Description
Neo Materials can be bought for less than 3x 2007 levered free cash flow. I expect free cash flow to grow at least 10% per year through 2009. We owe this opportunity to 1) uncertainty regarding a major patent expiration at the end of 2009, and 2) significant selling pressure from pre-reverse merger shareholders, in addition to basic small cap neglect and, under-promise over-deliver management. I believe Neo shares could at least double from these levels with negligible downside risk.
Background
Neo’s stock price is quoted in Canadian dollars, but the financials are reported in US dollars. Neo processes rare earth metals into intermediate products, principally for use in the electronics industry. Neo is a product of a reverse merger between AMR Technologies, the public predecessor, and Magnequench, an LBO led by Archie Cox that was spun out of GM in 1995. The successor is the only integrated rare earths processor.
Industry
Rare earths are naturally occurring elements (metals). There is no formal forward market for rare earths. Prices can be found at
http://www.metal-pages.com/ and
http://www.asianmetal.com/. Bloomberg doesn’t appear to track rare earth prices. Roskill writes the most respected research on the industry.
Despite their name, rare earths are quite abundant. However, at present China holds 90% of the world’s proved reserves and produces 95% of the world’s output of rare earths. Since bottoming in Q4 2002 Neodymium oxide, the most important rare earth intermediate, has more than quadrupled in price. This happened for two reasons: 1) rapidly growing demand, 2) inelastic supply coupled with Chinese mine production quota cuts. Virtually all of the rare earths have seen dramatic price increases since 2003, but none as severe as Neodymium. Because rare earths always occur together, there is perpetual oversupply of some of the metals as production is managed to try to keep up with demand for Neodymium. All this said Neo is effectively short Neodymium prices as it is a net consumer of them, and is thus not benefiting from high rare earth prices.
I see no fundamental catalysts for rare earth prices to come down. After the Chinese production quota cut this year it appears demand likely exceeded production at beginning of year prices, with little ability to change that dynamic in the short term except through prices.
Magnequench
Magnequench makes magnetic powders from Neodymium oxide. The magnets made from Magnequench’s powders are used primarily in electronics, but other applications are rapidly catching on, such as motors used in automotive interiors and air conditioners. All of Magnequench’s end markets are growth markets, some more than others.
All hard disk drives use neo magnets, and with the exception of counterfeits, all are made with Magnequench powders. So, the hard disk drive component will grow with HDD industry unit growth, which is actually respectable (~10% YTD). While HDDs in smaller applications are losing share to flash drives, flash is still far from economical in large drives where growth is being driven by the obvious multimedia applications. Legacy applications make up roughly two-thirds of Magnequench’s unit sales (HDDs are about 60% of that). In Q3 this component grew 11% YOY. New applications make up the other one-third of units – this component grew 97% YOY in Q3.
Neo magnets are smaller and stronger than traditional ferrite magnets. Trends toward miniaturization will drive market share gains for neo magnets for years to come. High prices for other raw materials is also an impetus for market share gains, as product engineers look to reduce input costs through smaller motor sizes. Management and other industry observers expect Magnequench unit volume to grow by 20% per year through 2010. In the last three quarters units have grown 26%, 19% and 42%. Absent the patent expiration this business would no doubt garner nosebleed multiples.
The primary input for the magnetic powders is neodymium oxide, effectively 40% of which Magnequench gets in-house, the rest it buys from the outside. Investors were understandably concerned about margin compression when Magnequench was lowering its prices 3% per year while Neodymium (currently 35% of cost of goods sold) prices were up almost 70% from Q2 to today. However, effective 10/1 Magnequench raised prices to offset the recent spike in Neodymium prices. Our channel checks suggest that the price increase was on the order of 15% for legacy apps and 10% for new apps. We expect some giveback to larger customers, so blended maybe 10%. Magnequench has 55% gross margins, so 35% x 45% x 70% = 11% Neodymium cost increase relative to revenue, so a 10% price increase should roughly cover the cost increase.
It should be noted that Magnequench uses FIFO, so as of Q3 they were likely working through inventory that was four months old – at prices prior to the 70% run-up – so the price increase won’t be a windfall from Q3, it will truly be a wash for gross profit dollars, though gross profit percentage will obviously get squeezed.
Magnequench consumes 1,300 tonnes of Neodymium oxide per year. Neo’s rare earth processing business produces about 850 tonnes of oxide per year, post the recent capacity expansion. At today’s oxide prices Neo is effectively short 450 x 20,600 = US$9.3 mill of Neodymium in the P&L. It is management’s goal to become Neodymium neutral at some point in the next couple of years, mainly through capacity expansions at AMR and getting exposure to mining.
As a side note, Magnequench will be able to accommodate their projected growth at almost no cost. AMR already owned a plant in Thailand that will double Magnequench’s capacity for a cost of less than $1 million in capex. This should be sufficient through 2010.
AMR
AMR produces rare earth oxides (intermediate products) as well as a few final products. Basically AMR takes a bag of dirt from the mines and separates out all the metals. AMR doesn’t own mines, as it is illegal for foreigners to own Chinese mines and there are limited resources elsewhere. This is a much more commoditized business than Magnequench. AMR accounts for only 20% of EBIT, so I will spend minimal time on it.
Prior to the merger AMR was ugly, primarily because its plants were underutilized and rare earth prices were considerably lower. Magnequench was considered a good fit because it could fill AMRs plants. AMR is now running near 100% capacity, so it was a 1 + 1 = 2.5 merger (if you can forgive the cliché).
In Q4 AMR will commission a 30% capacity expansion at AMR. It’s actually more convoluted than that, but it’s effectively 30%. Management expects to get the new plant up to capacity by FYE 2007.
It is worth noting that AMR is actually comprised of two businesses: rare earth processing and zirconium. The rare earth business is growing units at ~5% per year (excluding capacity growth). The zirconium business produces zirconium chemicals and oxide that are used in a wide range of industrial applications, most prominently automotive catalytic converters. AMR is taking share in the zirconium business and expects to grow unit volumes 15-20% per year. Zirconium also carries an EBIT margin 1.5x the AMR average. Zirconium was about 15% of AMR revenue when last reported in Q2 2005.
AMR is long rare earth prices since it sells them. Its input costs increase because they buy unprocessed rare earth from the mines, but because the finished products are always priced at market, AMR will never get squeezed.
Management
Constantine Karayannopoulos, the new CEO, was previously COO of AMR from 1994. An engineer by trade, Constantine is immersed in the tiniest details of his company but is at best average with the numbers, having only recently been placed in his current role from his previous role which was essentially technical in nature. Importantly, he admits his weakness. I’ve found Constantine to be very self-aware, intellectually honest, and open to others’ opinions, perhaps a product of not coming from a finance background. Constantine, the son of an electrical engineer, appears to derive his passion for the business from the science. Constantine is fond of saying, “the geek factor is very high at Neo.” Constantine has a unique creativity for deals. I expect him to deliver some meaningful deals in the future with minimal capital outlay.
Michael Doolan, the CFO, joined Neo after the merger. He was previously CFO at Falconbridge Ltd, but lost his job after Falconbridge was acquired by Xstrata. Both companies are based in Toronto and both companies sell metal, so it was a natural extension. Michael brings big league finance ideas to Neo. He was the impetus behind reporting return on capital employed in the quarterly reports after the merger.
I believe that both Constantine and Michael understand the value of under promising. As witness, they played up the impending costs of rising Neodymium prices, then subsequently delivered massive sequential growth in Magnequench EBIT, and grew AMR tonnage despite claims that they were running at 100% capacity and volume should decline sequentially. They also gave conference call participants no indication that they were planning to raise prices double digits, devoted one sentence to the topic in the Q3 call, and didn’t disclose the magnitude.
For curiosity’s sake let’s run through the valuation of the Magnequench acquisition:
Common: 27 mill x US$1.50 = 40.5 mill
Warrants: 17.7 mill x US$.20 = 3.5 mill
Options: 2.8 mill x US$1.50 x .3 = 1.25 mill
Debt: 57.5 mill
Convertible Feature of Debt: 22.6 mill x US$.20 = 4.5 mill
Total US$107 mill
$107 mill for a business that will likely produce more than $50 mill of EBIT in 2007. One could argue that maybe AMR got what they paid for because of the patent expiration. But I would contend that assuming the worst case, that Magnequench is worthless post-patent, AMR would make its money back 1.6x before 2010, so worst case it was a pretty average investment. The stock’s gone nowhere since the acquisition, so assuming management added any value, we’re not paying anything for it.
Magnequench Patent
Now to the crux of the matter. I believe that investors are concerned that in 2010 neo magnet powder prices and volumes will plummet on the magnitude of a branded drug post patent expiration and Magnequench’s franchise will be seriously impaired. I don’t think there’s any question the franchise will be impaired but I think the magnitude will be manageable, but as important, as new shareholders I believe that when we pay 3x run-rate free cash flow, there is negligible risk of loss of principal over a three year horizon, as in theory we should at least earn back our principal over that time, and receive residual value via AMR, assuming Magnequench is worthless. But obviously we have to be convinced of more than that.
Magnequench has spent nearly 20 years refining their formulation to achieve maximum performance and consistency. Performance is of utmost concern in most of Magnequench’s applications. Consider hard disk drives. Hard disk drives store some of our most valuable data. It would be a huge headache for consumers and an embarrassment to manufacturers if their drives failed because of a low quality magnet in the reader head. Magnequench has caught counterfeits in the past, and interestingly they’ve generally contained Magnequench powders, as evidence of the difficulty of getting the formula right.
Cost is also relevant. A typical $50 hard disk drive contains less than 3 grams of neo powders. At current market prices (2.4 cents per gram) this costs about 7 cents per drive. If I’m a Seagate product engineer, I’m not going to risk my job to save a couple of cents with No Name Chinese Magnetic Powder, Inc. No Name’s cost structure would also be no lower than Neo’s I might add since all of Neo’s production is already in China.
Lastly, consider that neodymium is already in excess demand, with no hope of increased supply within the next 2 years (with continued excess demand projected even after the scheduled new capacity comes online). The ability to start a new neo magnet business of any size is limited. Neo, on the other hand, has two guaranteed sources of supply. Neo’s partner in the JVs that own the AMR processing plants owns a mine, and in addition to having a commitment to supply rare earths, the partner also has an economic interest in supplying the plants. Another mine in northern China that controls 75% of market supply is also a committed supplier. As part of the quota system in China, export quotas were issued in accordance with each recipient’s export volume. Neo, being China’s largest exporter of rare earths, owns the most export quotas of anyone, and new ones can’t be issued. In exchange for allowing the northern mine to piggyback on Neo’s quotas, the mine owner has agreed to supply two thirds of Neo’s input requirements at below market prices.
Projected Rare Earth Market Supply & Demand
(000s metric tonnes) 2006 2007 2008 2009 2010
Projected Market Supply 87 85 88 90 91
Projected Market Demand 108 115 120 135 150
Source: BCC Research
Model
When modeling Magnequench I’ve chosen to present only the pessimistic view. That is that prices are reduced 6% per year through 2010 to discourage new competition and to drive new application adoption. This was the policy of the previous management team, but is 2x the policy of current management to reduce prices 3% per year (excluding “surcharges”). All this assumes flat neodymium prices. The key takeaway is that unit volumes are growing rapidly enough to offset some level of price erosion. Even assuming a catastrophic price war in 2010, 2010 EBIT could still easily be higher than 2006. In this model I’m frontloading the damage.
I’m arriving at the gross margin by using a back of the envelope index:
2006 2007 2008 2009 2010
Revenue 100 94 88 83 78
COGS 45 45 45 45 45
Gross Margin 55% 52% 49% 46% 42%
Magnequench 2006 2007 2008 2009 2010
Volume (tonnes) 3,945 4,800 5,900 7,000 8,000
Growth 22% 22% 22% 20% 15%
Price/Tonne (FYE) 26.75 25.15 23.65 22.25 20.90
YOY 7.4% -6% -6% -6% -6%
Revenue 98 125 143 162 175
Gross Margin 55% 52% 49% 46% 42%
Gross Profit (mill) 54 65 70 74 74
G&A 11 9 9.5 10 10
EBIT 43 56 60.5 64 64
AMR’s pretty straightforward. In 2007, I’m assuming they fill half of the new plant plus 7% organic growth – same for 2008. Blended market growth thereafter. I’m giving them no credit for gross margin expansion from growth in the zirconium business.
AMR 2006 2007 2008 2009 2010
Volume (tonnes) 6,100 7,450 9,100 9,600 10,100
Growth 28% 22% 22% 6% 5%
Revenue 67 84 102 108 114
Gross Margin 24% 24% 24% 24% 24%
Gross Profit (mill) 16.7 20 24.5 26 27
G&A 7.5 7.2 7.7 7.9 8
EBIT 9.2 12.8 16.8 18.1 19
Corporate 2006 2007 2008 2009 2010
EBIT -9.1 -9.7 -10.2 -10.7 -11.2
Consolidated 2006 2007 2008 2009 2010
Revenue 165 209 246 270 289
EBIT 43 59 67 71.5 71.5
Interest (Cash) -8.8 -5.2 -3.4 -1.4 .7
Pre-Tax 34.2 53.8 63.6 70.1 72.2
Net Inc1 27 43 51 56 58
Add: Dep & Amort2 12 13 13 13 13
Less: Maint Capex 4 4 4 4 4
Free Cash Flow 35 52 60 65 67
1Blended tax rate should be ~20%. Neo has $100 mill of US NOLs, pays ~20% in China & Hong Kong, 36% in Japan, 20% in Singapore. In the past eight quarters the effective tax rate hasn’t been above 24%, and has averaged significantly less than 20%. The tax situation is further complicated by transfer pricing schemes employed to reduce Chinese export taxes.
2About half of the D&A consists of patent amortization.
Capitalization
Basic Shares 79
Convert. Notes 22.6 US$50 mill principal convertible at C$2.50, 10% coupon
Warrants 17.7 Convertible at C$2.50, expire 8/31/08
Options 4.9 Avg strike C$2.38
USD/CNY Debt US$40 mill, 8% interest rate
Valuation
The capitalization is obviously extremely variable based on whether the convertible devices are in or out of the money. Below I take a stab at valuation in various scenarios using the treasury method for the convertible stuff. The calculations are circular so it’s difficult to make them transparent.
Today Base Best Worst
Price (USD) 1.84 4.25 5.50 1.50
FDSO 79 113 115 79
Market Cap 146 480 630 120
Lev. FCF 52 60 63 40
Multiple 2.8x 8x 10x 3x
So it doesn’t seem difficult to get significant upside with minimal downside risk. For the more adventurous, the warrants are freely traded, but they seem kind of expensive. To the best of my knowledge the convertible notes aren’t for sale, sadly.
Catalyst
Drying up of artificial selling pressure, good results