Description

full document here: https://www.dropbox.com/s/i8q75xrddaobwcb/Siltronic_quick%20writeup.pdf?dl=0

Thesis: this is a C+ business (which may be moving to B-) trading at an A price

Siltronic is attractively priced at 1.7x 2015 EBITDA (and probably somewhere between 2-4x 2016 EBITDA. WAF had 18% EBITDA margins in 2015 and spends roughly 8% on capex). While the industry (commodity silicon wafer production) is gut-wrenchingly cyclical, I think there are a couple reasons to own the stock today: 

• Valuation of 1.7x 2015 EBITDA (adjusted for forex hedges) and ~3.5x 2016 EBITDA assuming a halving of margins. Including pensions, which are being discounted at 1.65%, the metrics would be 3.5x/7.0x 2015 and 2016 EBITDA respectively
• Valuation of the entire sector pummeled due to decreasing demand from global slowdown in 2H’2015 (destocking), which has led to some pricing aggression to as utilization fell from 95%+
• Market consolidation has resulted in top 4 players now controlling 80% of supply, with no new supply having been added since 2007. Furthermore, global weakness has led to the #4 player (SEMI) being put in play as of February 19, 2016, and Siltronic appears to be the logical strategic acquirer. A merger of Siltronic and Sunedison would create 3 players each with 26-27% market share.
• In the Q3’15 Call, Siltronic’s CEO said : “Consolidation is a very interesting subject. Everybody is talking about it and has been relatively outspoken about the idea and about who will participate and who will not participate and we will stick to it that consolidation without Siltronic participating is not very likely to happen. On the other hand, we continue to focus on things we can influence on our own, which is technology performance, quality performance, and cost cutting performance. So yes it is a topic on our agenda but not the hottest topic”
• Demand for 300mm is finally catching up to massive capacity oversupply over the past decade and predicted to continue growing 7+% long term, with indications that demand will exceed supply in the next 1-2 years (which will require substantial price increases to induce new capacity additions)[1]. Ed Cunningham talks about tracking the capital budget cycle, and the wafer industry is now going on 10 years with minimal capex.
• Siltronic has the strongest balance sheet in the industry with a net cash hoard of ~156 million euros is 30% of the market cap, and is substantial firepower to withstand a prolonged downturn on the scale of a 2009 recession (SEMI will break its bank covenants, and SUMCO has 2.5x turns of net debt, 2.9x including pensions)

It is hard to stomach investing in an industry that hasn’t generated value over the last decade, but the structure of competition has changed significantly since the halcyon days of excessive billion dollar investments (with a capital response similar to what the oil industry is currently experiencing, but drawn out over 10 years). Despite the slowdown in silicon demand in the second half of 2015, if you believe that aggregate electronic demand will continue to grow long term, I think the current price of silicon manufacturers (SEMI, SUMCO, WAF) is very attractive.

Why Siltronic? The simple reason is its incredibly strong balance sheet (though it also has an incredibly large pension being discounted at 1.65%). But combine that with a small float, sub $500 million market cap, and currency hedges that are cutting EBITDA margins in half, and the equity doesn’t seem screen attractive. (I do think SEMI might have the best IRR over the next year but if a deal isn’t solidified it will free fall)

Industry background:

Siltronic is the third largest player in the manufacturing of polysilicon wafers (http://www.siltronic.com/int/media/documents/investor_relations/presentations/Siltronic_Investor_Presentation_February_2016.pdf). The wafer is the foundation of the vast majority of semiconductor devices (of which memory and microprocessors are the majority) and demand is tied to the proliferation of electronic devices:

[uses of polysilicon]

The industry has been difficult to invest in as the massive expansion of capacity in the 90’s and 00’s (similar to fiber capacity) led to negative/feeble returns on capital. Due to this excessive capital cycle, capacity has largely remained unchanged over the past decade and the industry has consolidated from over 20 suppliers to a handful today. Five suppliers account for nearly 90% of the market: Sumco (26%), Shin-Etsu (27%), Siltronic (16%), Sunedison Semiconductor (11%), and LG (10%).

In 2015, analysts predicted that demand would finally exceed supply (roughly 5.2 million wafers per month globally. Each wafer is ~110 square inches), and excitement built around potential price increases in an oligopolistic market (WAF’s utilization was >95% in 1H’15). The charts below illustrate the magnitude of capacity buildup up to 2007 (the same capital budget cycle hard commodities are experiencing at the moment), and the following decade of minimal capital investment needed to balance supply and demand:

[demand/supply] [capacity investments]

[estimated supply/demand balance]

Why hasn’t capacity been added since 2007/2008? Because it is incredibly expensive (Semi states that the reproduction value of their fixed assets is 3.5x their net book value in their S-1), and numerous data points suggest that adding capacity is simply uneconomical unless prices rise. I’m not exactly sure where 300mm wafer prices are at the moment, but I think an educated guess would be somewhere in the ~$60 region[2]

·         Siltronic has substantial “open shell capacity” up to 1m wafers per month (from ~780k), but indicate it will cost roughly $200 million per 100k capacity (as of 2014/15. In 2006, SUMCO spent $800 million for 200k capacity). On a unit metric basis, to get an ROIC of 10%, would require $200 in EBITDA per unit wafer capacity and $240 including 2% maintenance capex:

o    Estimating that ASP per wafer is around $60 (in December 2003, 12 inch blank wafers were ~$200[3], and have fallen some 75-80% as of 2014) and Siltronic is making 20% margins ($10), prices would need to rise ~15% to hit a 10% ROIC. While the $60 is undoubtedly a guess, the fact that you can buy Siltronic 12 inch wafers on ebay for $58/wafer probably signals a pricing in that region. A lower price at $50 would imply prices need to increase at least 20-30+%

o    In February 2016 presentation, Siltronic states “Should 300 mm wafer pricing improve to attractive levels , this shell capacity provides Fast and cost-efficient expansion option on an opportunistic basis to increase capacity”

·         Greenfield capacity requires $300 million per 100k capacity and would require substantially higher prices

·         SUMCO has also publicly stated that investing in new capacity would require price increases of ~30% to justifiy new investments

[ASP wafer prices]

Of course, there is always the risk of debottlenecking and some shadow capacity, but based on current fab expansions of the top 20 semiconductor manufacturers, 300mm fab capacity is growing ~7% a year, in line with Gartner estimates of mid/high single digit 300mm basic wafer demand growth.

What’s Changed

Since the middle of 2015, shares of semiconductor wafer producers have been pummeled, with SUMCO down 70%, Siltronic down 60% and Sunedison Semiconductor down 87% prior to the recent spike up on news of strategic alternatives (now just down 75%)

While things were looking up in 2Q’15 with stable/slightly increasing prices among small customers, the market soured quickly in 2H’2015 and demand weakened substantially as customers decided to run down inventories. As a result, demand weakened and pricing came under renewed pressure from (drumdroll) Siltronic (SUMCO does not believe Siltronic gained any share as they matched the price aggression, and is skeptical they will lower prices any further). Whether it was a poor decision by management or a calculated move to pressure a weak #4 player, it was met with an equal response from competitors. One would hope basic understanding of nash equilibriums will prevail in repeated quarters. We know SEMI probably isn’t the aggressor, as the CEO specifically stated in Q1 “broad improvement will require the price leaders to implement price increases across a larger customer base, and others such as us, will follow.”

[rising inventory days]

…in hindsight, the levels of inventory build probably should have signaled some caution at equity prices that were already pricing in price increases in 2016…

Sumco released full year results of 237 billion yen and 29.4 billion in EBIT (12.4%) and guided to Q1’16 revenue of 51 billion yen and 3 billion operating income (6% EBIT).  In contrast, Q1’15 saw 60.4 billion yen (down 20% YoY) and 8.85 billion yen (14.6% EBIT margins), implying incremental margins in the ~50% range. This guide reflects the Company’s belief that epitaxial wafer pricing is going to fall 5%-10% in Q1. SUMCO believes the inventory correction will largely be complete in Q2 and demand will begin to recover in 2H. As a result, we think EBITDA margins for all the players might be down in the 8%-10% range at least for the 1^st half of 2016 before recovering in the second half:

[SUMCO guidance]                

We don’t have a particular view on when the market will turn, but we think demand will return at some point, capacity will stay the same as it has over the past decade, and pricing will inevitably follow as supply constraints are hit.

Simple Valuation

Excluding hedges, Siltronic was doing ~20% EBITDA margins for 2015. Assuming prices revert over the next 1-3 years, you will have something like $180-$200 million in EBITDA, less 8% revenues for capex, or perhaps $110-$120 million in EBIT. On today’s enterprise value of $300 million, you are paying <3x EBIT. Adding in pensions of $290 million is no fun, but brings the ratio up to 5.1x.[4] This excludes any cash build as prices recover. I think ~5x seems a little too cheap. At 8x EBIT, the shares should be worth $26-$27. If prices don’t recover, the Company is trading at ~5x EBITDA and 29x EBIT, though I presume capex across the industry would take another significant leg down.

But then there comes all the optionality. If price increases that were expected in 2016 is simply postponed until 2017/2018, and we get a 5-10% increase in prices to encourage new capital investments, then EBITDA margins could expand to 25%-30%. At 25% EBITDA margin, Siltronic is trading at 3.6x EBIT (including pensions) and a 8x multiple would imply a share price worth $41. Then there the potential merger with Sunedison. If Siltronic bids $8 with equity (most dilutive, but safest form of financing), it will have 48.5 million shares outstanding, -45 million of net debt and ~340 million of pension liabilities. Combined revenues will be $1.7-$1.8 billion euros, and the stock will be worth mid $40’s (assuming Sunedison hits its internal target margins, and assuming no additional synergies). Baupost also owns 20% of SEMI and 9% of WAF.

I think you can get creative dreaming about the upside, but the main point is the current prices seem to be giving you some good long term optionality.

The Risks

·         Unless a merger takes place, Siltronic will always be operating less efficiently than peers: Because of scale, SUMCO and Shin-etsu have historically generated higher margins and captured a disproportionate share of industry profits. However, the gap has been closing as Sunedison and Siltronic have been re-focused on optimizing “footprint” and reducing costs (Sitronic has closed various production sites since 2012 and adjusted capacity downwards. SEMI is in the process of closing their Ipoh plant in Malaysia). Siltronic’s EBITDA margins, adjusted for hedge losses, were roughly equivalent to SUMCO’s margins in 2015. Regardless, SUMCO and Shin-etsu have larger plants and therefore will benefit from economies of scale:

o    Shin-etsu has 4 facilities in Japan and operations in Malaysia. The Shirakawa plant is gigantic and has 800k capacity alone[5]. This comes out in their margins (EBIT >15% in FY2015 vs. SUMCO at 11%, and WAF at ~5%)

o    SUMCO has large facilities in its Kyusha and Yonezawa plants (400k/300k capacity[6]) and closed its Izuno plant in 2012/13.

o    SEMI has its 200/300mm facilities across 4 facilities (Missouri, Japan, SK, Taiwan), but also has 3 other small diameter facilities

o    Siltronic has majority of operations across 4 facilities as well (below). The size of its 300mm plants are roughly in line with SUMCOs. I imagine a merger with SEMI could lead to utilization of the shell capacity to capture efficiencies

 [Siltronic fab overview]

·         Perpetual pricing pressure: Prices never increase and has declined by ~12% a year between 2011-2014, so a price hike would be a first in the 300m space. Per MS, “There were instances in the past when wafer makers tried to carry out price hikes, but related announcements came amid a worsening supply-and-demand balance and price hikes were not achieved.” This recurrence of this pattern might suggest (if not bad luck) that the players aren’t truly as rational as they should be in an oligopolistic market. However, in the past there have never been instances where demand has exceeded supply

[Price trend]

·         Weak demand: If demand for wafers is relatively weak (2-3% a year), there is perhaps enough debottlenecking/shadow capacity to move supply to meet demand. Growth in 300mm is projected to grow in the high single digits however, and so it doesn’t appear likely that debottlenecking will be enough to meet all this demand

·         450mm process: The 450mm process is already long overdue (perhaps as a result of the over-expansion of 300mm), but the earliest experts believe the transition will begin is probably in the 2020’s.

·         Process nodes continue to diminish at a faster rate than smartphone and other integrated circuit growth. Historically, per HSBC, correlation of global GDP growth to wafer shipments is .77. I think more and more ways to incorporate electronics into everyday use (i.e. Amazon’s weird flash buttons) can probably offset production efficiencies.

·         Operating profit could collapse in a severe recession. SUMCO in 2010 saw sales plummet 44% and operating margins hitting -40% before rebounding back to ~breakeven in 2011. A re-occurrence of a global economic collapse could eat away all of Siltronic’s excess cash if margins fall as dramatically. There doesn’t seem to be any real mitigant to this, though SUMCO’s margins back then might not be the best proxy as the markets might have just 3 players this time around, prices are 20% of wafer prices in 2008 already (kind of like how much lower can oil go form $30) so it’s unlikely to collapse another 50%+, and SUMCO was also in the solar polysilicon business in a large way prior to the recession.

Sunedison in Play

Sunedison Semiconductor looked particularly cheap towards end of last year (before Siltronic and SUMCO fell in tandem) especially as it has internal levers to pull EBITDA margins up from ~10% to a 15-20% (lower end from closing of facilities, upper end from a little bit of capacity tightening). However, as ASPs fell quite rapidly in Q4, the flow through to margins will inevitably reduce EBITDA substantially in Q4 and going into 1H’16.

What seemed like a pretty manageable debt load of 2x EBITDA became instantly likely to break maintenance covenants (2.0x total leverage). You can tell from the 8-K filings (refinancing with Korean banks) that the Company is aware of their debt situation, and I presume the fear of equity dilution contributed to the strong underperformance of the stock.

On Friday September 19^th, Sunedison reported that it had received “unsolicited preliminary indications of interest” and the board authorized the Company to hire Barclays to explore strategic alternatives. The stock jumped 51% and ended the day at $5.57.

Management has stated on numerous occasions that the industry needs to further consolidate. We think the logical acquirer here is Siltronic, simply because the combined companies will be #1) equal in size to SUMCO and Shin-Etsu, 2) will likely be able to realize synergies and move towards margins closer to peers, and #3) WAF’s management has basically stated it will be at least involved in any potential discussions.

Either way, I think Siltronic can’t ignore Sunedison being in play and a combination of the two would be the ideal outcome for both WAF and SEMI equity owners. If private equity takes over, the business will probably be much more focused on price (not to talk about minimizing capex spend). If a Chinese company takes over, that could mean trouble long term, but one would hope the oligopoly players would realize the threat and make the bidding process more interesting.

It will be interesting to see where things “shake” out.

Catalyst

Potential merger talks flourish with SEMI

Demand returns in 2H'15