Canadian Solar: Value Diamond in the Alt. Energy Rough

Canadian Solar (CSIQ) reported a very nice 4Q '07 Wednesday. Although CSIQ had previously guided to expected 4Q revenues of $110 to $120 million, it actually reported $127.5 million in sales despite having lost about 5 megawatts of production due to the severe weather in China recently. Had that weather impact not occurred, revenues would have been in excess of $140 million, a blowout by any measure, especially considering that 3Q '07 revenues were $97 million and given that 4Q '06 revenues were $24 million.

Here are the sequential revenues for CSIQ's 4 quarters in 2007:

$17M, $60M, $97M and $127M

CSIQ reaffirmed revenues of $650 to $750 million in 2008, but this did not include approximately $100 million in sales of a special "UMG-silicon" type of panel that CSIQ expects to sell in 2008. Indeed, it seems much more likely that CSIQ will actually sell over $750 million of solar modules in 2008, and if CSIQ exceeds the best expectations, revenues in 2008 could approach $1 billion, putting CSIQ's sales at the same level as First Solar (FSLR).

CSIQ also did well on the income side, reporting 20 cents of earnings per share, a very nice sequential increase from 2 cents per share in 3Q '07 (and an even nicer change from the loss in 4Q '06).

In the release and on the conference call, CSIQ reaffirmed that they have secured essentially all of their silicon for 2008's production, and also stated that ASP's in the first half of 2008 will actually be slightly up from 2007, followed by what sounded to be a few percent decrease in ASP's in the second half of 2008. CSIQ management indicated that none of their customers were seeking double-digit ASP decreases in 2009, suggesting that ASP's in 2009 will decrease in the single-digit range from 2007's panel prices.

However, because the degree of manufacturing cost decreases (due to increases in manufacturing efficiency and decreased silicon costs) will exceed decreases in ASP's, CSIQ guided to improved gross margins from about 11.5% in 4Q '07 to about 14% in 2008. Utilizing this information, one can roughly estimate that CSIQ should make about $2/share in 2008, as follows.

Gross profit in 4Q '07 came in at $14.6 million, yielding net income of about $5.5 million. If gross margin had been 14%, gross profit would have been about $18 million, and assuming that half of that additional profit dropped to the bottom line would have yielded about $7.5 million of net income on $127 million in sales. Assuming sales of $750 million in 2008, that would yield net income of about $45 million in 2008. This does not include a penny of income to be derived from sales of UMG-based solar panels, which actually have greater profit margins than normal silicon panels.

If CSIQ sells the mid-range of its guidance in UMG panels, that should add $100 million in revenues and $10 million in net income for 2008, yielding total net income for 2008 of approximately $55 million. Since there are about 28 million shares outstanding, this yields a fair estimate of about $2/share in 2008. I believe a case can be made that earnings will be lower or higher, but I think $2/share is a very fair estimate.

Given $2 in 2008 earnings, CSIQ continues to be by far the best value in the solar space bar none, because you have here a company that will almost triple sales from 2007 to 2008, will grow income from essentially nil in 2007 to $2/share in 2008—and yet is trading at a PE of 10 against 2008 earnings? A PE of 10 may be reasonable for a boring company growing at a few percent per year, but it certainly is not reasonable for a company growing at over 1,000%.

Undoubtedly, the naysayers will raise various arguments about CSIQ, which I will address in the comments section as needed, but there is simply no logical reason to value this company at a PE of 10 against 2008 income while FSLR—which will grow both its revenues and its earnings less quickly in 2008-is being valued at a PE of nearly 100.

Given the foregoing, I believe CSIQ should be trading at $40-50 within 6 months, if not sooner, based on $2 in 2008 earnings and a very modest PE of 20-25. Of course, by mid-2008, 2009's earnings should be much more visible, and if they are at about $3 (which is what I am modeling), a forward PE of 20 against 2009 earnings could very reasonably yield a stock price of $60 or more after 2Q '08 is announced this summer.

Along with TMA and HTE, two of my other favorite plays for 2008, this could be a very good year indeed. As for CSIQ, this is even more likely if the solar tax credits in the US are renewed (very likely) or even enhanced (reasonably possible, especially if one takes new state incentives into account), and of course, all three stocks will benefit from a recovery in the economy and stock market which many (including myself) anticipate will be well underway before the end of 2008.

Disclosure: Long CSIQ

Comments

[Zenfar:]

Do you have any more details of this UMG panel?

2008 Mar 06 10:36 AM

[David White:]

CSIQ seems like a decent stock, but this is too much hype. A lot of institutions own it, so it is getting more play than it probably should. It has some of the worst margins in the industry. If there are significant pricing problems in the industry, as many are predicting by the end of 2008, this stock could fall through the floor. Even SOLF, which is in a similar position, has had better margins. It also has a better PEG ratio .35 (SOLF) to .51 (CSIQ).

In my mind, although it has had significant bad press, LDK is a far superior stock. It is building its own polysilicone plant. It has so far kept gross margins above 30%. It will grow those margins once the polysilicone plant is built. It has a PEG ratio of 0.17. It will be a survivor in the solar wars. It is not clear that CSIQ will be. TSL is another company that has plans to build its own polysilicone plant. It doesn't have the negative publicity of LDK. It has better margins than CSIQ. It has a PEG ratio of 0.24. What is so much better about CSIQ?

Even what you say about FSLR is somewhat false. FSLR has long term gross margins of 45%. It is in a great position to weather a price war. It may be overvalued, but the company is unlikely to be at risk of being driven into the red by the coming price wars. CSIQ is momentarily hot. Enjoy it while it lasts. It probably won't be for too long.

2008 Mar 06 12:12 PM

[David White:]

I should have added that one reason for CSIQ's recent success has been its exposure to the hot solar markets of Spain and Germany. These markets are currently being predicted to cool in the near future. This does not bode well for CSIQ. Other solars such as LDK and FSLR have much more diversified sales.

2008 Mar 06 12:18 PM

[Jack Yetiv:]

Here is what I know about UMG (metallurgical silicon):

It is significantly cheaper to buy, but achieves somewhat lower efficiencies (16% versus 17% in CSIQ's conference call) and for this reason and due to some doubt about its longevity, it sells at a discount. However, because it costs less, the gross margin is better. CSIQ has indicated that it believes it will sell 30-40 MW of UMG panels in 2008, but this is not included in the guidance of $650 to $750 million. I think UMG could be worth about 20 cents/share in extra earnings in 2008. Without UMG, I think CSIQ will make about $1.70 to $1.80/sh in 2008.

To Mr. White:

First of all, "hype" is a contraction for "hyperbole," meaning exaggeration. What fact have I exaggerated? Frankly, if we want to discuss "hype," I think the most "hyped" stock in the solar space is FSLR.

Second, I know you are very knowledgeable about solar. What earnings do you come up with for CSIQ in 2008 and 2009? I really would like to know.

Third, although facts are facts, we can legitimately disagree as to a reasonable PE. You appear to think that a PE of 100 is OK for FSLR. What PE is reasonable if CSIQ makes 30-40 cents in 1Q08 and 40-50 cents in Q2, and appears to be well on its way to $2 in 2008? I would again be interested in your viewpoint on that.

Is it really rational for FSLR--which will have about the same revenues and maybe 10-20% more earnings (per share) in 2008 compared to CSIQ--to be worth TEN TIMES more than CSIQ?

Fourth, I agree with you that all things being equal, higher gross margins (and even more importantly, higher operating margins) deserve a better multiple, but I personally think that earnings are a better metric. In addition, higher gross margins do not deserve a multiple of 100, as FSLR gets.

It seems your big concern is that there may be "pricing problems" by the end of 2008, and they will hurt low-margin operators more than big-margin operators. Generally, I would agree, but I am not so sure about that in this case, and here is why. I am not convinced that investing in building polysilicon plants today will yield the best ROI in 2009 and 2010 because I believe poly prices will drop significantly in 2009 and even more so in 2010, and being integrated may not yield quite the advantage that is being touted for it today.

Indeed, I believe that as poly drops under $100/kg, as I think is possible (if not likely) in 2010, the pricing advantage that the thin-film companies (like FSLR) enjoy may be substantially reduced, and possibly even eliminated since FSLR has to build TWO panels to produce the same amount of electricity that is produced by ONE silicon-based panel.

I agree with you that CSIQ's concentration in Spain and Germany is a negative for the stock, but it is my belief that the market both in Europe and in the US and elsewhere will grow sufficiently to absorb all production well into 2009 (for example, in Italy, Greece, and US). I base this opinion on Mr. Qu's statement on the conference call (and CEO's of other solars that have recently reported seem to second this view) that as they are writing contracts to sell their production in 2009, no customer is asking for 10% of greater discounts. This tells me that supply is currently looking balanced with demand in 2009. This could absolutely turn out to be incorrect, and if it is, all the solars will be hurt, not just CSIQ.

I also agree with you that SOLF, TSL and LDK offer potentially good values, but all three (and especially LDK) have gotten significant negative press and seem disfavored. This does not seem to be the case with CSIQ.

Of these three, I like TSL the best, and I will explain the reason why in a future post. I will say that the PEG's I have calculated do not come out the same as yours, and of course I have noted above that I believe in 2009, we will come to the realization that building one's own poly plant is not the positive it was believed to be, and this could impact LDK and TSL negatively. But I acknowledge I could well be wrong about this.

Finally, I am confused as to what is false about what I said about FSLR, especially given your statement that it "may be overvalued." I have NEVER suggested that FSLR is not a great company; instead, my position has always been that it is overvalued. If we agree on that, it seems to me that is the end of the inquiry.

Am I missing something?

Jack Yetiv

2008 Mar 06 08:25 PM

[MrT:]

hey gents-

why no analysis on STP and SPWR?

Both of these companies appear to be well positioned long term with excellent management, balance sheet, diverse product, and services.

BTW---FSLR is the short of the year.

2008 Mar 06 11:29 PM

[R D Olivaw:]

What would CSIQ's earings for 2009 be at a polysilicon price of $30/Kg and ASP of $2 per module?

2008 Mar 07 05:38 AM

[User 161444:]

Hi Jack,

Thanks for creating some excitement about this stock and allowing me to sell. What you're missing in your analysis is three things:

1. CSIQ has no competitive advantage in it's business. Nor do any of the Chinese silicon-based PV module makers. FSLR is entirely different story, since the barriers to entry in thin-film solar are huge. There will at most be a handful of thin film companies, while there will be hundreds of silicon module makers. Company's with no advantage in industries with zero barriers to entry, do not deserve P/E's above 10, let alone 5 in my opinion.

2. More to the point, CSIQ's earnings are a complete mirage. They really don't make any money. Yes they may report earnings, but when you actually bother to analyze the financial statements you'll see continued huge cash outflows. The business still needs a huge amount of financing to survive and there's no telling how it will do in a tougher financing environment. What's the proper valuation for a business that makes no money, still requires huge capital investment, has low odds of every attaining higher margins, and has zero competitive advantage?

3. Finally, in general, when looking at Chinese solar companies you need to adjust earnings to compensate for their low level of taxes. Without doing that you run the risk of faulty comparisons to non-Chinese solar companies, i.e. assuming the Chinese companies are undervalued, when in fact when you adjust for taxes, they are extremely overvalued.

2008 Mar 07 10:35 AM

[semperpax:]

How does snow in China in the first quarter of '08 affect revenue and earnings in the fourth quarter of '07?

2008 Mar 07 12:52 PM

[vitamin_j:]

I actually think (self-justifyingly, I have to admit, since I own some STP) that Suntech is the real value story right now, even though its trailing P/E is currently around 33, forward around 23. Investors have been thrashing this stock since a 4Q call that missed earnings by a small amount, and because they maintained revenue/production guidance for 08 instead of raising it (a wise long-term move, I believe). I think investors are wrong, and that STP will be a long-term winner.

I would ask Mr. Yetiv, why is CSIQ a better deal than STP on any other front but the multiple (since multiples can be fleeting in such a high-growth, highly volatile industry)?

- STP has already grown its capacity, and is now I believe the third largest panel maker in the world. They've got scale, and scale is an advantage now, and an even bigger advantage going forward. Why do I want to buy a panel maker that's not as big, that has to keep putting capex into getting as big, in a market that's increasingly going to favor the big players, and all when further capital for said capex might not be as forthcoming as before?

- Suntech's gross margin was 20% for '07. Its net was 12%. That makes its 07 net better than CSIQ's 07 gross. Why do I want to buy a company with currently low margins, that's going to try really hard to improve them in '08, when I could own a bigger panel maker that also has much bigger margins and much more of an established history of profitability? Please consider margins not just as a multiple of share price, but also as an indicator of company performance, health, and management quality.

- CSIQ is a one-trick pony. They make silicon PV panels. Now they also make silicon PV panels that might not work as well (but are, I assume, still under an industry standard 25-year warranty). Suntech too has their main business in silicon PV panels, however they also have a thin-film line; they have a BIPV business; they're manufacturing and distributing Akeena's Andalay panel (which cuts costs at installation level); they have a high efficiency panel going into production which will put them at the top of the efficiency heap; they have significant R&D and research relationships in Australia and New Zealand. If I want to buy a pure-play solar, why do I want to buy a one-trick panel maker when I can get a much fuller suite of solar products?

- If you're a long-term investor in this industry, you know that solar businesses aren't really competing against each other, but against current grid electricity providers. Suntech has laid out its plan for grid parity by 2012 (if my memory serves). What's CSIQ's plan? Are they thinking along those lines? I haven't heard (doesn't mean they're not), but I definitely want a solar that's got an eye on grid parity.

Now none of the above necessarily precludes CSIQ from being the value darling that Mr. Yetiv believes it to be. However, while their earnings progression looks impressive over 07, I would take such a figure as only the start of why a company is a "value" company. I want to hear why such a company is a great company that's going to excel in its industry. Maybe it is. I'd love to hear the case, qualitative and quantitative. However, I happen to believe that Suntech is in fact the real value story in the market right now, and I'd be interested to know why CSIQ beats it.

2008 Mar 07 03:57 PM

[jjason:]

Jack, Do you know of any companies in N.Y.,N.J. or PA. that sell csiq's products? I am interested in learning what their panels cost. Thanks for the article.

2008 Mar 08 01:26 PM

[Jack Yetiv:]

To JJason:

I do not believe CSIQ sells very much in the US. I know that they are in the process of changing that, but I doubt their distribution channel is set up in the US. That may change by the end of the year, I suspect.

To Semperpax:

Thanks for correcting my mistake--obviously, the Jan weather events did not have any impact on CSIQ's 4th qtr. I apologize for my error.

To User 161444 and Vitamin:

Thanks for your perceptive commentary and your civil means of expression. Here are my responses:

1) First, to User, I doubt highly that what I write can create enough excitement to move the stock very much when it's trading millions of shares per day. But I am glad your investment goals were achieved.

2) Second, I disagree that CSIQ has no advantage over peers. Based on sales ramp through the 4 qtrs of 2007 and guidance for 2008, it looks to me to be the fastest grower in this field. If you know of another company that has had a faster sales ramp from 2007 to 2008, please tell me which one it is because I have missed it. To me, that sort of sales execution IS AN advantage over peers.

3) It is my opinion that FSLR's technological "thin-film" advantage will decrease tremendously if not disappear completely when silicon goes to $40-50/kg. There is very little doubt silicon prices will get there--the only realistic argument is whether it will happen in 2009 (doubtful), 2010 (more likley, but probably not), 2011 (50:50 chance), or afterwards (pretty much guaranteed).

Remember, FSLR has to build TWO panels to equal ONE silicon-based panel. The aluminum, other materials, etc, for PANEL #2 will INCREASE in cost over the next few years, while the silicon in CSIQ's panels will DROP substantially in price.

4) Your point #2 about CSIQ's not making earnings and needing huge amounts of capital to continue going confuses me. I have studied their recent earnings release and they look like real earnings. Perhaps you can explain to me what you mean. In addition, in their recent release, they stated they have ALL the capital they need to fund all of their expected expansion in 2008.

5) I also am unclear about "adjusting for taxes." Can you tell me exactly how you would make that adjustment?

To Vitamin:

Although I like STP and agree with much of what you say about it, STP is the PERFECT example to underscore my concerns about high-PE stocks in the solar space.

When I wrote my article comparing the solars about a month ago, CSIQ was trading at $18.56 --a PE of 11 against expected 2008 income--and STP was at $51.70, a PE of 30 against expected 2008 income. STP's earnings report wasn't bad at all, yet it sold off. CSIQ's report was very good, but not really a blowout, and it went up.

Why? Because STP HAD to BLOW OUT to justify a PE of 30, and when it reported well but did not blow out and even suggested a leveling of sales growth in 2008, it got pummeled.

The same point can be made about SPWR ($73.29 when I wrote my article, about $61 now), and JASO ($60.38 when I wrote the article, $42 now, split-adjusted) and LDK ($39.84 when I wrote the article, $22 now).

Whereas CSIQ is a bit up since a month ago, most of the other solars are down. The lone exception is FSLR, which was $171.28 when I wrote the article and is $197 now.

However, if next quarter or the quarter after that, FSLR fails to BLOW OUT earnings or issues sub-phenomenal guidance, I would not be shocked to see it shed 50% in stock price over the next few weeks, dropping to a PE of "only" 30 or 40.

A good example of this phenomenon is AAPL's drop from $200 to $120. Apple is still a fabulous company doing great things (just like FSLR), but expectations had risen so much it was bound to disappoint. I believe the same will happen with FSLR sometime THIS year.

Finally, I have addressed gross margins above, so I won't repeat here, but in general, my opinion is that although gross margins are important and worth some value, they DO NOT justify a doubling or tripling of a PE or PEG. And even today, STP trades at a 2008 FPE of 20 versus 10 for CSIQ. Also, keep in mind that it is much harder to double $1 billion in revenues than it is to double $300 million in revenues.

Here are STP's last 4 quarters of sales:

$247M, $317M, $387M and $398M.

And if I recall, they guided to relatively flat revenues in 1Q'08.

I hope I have shed some more light on why I like CSIQ.

Jack Yetiv

2008 Mar 08 09:53 PM

[User 161444:]

Hi Jack,
Just because a company reports accounting earnings, doesn't mean it actually makes money. You need to look at cash-flow statements, and when not available, compare balance sheets between quarters. Don't understand adjusting for taxes? Sorry,

2008 Mar 09 10:38 PM

[User 161444:]

not a place for financial accounting lessons.

Let me ask you a question, I have an apartments for sale. I collect $1 million a year rent on the apartment, and that's up from $200K last year! I want $10 million for the apartment. It's a great deal.

Yes, the apartment requires $5 million a year in maintenance, rent is usually overdue for six months at a time, and there are plenty of other apartments for sale in the neighborhood that are exactly the same and require less maintenance. But, not to worry, my rent went up the most last year, I can assure you of that. So that means I have the best apartment in the hood. Why else would my rent be going up so much? So how about that $10 million? It's a great deal.

2008 Mar 09 11:14 PM

[absurd:]

It's obvious you're invested in CSIQ and therefore hyping it to the max. The Q1 storm afecting Q4 revenues?? was very funny, it also tells me you're not that good with your analysis and very biased. Now, I do agree with David White about LDK -- a winner, however in the short term CSIQ has the momentum. I am afraid it's a "pump and dump" technique by MMs, so please buy low sell high, don't get overlyexcited with a stock, especially in an environment like this. The solars will go up, no doubt about it, but not until the financial crisis storm goes away. Next week might prove to be very good for traders, lots of bargain hunters will take solars up, for how long? who knows. Good luck

2008 Mar 10 07:28 AM

[zawy:]

Jack's comment about the storm probably resulted from CSIQ mentioning that the storm had no effect of 4Q results.

The "value" of a company is pretty much DEFINED by Jack's method. All the things David White and User mention are already INCLUDED in Jack's methods, except for price wars, which Jack counters with the drop in silicon price argument. I would like to know how much their costs decrease with a 30% decrease in silicon. CSIQ is trying to produce as much as possible while acheiving an acceptable profit margin that keeps investors happy. This emphasis on production over short-term profit is why their yoy growth in watts produced exceeds 2nd-place winner FSLR. Their primary goal seems to be to supply the world with massive amounts of good solar. I like that, especially for an industry that has no proprietary advantage. The CEO is basically a PhD solar techie, plenty wise in the financial things, and not playing games with anyone, let alone the market. They want to be the Chiquita of bananas. FSLR does too, but FSLR is less likely to install residential, basically killing 30% of their watts produced. FSLR also has little desire to vertically integrate, where CSIQ could make all its profit if it decides to squeeze out new players by reducing cell price to zero margin.

CSIQ does not have the contracts to support $1 billion this year.

2008 Mar 12 10:45 AM

[Camden:]

Jack, you've made a compelling case about why CSIQ is the safer stock investment based on valuation but I don't think you've made the case for why it is a better company or has better long term prospects. Sales growth over the last two years only looks backward. Do you have information on how CSIQ's signed contracts (backlog) compare with others like FSLR?

2008 Mar 12 12:50 PM

[dj10:]

User-

"not a place for financial accounting lessons."............

Thanks for the hypothetical apartment example. Can you list
actual cash flow figures for Canadian Solar that support your contention? Or provide a reference or a link?

2008 Apr 09 09:40 PM