Friday, June 16, 2023

Rare Earths: Lynas

Theres 2 non-China REE producers in the world.

  • MP Materials is expanding their production, so I can't value them.  No point looking at them now.
  • Lynas is more interesting, but not cheap.
Here I look at Lynas


Lynas mines REE ores from its Australian Mt Weld mine and ships the concentrates to its separation plant in Pahang, Malaysia.  From there, Rare Earth Oxides are shipped for sale.  Lynas is the largest supplier of NdPr to Japan, who is a partner (through JARE) with priority rights to the material (p87).

Lynas' Malaysian separation facilities have been controversial from the start:

  • They were awarded by the previous BN government which ruled Malaysia since the 50's, despite strong opposition.  
  • The opposition PH campaigned to close the plant in 2018, but after they won they allowed it to continue, upsetting their electorate.
  • Energy, Science, Technology, Environment and Climate Change Minister Yeo Bee Yin, a staunch Lynas critic, had initially insisted the Australian firm repatriate its waste (slightly radioactive waste from Lanthanum).
  • In early 2019, Lynas announced they would move Lanthanum processing to Australia, building a new site in Kalgoorlie.  In late 2019, the Malaysian Prime Minister announced they can build a permanent disposal facility for their existing waste, instead of having to repatriate it.  
  • In February 2023, PH announced the plant needs to be radioactive free by July.  Later extended to 1st Jan 2024.  After this, no more radioactive waste will be produced.  Hopefully this puts the matter to rest.
  • The license is due to be renewed on March 3, 2026.
Does Lynas sell at market price?  While they mention that they promote "fixed pricing" and "long term contracts" to some customers (p21), their historical ASP varies:

It correlates with the Nd historical price (note that their Financial Year ends 30th June):

So they seem to be selling at market price.  Remember that their ASP is for a "basket" of REs (with Lanthanum & Cerium extracted first) - they do not break them down individually.

Profitability, Production Volume and Cashflows:

Lynas' COGS doesn't vary much year-to-year, so revenue (based on REO prices) is the biggest factor determining their profitability:

Operating expenses (G&A and forex) vary randomly, while financial expenses have dropped as debt was lowered:

But these are both too small to affect profit much.  The end result is wildly fluctuating profits, following REO prices (the first graph above), with monster profits in FY2022:

Production output has varied:

They aim to increase production to 12,000t NdPr by end 2024 with their Mt Weld Expansion project.

CFO has always been greater than profit, except in FY2022.  The company does not reconcile cashflows with profits, but in most years, depreciation (eg: on p72) is enough (or more than enough) to account for the difference.  CFI is lower than CFO for the past 7 years:

In summary: everything depends on REO prices.

Balance sheet

As of June 2022, balance sheet has AUD 900m net cash.  There is 1.2bn total capex expected in FY2023 and FY2024 combined, for the Mount Weld Expansion and the Karlgoorlie Lanthanum separation facility. 

The company has a history of issuing new shares:

Given this, lets look at their EPS, rather than their PAT (2 charts above):

Seems OK.  I don't like companies issuing equity, but for a highly cyclical mining company, its safer than taking on debt.


They increased their reserves by 60% in 2018, and have since been mining them down:

As of June 2022, they have 72 years production left, at FY2022 levels.  If they increase NdPr to 12,000 tonnes/year, assuming a proportional increase in REOs, that gives 33 years of production.  The REO's given in the reserves here are for all REEs, not just the valuable ones.

Trying to guess the percentage of NdPr based on their production figures:


Hard to value a commodity company, as you need a price for the underlying commodity.

Back of the envelope calculation:

  • I expect demand to rise for REEs due to EVs and windmills, but the FY2022 price rise was too large.  I'll take the FY2021 price to be a 'normal' price.    
  • That gave us an EPS of AUD 18c, from production of 5461 tonnes of NdPr.
  • Assume they reach 10,000 tonnes NdPr in 2024 (vs their 12,000 tonnes target).
  • Gives an EPS of 33c.
  • At AUD 8 per share, its a PE of 24.  

Still too expensive for me.


Good company, but highly cyclical.  Dependent on REE prices, that are controlled by China.  May be some remaining political risk in Malaysia.  If China attacks Taiwan, this will be one of the few stocks that goes up.

My Positions: covered Tech shorts

Covered my tech shorts at a loss.  It was a mistake to hold onto SMH and XLK for so long with the AI bubble, they turned bullish weeks ago, should have cut then.

Still expect a downturn, but the bubble can go higher for a few weeks/months first.  Plan to buy some more puts next month, then add more shorts heading into August.

Plan to slowly add more short positions, before the next downturn.  And wait for the downturn to add some longer term investments.  Right now I'm doing nothing.

Monday, June 12, 2023

Bought Puts

I still think we're headed for a recession, now the debt ceiling is resolved the Fed's Net Liquidity should fall as they refill the TGA.  With the VIX at 3 year lows, and SPX concentrated on a record few number of stocks, I bought 4 SPY puts (strike 410 15thDec23) at 10.08.  Its 0.5% of my portfolio, with 4X upside if SPY drops 17% to its 2022 lows.  Aim to buy a few more puts next month, and the month after.  Aim is to buy at VIX 13 and sell in the 30s.

Why 6 months out?  The Fed's Net Liquidity probably takes longer to reduce than expected.  Probably pushed out till July/August.  The numbers below show the Fed's Net Liquidity every Wednesday: in the week after the debt ceiling resolution, the TGA increased but RRP decreased by almost triple that amount, so Fed Net Liquidity actually increased.

70% of options expire worthless.  And I was expecting a bear market 6 months ago.  So maybe I'm just burning money.  But the risk/reward is worthwhile.  Everything takes longer than we expect.  

Saturday, June 10, 2023

Rare Earths: An Introduction

(Updated: 29th June 2023)

Quick notes on Rare Earth Elements (REEs) which are are used to make magnets used in hybrids, EVs and windmills.

It takes years to enter the rare earths business, because:

  • As you're extracting 7 or 8 elements, processing is far more technical and complex, than an iron ore or copper mine for example.
  • There's no benchmark pricing, so its hard to new entrants to get long term contracts needed for financing.
  • Production involves several steps.  And for each step, the product has to go through a rigid qualification process.  So you usually complete one step at a time, each of which takes one or two years.

Processing Stages

The steps involves are:
  • Mine the ores (which have < 10% concentration of Total Rare Earth Oxides (TREO)) and upgrade it to concentrates of at least 60% TREO for further processing.  Starts with a mechanical process, followed by a chemical process.

The rare earths we want for magnets are Nd, Pr, Tb and Dy.  The elements on the left are low priced money losers, and the heavier ones on the right are precious.  So the concentrations of different light and heavy elements matter.  For example, typically you may have 25% La, 25% Ce and 1% Dy.  Current prices (USD/kg, early June 2023) are in pink - see the wide variation.

Scandium (Sc) is not really an REE (its like aluminium), so is not subject to the above extraction order.  Its grouped with REEs because its often found with them.

This is the hardest stage of production.  It is challenging for a newcomer to prove they can produce the correct amounts of oxides consistently.

  • Produce Alloys from the oxides.  Low margin business which requires strong environmental safeguards.
  • Produce Magnets.  Needs to be produced under license, using patents for one of 2 processes.  Both processes were developed in the 80's: one by GM is now held by a Chinese company, and  another by Sumitomo which is now held by Hitachi. So you need to license from Hitachi.

Industry Structure

There are 5-10 producers within China, and 2 outside it (Lynas and MP):

Other Chinese companies (not shown above) are China Rare Earth Group (a merger of 6 SOEs), Shenghe Resources (who buys from MP above), GuangSheng Nonferrous, Yuijing & Huicheng Environmental Protection.  A lot of China's rare earth production is a side-product (eg: from iron ore), so it does not have to be economical.

China dominates rare earths metals at every stage.  Its been hard for other junior producers to start up, as they have to build a mine and separation facilities (around $25-50m), while arranging off-take agreements to get financing, under the shadow of China holding prices below production cost.  This prevented new players emerging, at least till the 2021 bull market.  It takes around 3-5 years to get mining and separation facilities up and running. 


I haven't been able to find free long term charts of all the rare earth elements.  But they seem to follow each other:

La, Cr, Nd and Pr: 2003 to 2015:

Source: Resaerchgate

La, Nd and Pr: 2008 to 2018:

The only recent chart:

Source: tradingeconomics 

Prices peaked in 2011 after China blocked exports to Japan in 2010, for a week.  They peaked again in the "everything bubble" of 2021/22, before falling 55%.  A bet on REEs now is a bet that increasing EV penetration and "green energy" drives demand up, while supply takes time to come online (takes years to start up mines and separation facilities) despite US government subsidies.

My Positions

 Few changes:

  • Covered XRT as it rose 8% in a week on higher volume.
  • Replacing it with XLI

Like a broken record, I expect a recession and a final leg down for the market.