Disclosure: This is not financial advice; always do your own homework. Some of us own this.
Recently we wrote about using 13F reports to find new investment ideas (where investment managers who manage over $100MM in public securities must file with the SEC). As we noted, Pabrai Funds recently took a large new position in Graftech International Ltd. (large being ~10% of reporting holdings, where Mohnish Pabrai is very concentrated and puts in a maximum of 10% into his highest conviction ideas…). We have a big soft spot for him – but also his record - and so of course this piqued our interest. It’s had a bit of a run in the last few weeks but it is generally in line with Mohnish’s likely cost base (see original article - http://www.canadianvalueinvestors.com/home/2019/9/5/cloningthose13fs ).
Market cap of ~$3.93B as of this date.
“So what would you say you do here?”
Graftech describes itself as “a leading manufacturer of high-quality graphite electrode products essential to the production of electric arc furnace (or EAF) steel and other ferrous and non-ferrous metals.”
When you first hear about this business, you likely feel like one of the Bobs from Office Space, but thankfully there’s a lot more substance to what they do than that angry bumbling Initech employee (pure gold https://youtu.be/m4OvQIGDg4I ).
Here’s the gist – They are one of the largest producers of a niche product, graphite electrodes, that is essential to make EAF steel. They also are also the only vertically integrated manufacturer that also makes the main input to graphite electrodes, petroleum needle coke. The Company was taken private by Brookfield Asset Management Inc. and put through an efficiency/capacity focused reorganization. Now, after going public (again) in 2018, they’re trading at ~5x cash flow and still majority-owned by Brookfield.
Now before we get too far along into the Company, let’s talk about the industry.
Making that Steel - Industry overview
There are 2 types of steel manufacturing plants: 1) basic oxygen furnace (BOF - steel from scratch/smelting iron), and 2) electric arc furnace (EAF - from recycled steel). Generally, once this steel reaches its end of its useful life it is then recycled into new steel in EAF plants, which tend to have lower capital and operating costs but also require more specialized materials (below). BOF plants rely on virgin iron ore and metallurgical coke (from metallurgical coal). While in the EAF method, scrap steel is melted and recycled.
In developing countries, BOF steel is produced primarily given lack of recycled steel available, while there continues to be an ongoing global trend (two decades plus) of transitioning towards EAF steel (approximately 45% of steel production globally, excluding China which primarily uses BOF for now). There are several benefits to EAF steel, including 1) it allows steel to be made from 100% scrap metal, 2) requires less energy input 3) allows for variable production vs a BOF furnace, making it easier for the plant to match actual demand.
Here’s a visual (sales pitch) for an Electronic Arc Furnace - https://youtu.be/8nbsPXqr_4A
How Graphite Fits In
To make steel, you need graphite electrodes. These are put into EAF furnaces and act as a conductor to shoot electricity into the furnace, creating heat to melt scrap metal. Graphite electrodes are manufactured from petroleum needle coke and adhesives, where they are assembled into columns and sold to steel manufacturers for use in furnaces.
There are generally 2 types of graphite electrodes ("GEs"): 1) ladle GEs, and 2) ultra high powered (UHP) GEs.
Ladle GEs, which are used to maintain temperature of liquid steel in BOF plants, are made from pitch needle coke which is a by-product of coal/coke production. It is relatively easy to make pitch needle coke and ladle GEs.
UHP GEs, which are used to melt scrap steel in EAF plants, are made from petroleum needle coke (PNC) which is a by-product of a certain type of oil refining. PNC and UHP GEs are 1) technically difficult to make; specific types of heavy low-sulfur oil must be used, and then the given refinery (which typically focus on gasoline/diesel) must configure its operation to produce the "right" type of bottom-of-the-barrel by-product that can then be used to create PNC; and 2) time-consuming to make, with PNC taking 3 months and UHP GEs another 3 months. Phillips 66 is by far the #1 PNC producer with ~50% of the ex-China market.
Big picture – “In a typical furnace using alternating electric current and operating at a typical number of production cycles per day, three electrodes are fully consumed (requiring the addition of new electrodes), on average, every 8 to 10 operating hours. Graphite electrodes are consumed at a rate of approximately 1.7 kilograms per MT of steel production.” (2018 Graftech prospectus)
The industry generally refers to ex-China PNC/UHP GE production because information from China is opaque, and it is generally thought that China's GE producers do not currently have the required technical expertise and therefore have essentially nil UHP GE production (even though they market their GEs as UHP). Regarding technical expertise, Showa Denko (a long-time Japanese UHP GE producer, and #3 in the industry) owns the only 2 new green and brownfield ex-China UHP GE plants that have been constructed in the last few decades, and after 10 years have not yet fully ramped up because of ongoing operational challenges.
PNC has historically been used mostly for UHP GEs but in the last few years, electric vehicle battery producers have started using PNC to improve battery performance. EV/battery demand now makes up ~10% of PNC demand, up from essentially nil a few years ago.
In the early 2010s, China's BOF steel was dumped into the European/North American markets, causing significant declines in UHP GE and PNC pricing. This seems unlikely to continue given a few big changes, namely China rationalizing steel production (sources behind paywall, but here’s a semi-public analysis - https://www.economist.com/leaders/2017/09/09/making-sense-of-capacity-cuts-in-china )and trade tariffs. From 2014-2016, ~20% of the UHP GE capacity was permanently removed (from ~1m metric tons to ~800k). The UHP GE industry has increased capacity by ~5-10% since 2016, primarily through debottlenecking projects; brown/greenfield projects have a long lead-time, are capital intensive, and also require secure PNC supply which has been difficult to obtain irrespective of price. As a result, there are no new UHP GE brown/greenfield projects currently contemplated.
Historically this has been a tough business. However, this is the Company’s PR department’s thoughts have been summarized in a nice one-page table:
Source: 2018 Prospectus
GrafTech makes UHP GEs and is the only UHP GE manufacturer that has its own supply of petroleum needle coke. After Phillips 66 (~50% of market), GrafTech (via its Seadrift subsidiary) is #2 (~20% of market) in PNC production capacity (ex-China), but GrafTech produces enough PNC to support ~3/4 of its UHP GE production capacity, which is ~20% of the UHP GE industry capacity (ex-China).
Since 1) PNC is generally in short supply due to (a) general trend towards EAF steel as countries develop, (b) low production capacity additions due to the technical difficulties and relatively small market size vs gasoline/diesel refined products, and (c) emerging EV battery demand, and 2) GrafTech is the only vertically integrated UHP GE producer. This has all come together to enable GrafTech to sign 3-5 year take-or-pay contracts with many EAF customers. No other company in the industry can do/has done this, as they do not have a guaranteed supply of PNC. GrafTech, if operated effectively, should also have a low-cost advantage due to its vertical integration.
A tier 1 EAF plant generally requires a constant supply of UHP GEs. UHP GEs are consumed in 8-10 hours and without them, the EAF plant may have to shut down which is extremely expensive (estimated at $1k/minute – a bit more expensive than a good lawyer’s billable rate). Yet, UHP GE prices make up <5% of an EAF plant's total costs. In 2H 2017, EAF producers were paying ~$25-30k/ton for UHP GEs in anticipation of a potential shortage. UHP GE spot prices are now currently ~$11-12k/ton.
The Company Turnaround
Coupled with these structural changes comes a turnaround done by Brookfield. Here’s what they have done in their own words (to be taken with a rounded spoonful of salt of course):
”We have achieved annual fixed manufacturing cost improvements and capital expenditure reductions of approximately $190 million since 2012, while also improving the productivity of our plant network We have strategically shifted production from our lowest to our highest production capacity facilities to increase fixed cost absorption. In 2018, we expect to produce a greater quantity of graphite electrodes from our three operating facilities in Calais, France, Pamplona, Spain and Monterrey, Mexico, than we did from our six operating facilities in 2012. As a result, we have achieved significant operating leverage at higher capacity utilizations. In our experience, high capacity manufacturing facilities can have operating costs of more than $1,000 per MT lower than low capacity manufacturing facilities. In addition, we have streamlined fixed costs across our plant network, including a 50% headcount reduction at Seadrift since 2014 and an optimization of Seadrift's systems and manufacturing process to reduce capital expenditure requirements. As a result of these actions, by the end of 2016, we had reduced our annual fixed manufacturing costs by approximately $80 million and our maintenance capital expenditure requirements by approximately $45 million since 2012.”
The Recap – Where we at with that thesis?
Here’s a few highlights:
Lollapalooza - Essential product, small part of cost of overall production, high risk from bad product
Graphite electrodes typically represent less than 5% of the total production cost of steel in a typical EAF, but they are essential to EAF steel production. Graphite electrodes are currently the only known commercially available products that have the high levels of electrical conductivity and the capability to sustain the high levels of heat generated in EAF steel production.
Concentrated producers, diverse customers
The graphite electrode and petroleum needle coke producer markets are relatively concentrated, while the customer bases are diverse (Graftech themselves have over 100 customers).
Interestingly, when Graftech bought its (now) petroleum coke subsidiary, it resulted in an antitrust complaint by the Department of Justice, worth reading here - https://www.justice.gov/atr/case-document/file/497821/download
As per the Company’s prospectus documents (i.e. not us) – “Significant amounts of graphite electrode industry production capacity have recently been removed from the market globally. We estimate that approximately 20% of industry production capacity (excluding China) has been closed or repurposed since the beginning of 2014. Some of these closed manufacturing facilities have sold off equipment, been demolished, undertaken long-term environmental remediation or been repurposed for other manufacturing uses. Accordingly, we believe the majority of these closures represent permanent reductions. As part of this overall industry rationalization, we permanently shut down two plants and temporarily idled our St. Marys plant, reducing our electrode manufacturing from six operating facilities in 2012 to three operating facilities in 2017. Also, in October 2017, the third largest graphite electrode producer acquired the second largest producer.”
While this could be considered Company marketing mumbo-jumbo, it has translated into truly higher prices that are locked into long-term contracts. There’s nothing like a graph to really drive home the step change in market prices. Shown in the below now-a-bit-dated-chart, this isn’t the graphite electrode market of the 90s.
The Company continues to push take-or-pay agreements, which they can do now given the tight market and their apparent ability to deliver a consistent high quality product.
Company structural change
Analyzing this requires a bit of Kool-Aid and your own numbers.
The Awkward Public Offering (and now what do you do with all that cash?)
When the Company went public in 2018, Brookfield sold approximately 20% as part of the go-public offering (and still owns the rest).
Now this is awkward, as the Company is “targeting ~50 –60% of 2019 free cash flow to be returned to shareholders. The Board has authorized open market repurchases of up to $100M of common stock or ~15% of public float”. So the Company is authorized to buy back part of a relatively small float, being about 20% of it. Now we would coffee bet that Brookfield intended to sell more over time, but why sell when a Company is doing great and you don’t have to sell? But then what do you do as a management team…?
We won’t be doing a full valuation here – we have to leave some fun for you (and we also don’t provide financial advice). But here’s a bit of data. GrafTech is currently producing ~$750-800 million of free cash flow before $125 million annual debt amortization (vs market cap that was ~$3.3b and now ~$4B + $2b of debt), of which ~3/4 are supported by contracts through to 2022. The remaining spot/short-term sales are supported by third-party PNC purchases. GrafTech will be negotiating new/extensions to its contracts later this year. As the only producer that can guarantee its EAF customers UHP GE supply (by way of its guaranteed PNC feedstock - assuming its PNC manufacturing facility does not have a failure) it isn't unreasonable to assume they can extend these contracts at or near the current contracted prices of ~$10k/ton, which are ~15% below current UHP GE spot prices.
Graftech Aug 9 2019 – Form 424B4 https://www.graftech.com/investors/reporting/sec-filings-details/default.aspx?FilingId=12907318
Anti-trust complaint - https://www.justice.gov/atr/case-document/file/497821/download
EVs and Needle Coke - https://www.mining.com/web/imo-2020-evs-steel-perfect-storm-needle-coke-sector/
Never forget to go to the next level - https://www.quora.com/What-is-decant-oil-used-for
An EAF in action - https://www.youtube.com/watch?v=_j2jESz7Zl8