and have been dropping since then. Today's spot price confirms that.
  •  
    zawy
    Feb 08 12:06 PM
    In 2009, the indium report says there will be a shortage of 43 tonnes assuming 64 tonnes increase in non-ITO uses over 2007. The report does not give an estimate on future solar needs. 3 GW in 2011 will require 100 tonnes more than what they seem to predict, so it's 143 tonnes short (demand exceeding supply by 10%) in 2011 if the shortage remains about the same and if they are not sufficiently bullish on CIGS needs. Their charts do not go out to 2010 which is where I'm saying the problem could begin. Judging from their comments and graph, a 1% shortage caused a quadrupling of price. Their comparison to silver is absurd. The concentration that is in the Earth's crust is not related to minability. This type of reasoning is often use to make a false argument in metal discussions. The theoretically minable silver is 100 times more than the theoretically minable indium. The report shows that almost all increase in "production" is actually from LCD reclamation. There's only a 10% increase in virgin production from 2007 to 2009. The doubling of reclamation does not add up very well with the comment that reclamation is already at 60-65% of the 70% in LCD "targets" (waste). 100% reclamation only increases 35-40% of the 70% which is about 23% overall increase in reclamation instead of 100%. It's not a small discrepency since reclamation increase shown in their table is 90-95% of the total increase in supply. The difference is another difference of 25% between supply and demand. Although the table shows a great increase in reclamation, it does not show what happens when there is no more available efficiency increase in reclamation. That is, their table interestingly stops before the "end of the line" for significant indium reclamation increase...NEXT YEAR. The only valid point they make is that smelters can now afford to transport waste slag etc further from small producers. Most large producers already have it on site.
  •  
    ecoman12453
    Feb 18 11:43 AM
    Going back to TSL, I can clarify how TSL will at worst meet average consensus 08 Estimates, and at best blow out the highest number.

    TSL has a net profit margin of $0.35/share for the last 2 quarters, Q2, and Q3 2007. In these quarters, TSL's gross margins were at some of there lowest levels in the last few years averaging lower 20%. Also, last Q, the net profit margin was shrunk due to high OPEX, 11% of Revs, which is sure to go lower. Going into 2008, ASP's are expected to decline from 2007 average by a smaller amount, but the average for 2007 will be slightly higher than the ASP in Q2, and Q3.....Prices declined from Q1 to Q2, remained rather lower in Q3, and then increase in Q4. So what I am saying is Q2 and Q3 saw the ASP's that are low enough to account for most of the decrease into 2008. Now Poly is expected to increase 10% for 2008 but the 10% increase is said to be a peak in 2008 then prices drop in the latter half like YGE said. TSL is expected to cut POLY usage 10% for the 2nd half of the year so they should remain equal on poly costs year over year.

    Last, TSL's processing costs/watt are only going to go down from efficiency gains for modules' conversions, and economies of scale, and moving toward 100% in house cell processing.

    From $0.35/watt profit margin in Q2 and Q3 2007,

    Poly should remain the same for 2008, increased price will be mostly in 1st half but offset by less poly usage,

    ASP's will probably be about the same for 2008 as they were in Q2, and Q3 when they dipped to the lowest levels of 2007.

    Processing costs will only go down so at the worst case, TSL will maintain the same profit margin of $0.35/watt, (This is sure to increase due to less processing costs, but they will be taxed at a higher rate so I'll say profit margin is staying equal.)

    On 200MW (lowest part of guided range) and $0.35/watt profit margin, TSL does $70 million net income. This comes out to $2.75 EPS,


    For 2009, TSL has more growth than most other players as they plan to ship double 2008, 400MW, and there's no reason to think they will maintain the same profit margin, $0.35/watt.....ASPs will surely decrease significantly for 2009, but poly will start to trend lower, and they will be cutting poly usage once again, more cell efficiencies toward 19%, and more gains from economies of scale....so if they
    maintained the same $0.35/watt profit margin (I think this may go up a little due to the only downward pressure being ASP's but upside pressure being lower poly costs, less poly usage, and efficiency gains, processing cost reductions and economies of scale) If they do $0.35/watt on 400MW for 2009, they make $140 million
    or $5.51 EPS for 2009.
  •  
    theotherjimm
    ac
    Feb 29 01:27 PM
    I think you may need to adjust your "best value" model. SOLF has been oversold to $12 or less recently, which puts it at perhaps 12x annualized eps Dec quarter 2007 $1.00 (they may report $0.25 more or less in first half March), and less than 10x 2008 eps on 100% sales growth.

    Regarding your critic's assertion that CSIQ (and SOLF in fact) appear to be undercapitalized to grow as fast as they expect, he is missing the fact that SOLF, for one, has stated in their Feb 20 Piper Jaffray presentation that they are on much better terms with their large customer base now (25 customers now), payment terms are 30 days and even prepayments. I don't really know what Finnerman of Fast Money means by lack of disclosure. She must mean lack of quarterly reports. But who cares? Those are unaudited any way, and they do have their 2006 20-F on file. They've stated they sold solar modules for $3.68 per watt in 2007 q2, $3.66 per watt in q3, and expect to sell for $3.80+ per watt in q4, and likely $3.80 per watt in 2008. What we don't know is what portion of their 2008 silicon needs are secured, unlike CSIQ, which asserts 90%.

    CSIQ and SOLF are working on vertical integration and sourcing more silicon, it's true, they're at a competitive disadvantage now relative to FLSR, STP, SPWR, YGE, even to TSL. But they can still make better investments at $19 and $12 respectively, as they may be oversold on excessive pessimism, while the other players trade for higher multiples, some at astounding 20x 2010 eps expectations (FSLR), whichi is also 100x 2008 expectations. Like when EBAY went public during the dotcom boom, and went vertical, then sideways for years.

    This is about finding growth at a discount. It is highly unlikely anyone buying FLSR at $200 will grow their money 50-100% in a year, but if in 2008 CSIQ $19 proves they're heading for eps $1.86 and SOLF $12 for $1.44, with rising margins and progress toward vertical integration, then they are more likely than FSLR to rise 50-100%, toward 15-20x eps from current 10x.

    FSLR, on the other hand, would need to trade for 150-200x 2008 eps(e) $2.00+, and 30-40x 2010 eps(e) $10.00 to rise 50-100%. And FLSR's margins will contract as they ramp new capacity in early 2008.


  •  
    User 162614
    Mar 11 12:40 AM
    Impressive article and discussion.

    Before commenting on CSIQ I believe everyone should see this short presentation they copied onto the web (of course it sounds like many of you have already seen this):

    library.corporate-ir.n...

    Thanks for the article, it was a good read.
  •  
    User 162614
    Mar 11 12:42 AM
    Sorry, link got cut; this will work:

    tinyurl.com/2wkpau
  •  
    hcl1202
    May 13 09:55 AM
    Thanks, Jack.

    I purchased CSIQ after reading your article - my first reading/encounter of fundamental analysis. Took me a while to grasp it. This article was very well written and I recommended it to a friend of mine. Now, I am enjoying 99% unrealized gain - it is increasing constantly! Trading volume was more than 2 millions 7 minutes into the trading today.
  •  
    aqua
    May 23 02:01 PM
    Nicely written. I'd love to see an update, now that we've gotten thru earnings season
  •  
    aqua
    May 23 02:03 PM
    Oops. Just found your follow-ups. Nevermind :)
Loading...
Symbols:
Authors:
Quick Links:

ETFs In Focus

  • Long Ideas

  • Short Ideas

  • Cramer's Picks