GrafTech: Contracted Cash Flows Covering 60% of the Market Cap

Written by Miguel Neto. 

 

There are some pictures I couldn’t put in this write-up that I included in the write-up I put up on my blog (https://netosnotes.blogspot.com/2019/09/graftech.html)

*****

 

GrafTech (ticker: EAF) has been in operation since 1886. It went public in 2018 when Brookfield sold 15% of the business, after being taken private in 2015. GrafTech is a backward vertically integrated producer of ultra-high power (“UHP”) graphite electrodes (“GE”), an industrial consumable product used primarily in EAF (“electric arc furnace”) steel production. Around 46% of the world’s steel production is produced through the EAF method, the other 54% being through the BOF (“basic oxygen furnace”). These numbers exclude China, where the BOF accounts for 91% of steel production.

 

GE’s are an industry consumable used to conduct electricity in a furnace, generating an electric arc (a lot of heat) to melt scrap metal to the point where it is a liquid. UHP GE are made of >95% needle coke, whereas ladle electrodes, which are used in BOF’s, are made up of 20-30% needle coke. Ladle electrodes are of lesser quality because they only need to maintain the scrap metal in a liquid, as opposed to UHP GE which actually need to melt the metal. Furthermore, UHP GE takes around 6 months to produce, despite being used up in a single 8-10hr shift. Most EAF’s use 3 electrodes.

Graphite is used because it is the only material that has the chemical properties to handle a current consistently at temperatures needed to produce steel in an EAF – thus there it cannot be replaced by another product. Even at today’s prices GE only accounts for 1-5% of the production cost of steel. This bodes well for GrafTech as customers will be far more concerned with the (1) quality of the product and (2) security of supply, rather than price.  Lower quality products lead to higher rates of breakage, more production downtime and thus increases production costs.

Worldwide capacity for GE is about 850,000 MT per year. Demand for steel is about 930 million MT per year. Assuming around half of that is made through the EAF, around 1.7kg of GE is used per metric ton of steel. Outside of China, only one greenfield and one brownfield project have been completed since the 1980s.

GrafTech makes UHP GE in 3 plants (+1 idled). There are 6 producers of GE worldwide and GrafTech is the 1^st or 2^nd largest in the world, with about 25% share. The US’s dominant method of making steel is through the EAF method, as opposed to globally where the BOF method is still dominant. EAF pollutes a quarter of BOF and is a simpler process, incurring less overhead.

Today the GE market is operating at 100% capacity, after companies closed or repurposed facilities. GrafTech produces more from their 3 facilities (plus one which is idle) than they did from their 6 facilities a few years ago.

Industry Background & Recent Events

• Petroleum Coke

Pricing for GE has historically been cyclical, reflecting the demand trends of EAF steelmaking industry and supply of GE. Petroleum needle coke represents a significant percentage of raw material cost of GE, and therefore GE has typically been priced at a $3,000 spread to petroleum needle coke. Between 2008 and 2017, the weighted average. price of GE was $4,500 per MT. During the most recent demand trough, in 2016, the price was $2,500 per MT. As of Feb 2019, spot prices were approx. $12,000 per MT.

Why will the spread persist and why is it different this time?

The spread will persist for at least the next 5 years because previously petroleum needle coke was the primary input for GE and that was its use. Today however, petroleum coke is also used to make lithium-ion batteries thus limiting the supply of petroleum coke needed for GE’s and therefore potentially limiting steelmakers ability to produce steel.

• Steel Industry

First, the steel industry when through somewhat of a crisis beginning in 2014. EAF steel production grew at 3.5% per year from 1984 to 2011, encouraging growth of GE capacity. Chinese steel production, through BOF, created an oversupply in steel which led GE producers to rationalize production capacity and consolidate between 2014 and 2016.

According to the company’s filings “In 1984 China produced 21 million MT of BOF steel, which by 2017, had grown to 754 million MT [91% of its steel production]. “Growth in capacity surpassed growth in demand, resulting in excess capacity within China and increased exports into global markets. China net steel exports peaked at 112 million MT in 2015.”

In 2015 China began restructuring its steel industry by encouraging consolidation and shutting down excess capacity. Implemented environmental regulations thus reducing BOF steelmaking. North America and Western Europe have implemented trade decisions against BOF steel-producing countries to protect their domestic steel industries against imports. Furthermore, there has been a long-term trend whereby EAF has accounted for an increasingly bigger proportion of steelmaking, as can be seen by the following table which represents global steel making by method:

	2000	2017
EAF	32%	46%
BOF	68%	54%

 

• Graphite Electrode Industry

Second, and as alluded to before, the GE industry went through a rationalization. In 2013 the GE industry (ex. China) had capacity for 1 million MT across 30 plants. From 2014 to 2016, the industry repurposed approximately 20% of global production capacity (ex. China) consisting of smaller, higher cost facilities. Additionally, in 2017 Showa Denko (3^rd largest producer) acquired SGL Carbon (2^nd largest producer). High capacity manufacturing facilities can have operating costs that are a lot lower than low capacity manufacturing facilities (fixed costs are spread, resulting in operating leverage), encouraging producers to consolidate to reduce costs. The majority of production capacity reduction was permanent due to the demolition, long-term environmental remediation and repurposing of most of these lower capacity facilities.

Petroleum Coke Industry: New Market

Today there is about 750,000 MT of petroleum needle coke made per year. Philipps 66 accounts for 56% of that, and Seadrift accounts for 18%. Of the total demand for petroleum coke, 90% goes into making graphite electrodes, however in the last few years another huge and growing end-market has been created for petroleum coke: electric vehicles. Petroleum coke is used to make lithium-ion batteries and already accounts for 10% of demand of this material. This is up from basically zero a few years ago, and demand is growing very quickly. According to BMO Capital Markets, lithium demand for EVs grew 33% from 2017 to 2018. Lithium demand for EVs accounts for about 33% of total lithium demand as well. There are about 5M electric vehicles worldwide, and EVs sales account for about ~1.5% of vehicle sales in the U.S.

This secondary market is set to put pressure on petroleum coke demand and thus increase petroleum coke prices and GE prices (as petroleum coke accounts for a large percentage of the raw material cost to make GE), allowing GrafTech to sell GE years in advance to steelmakers who want to make sure they have the essential raw materials to make steel.

So, the primary raw material for (high quality) GE is needle coke. There are two kinds of needle coke:

1. Petroleum needle coke (“Pet coke”) – byproduct of a refinery.
2. Pitch needle coke (“Pitch coke”) – made from coal tar pitch, a byproduct of cooking metallurgical coal

Why is it hard to increase pet coke supply?

Making pet coke: In a refinery the oil undergoes two processes. First, the oil is distilled into various products, separating the light parts of the oil (gasoline vapors, etc.) from the heavy parts. The heavy oil is then processed through a cooker to make pet coke – this is what Seadrift does and where the vertical integration comes from. Heavy fuel oil accounts for about <10% of a WTI crude oil barrel. Then, you take the heavy oil and put it through a cooker, which creates various kinds of pet cokes. About 80% of which are fuel-grade¹ and therefore not good enough to make electrodes.

Overall, pet coke is made from ~2% of the a WTI crude oil barrel. Even if pet coke prices go up materially, as they have, the refinery is not going to make capacity expansion decisions based on 2% of a barrel’s content. There are five producers of pet coke globally and one of those (Seadrift) was bought by GrafTech in 2010. No pet coke capacity has been added in the last 10 years, outside of China.

Pet coke, not pitch coke, is essential to make UHP electrodes. It can take about 2 months to manufacture pet coke. 

So, debottlenecking doesn’t pose a big risk given that it’s 2% of the output of a refinery, and capacity expansion from Chinese producers – which produce sub-par anodes – is not good enough for the UHP GE steelmakers need. Furthermore, the vast majority of new capacity that has been announced in China has come from new players, with limited track record producing pet coke – much less the quality needed for UHP electrodes. GE are cheap relative to the total cost of making steel, so the quality and production line are not going to be sacrificed because of a cheaper price from competitors. Even so, GrafTech estimates it will take approximately 5 years for meaningful needle coke capacity to come out of China.

A further sign of how strained the pet coke industry is, is shown by the fact that pet coke producers used to agree to supply pet coke in 12-month contracts, in 2017 they changed to 6-month contracts and now they’re doing 3-month contracts. As a result, competitors – who need to buy pet coke to make high quality GE – must continually renegotiate supply agreements through changing market conditions

What about GE supply? Notwithstanding adding pet coke supply, why would it be hard to replicate GrafTech’s GE plants?

Greenfield GE manufacturing projects have been difficult to develop due to significant capital costs, long lead times, technical know-how, and tough regulatory regimes. Management estimates the time from initial permitting to full production would be >5yrs and cost as much as $10,000 per MT.

So, in the meantime, GrafTech is the only vertically integrated GE producer worldwide, because it owns Seadrift. The chart below shows GrafTech’s advantage of being vertically integrated, because whilst all other competitors have to buy pet coke on 6-month or shorter contracts – therefore being susceptible to rising prices – GrafTech doesn’t, to a large extent, because Seadrift supplies 70% of its pet coke needs. GrafTech’s vertically integrated total cash cost per ton is less than what competitors are paying just for pet needle coke.

Seadrift supplies 70% of its pet coke needs. GrafTech’s vertically integrated total cash cost per ton is less than what competitors are paying just for pet needle coke.

There are only 5 global facilities that produce the quality needle coke needed for UHP. There is a new market and there is no easy way to add supply, and this is why spot prices have risen so dramatically.

Take-or-pay Contracts

Now, to really prove that this business, temporarily, is not commodity, in 2018 management came up with these new take-or-pay contracts. Due to the shortage of UHP GE’s steel makers have been happy to sign these contracts to guarantee supply of GE (so they can make steel) five years out (2018 to 2022). And management has been able to secure this supply at prices that represent +65% EBITDA margins.

Management has actually secured ~65% of their cumulative capacity (which is around 674,000 MT) from 2018 to 2022. Since they sold 133,000 MT in 2018, there’s 541,000 MT in contracts left. On average, the contracts are secured at $9,700 per MT. This represents $5.2bn in revenue and around $3.4bn in EBITDA.

Given that GrafTech is vertically integrated they are in the unique position of being able to secure supply for the steelmakers at pre-determined prices, because they had (1) a secure supply of pet coke, and (2) have hedged the cost of heavy oil (primary raw material for pet coke production). The probability that they’re not going to be able to actually supply what is contracted is low, given that they have limited the contracts to an amount that can be wholly supplied with Seadrift’s pet coke.

Another positive for GrafTech is that there isn’t much concentration of customer purchases in the contracts, with no customer accounting for more than 8% of the contracted capacity. Furthermore, per the company’s 10-K, the contracts are such that in the event the customer wants to cancel the contracts, termination fees range from 50-70% of the remaining contract revenue:

“In addition to defining annual volumes and prices, these three‑ to five‑year take‑or‑pay contracts include significant termination payments (typically, 50% to 70% of remaining contracted revenue)”

 Debt

First and foremost, the company must be able to service debt: In 2018 the company got into a loan for $1.5bn, which has an interest rate of 6% and matures in 2025. Furthermore, the loan amortizes at 5% per year. However, there was a promissory note with a maturity of 8 years that totaled $750M to Brookfield. The company opted to increase the loan a further $750mn to pay for the Brookfield note. So, now there’s $2.25bn in debt, which translates to around $100mn in interest expense and $110mn in principal amortization.

With $200mn in cash and a backlog of $5.2bn which would generate around $3.4bn in EBITDA, I think the company will be able to service their debt.

In terms of the other 35% of capacity, I believe that if prices remain above $3,000 the company will very likely be profitable.

Valuation

So, on the contracted part you have $5.2bn in revenue and around $3.2bn in discounted EBIT. Furthermore, if I assume the company runs at almost full capacity (as it has), then there will be room for 179,000 MT of cumulative capacity for uncontracted business from 2019 to 2022. This represents another $1.8bn in revenue and some $940mm in discounted EBIT. Therefore, in 2023, the company will have produced $4.1bn in EBIT.  That equates to approximately $2.9bn in discounted unlevered FCF, or three-quarters of the current market cap.

As for 2019, the company has take-or-pay agreements for 148,000MT for 2019 at an average price of $9,800, and has capacity for 180,000 which has been running at ~99% utilization. Assuming 95% utilization on that and an average price of $10,000 per ton, there’s $1.7bn in revenue and $760mm in FCF. Commodity businesses like steel companies trade for >4x EBITDA, whilst, commodity companies with more earnings visibility trade at >5x EBITDA and around 9-10x FCF. So, a 10x multiple on FCF gets you to a value of $7.6bn. Or $26 per share. The company is trading at 5x 2019E FCF.

Risks

• In the past China has be known to oversupply certain industries, and then dump the oversupply globally by exporting. However, when it comes to the UHP GE this is more unlikely as China doesn’t produce these, but instead produces sub-par ladle electrodes, which are not a substitute. Domestically, China doesn’t have access to high quality pet coke. The risk is further mitigated by the lack of actual facilities, technical know-how to produce UHP GE and pet coke, and the time it would take to build these facilities and run them at full capacity (5 to 10yrs).
• Decrease in the spot price of UHP GE. ~35% of GrafTech’s cumulative capacity until 2022 is subject to this risk, since this capacity will probably be sold at spot prices.
• Brookfield dilution by selling more of their 85% stake. 

Conclusion

In terms of the graphite electrode industry we can sum it up by pointing out that there are 5 producers of graphite electrodes worldwide. Only 4 or so produce the kind of UHP graphite electrodes that GrafTech makes. It would take more or less 5 years to build more graphite electrode capacity, and the supply from this would be incremental, i.e. the plant wouldn’t be operating at 100% capacity in the first year of operation.

Secondly, we can point out that the main product for making GE is petroleum coke. In this industry too there are only 4 major players, with Phillips 66 accounting for 56%. Petroleum coke is only something like 2% of the output of a refinery. So more refineries, refinery debottlenecking and capacity expansion from Chinese players which produce subpar anodes is not a threat for now. And again, it would take something like 5 years to build a petroleum coke plant.

With all these factors in mind, what we can sum it is that it’s quite hard to add capacity immediately.

Because petroleum coke now has a potentially new and huge market (EVs), and thus its supply is constrained and will probably be constrained for the next 5 years, and because that will have an impact on the supply of GE, GrafTech went to their customers and offered them 5 year take-or-pay contracts so as to allow steel companies to make sure they can produce steel. Also, GrafTech is the only company that could do this at pre-determined prices because they’re the only backward-integrated company in this industry, thus controlling the cost of their raw material. Competitors may have to pay very market prices (high prices) for petroleum coke. Steel companies were eager to secure their supply of graphite electrodes so as to not run into the risk of halting their production – it surely helped that these electrodes are essential for steel production, yet only account for less than 5% of the total production cost of steel.

These contracts allowed them to sell two-thirds of their 5-year capacity, the remainder will be sold at spot prices.