Friday, June 16, 2017


Potash has been dropping for four years now.  Is this a normal commodity cycle, where prices stay low enough that new investment is discouraged and marginal players are squeezed out?  If so, where are we now?


Potash is a fertiliser which provides crops with potassium - one of the three primary nutrients for plants.  It accounts for 70 to 90% of potassium fertiliser used, and is commonly used for bulk crops or grains (1).  There is no commercially feasible substitute, although some are working on it (12).

Potash is plentiful, but the industry has long since been an oligopoly, due to the enormous capital requirements in setting up the mines and transport.  In 2013 the potash oligopoly broke up, with one of the players flooding the market.  Prices have headed down since:

(Source: SeekingAlpha article: Potash - a 2018 story?)


Annual demand is around 55-65 million tons.

In the long term demand is constant.  A known amount of Potassium is depleted in the soil for each crop and must be replenished.  But usage varies year by year, as replenishment can be delayed a few years due to crop rotations, rainfall, crop prices, etc:

(Source: Uralkali 2015 investor presentation)

Long term demand is has grown by 2% since 1994:

(Source: 2016 Agrium Inevstor day presentation)

Ignore short term demand since its unpredictable.  Just assume long term demand growth of 2%, and concentrate on looking at supply instead.


The latest cost curve I have, showing actual production estimates, is from 2015:

(Source: Morgan Stanley: March 2016 Report: Global Chemicals: Potash S&D Update)

Transport Costs

The above excludes transport costs, which are significant for a bulk commodity like Potash.

From Canada:
From Russia:
  • 80% of Uralkali's exports are shipped through St. Petersburg.  Total transport cost (rail and freight) for this was USD 28 per tonne in 2015 (p29).  Most of the remaining potash was transported to China by rail, which cost USD 39 per ton.
From Belarus:
  • Belaruskali gives no transport costs, but they must be lower since Belarus is so much closer to the sea that Uralkali's Perm region mines.

Future Supply

Supply is quite easy to predict.  Since potash mines are massive projects that take 5-7 years and $1-2 bn to complete, most companies' expansion plans are well known.   Most projects go belly up, especially those from small companies.

For production that is likely to be built, net depletions:
  • I get 11 mt/pa to be added by 2021.  8.5mt if you exclude Yancoal, for which the project has not started development yet.
  • That is assuming mines operate at 90% of capacity. 
  • Also excluding any additional capacity from Potash Corp (POT).  Assume they won't increase production even if they have capacity, as they continue to act as the swing producer in a low priced environment.
  • Also assumes Jansen dies.


Even taking the lowest figure of 8.5 mt net extra capacity by 2020, with the conservative assumptions above - with long term growth of 2%, it will take 5.5 years (from now) to absorb the extra supply.  That seems balanced.  We are not at a place where we can say that Potash prices will rise due to constrained supply in the future.  We're at a place where we can say they may stop going down.

Another thing to watch for are potential African projects.  Ethiopia (Dankalli) seems dead, with ICL pulling out due to lack of infrastructure.  But Republic of Congo and Gabon look interesting - very low production costs, no rail required, and easy shipping to Brazil.  Still too early to tell if they produce, but if they do hit their numbers, they'll be the cheapest producers in the world.

Appendix - List of future Potash projects

I judge the likelihood of new capacity plans to be actually built based on the track record of the company, their financial backing, and the cash costs of the project.

New production capacity that is likely to be built is below, listed from lowest-cost to highest-cost (left side of cost curve first):

Net production addition
Cash Costs/mt
Increase from 10.8mt KCl (production) in 2016 to 14.4mt (nameplate capacity) in 2020  That includes a 2mt loss from existing mine depletion, and 2 new Mines (Solikamsk No 2 for 2.3mt/pa and Ust-Yayvinsky for 2.5mt/pa).  (Source: E&MJ, 2015 article)
2.2mt by 2020, assuming operating at 90%capacity.
Same as existing Uralkali operations
2mt.  Production scheduled to start from 2021 to 2023.

Mine under design in 2015. 

This company seems very profitable, but this is their first potash mine.
2 mt capacity.

1.8 mt if operating at 90% capacity.
Same as Uralkali - in Prem region.
Usolskiy mine expected startup in 4Q 2017.
Volgakaliy expected mid 2018 startup.
Both projects have 8.3mt KCl production potential.
4.2 mt nameplate capacity around 2020. 

So 3.8mt if operating at 90% capacity.

8.3mt nameplate capacity afterward?
Same as Uralkali

That includes Petrikovskoye, which will add 1.5mt KCl.

Little information available about this company.
1 mt if operating at 90% capacity.
Same as existing Belaruskali operations
Produced 2.2m in 2016, which is 73% of nameplate capacity of 3mt (p8).  Expect to ramp up in 2017 .
0.5 mt, if operating at 90% capacity.
Same as existing Agrium   operations
Legacy: started producing in May 2017, expect to reach 2m production by end 2017.  Plan to sell 0.7mt to US, the remainder offshore.
Started production May 2017

K+S estimated to currently produce 6mt in Germany.  High cost operations.
2mt by end 2017

2.9 mt by end 2019
Slightly lower than Potash Corp
Potash Corp (POT)
9.3mt production in 2016.  They closed high cost mine (New Brunswick) but opened a cheaper one (Rocanville).
Total planned expansion is 10.1mt.    Morgan Stanley estimates 2015 had 9.3mt operating capacity, with 5.5mt capable of restarting in 12 months.  They model 13.4mt capacity increase by 2020.

Not all their capacity is used; POT acts as the swing producer.
Additional 10mt capacity, but may not be used if process remain low.

Part of Yanzhou Coal (HK:1171), which has been profitable for the last 5 years (p10) , but with high debt (40X 2016 income, p11). 

Are Chinese companies operating from a strategic rather than economic perspective?  ie: acquire resources, don't expect profits.

Economics should be the same as any Ssachkatchewan project. 

2.5 mt if operating at 90% capacity.
Similar to POT

Production the is likely to be removed is listed below:

Net production removal
Cash Costs/mt
Cease UK production of 1mt/pa in 2018, produce SOP instead.

High-cost US miner.  May go out of business.  I’ll be conservative, and assume they do.

Sigmundshall mine depleted by 2020.  Estimate its production is 0.3mt (p25p8)

0.5mt.  Depleted by 2018
-0.5 mt

Speculative projects, as of now, unlikely to be built:

Nameplate Capacity
Cash Costs/mt
Additional 2.8mt from panned Polovodovo mine.

Costs 1.6bn, they may not have the money.
2.8 mt by 2023
Same as existing Uralkali operations
Jansen.  8mt pa production, (10m nameplate capacity). 

I don’t think it will ever produce.
10 mt
Need potash price of over USD 400 to be profitable, says POT.
Sirius Minerlas
York project, under a national Park.  Nameplate capacity 20mt/pa.  Currently under design and site preparation.
Produces polyhite, not KCl.

May be a different product.  Polyhite is probably more useful for producing SOP  It has 4 plant nutrients, so may be better sold as as blended fertilizer that has no Cl.  Has less potassium than SOP.

LSE listed.  Stock price up to ~1.4bn pounds market cap.  Gina Rinehart has a stake.
Probably not relevant, supplying a different market.
Encato Potash
2.8mt capacity in Saskatchewan. They bought native rights. 

Encato has a market cap of ~50m.
2.8 mt
Similar to POT
Circum Minerals
New Dankalli mine (Ethopia) schedued to produce 2mnt by 2021

ICL quit the project in Oct 2016.
2 mt
USD 38/t operating costs.  But 500km rail/truck to Dijoubit. 
Dankalli/Colluli (Eritrea)
Produce SOP not MOP, not relevant here.

Elemental Minerals (Kore Potash)
Kola project. DFS to end in 2018.  Expected start of production 2022.   DFS to focus on 2mt.
Yangala (Dougou extesion)

ASX listed, small company
Very Cheap: Life-of-Mine cost $68/ton. 
Cheap transport: Only 36km from coast – use conveyor belt
In Gabon (near KorePotash, north of Republic or Cngo)
2 projects, no estimate for production capacity.

ASX listed, small company

African Potash
Lac Dinga: In Republic of Congo.  No estimate provided for production capacity.

Small AIM listed company.

Highfield Resources
1.6mt production for 2 Spanish projects. Plus 1m for another project.

ASX listed, small company
3.6 mt

Sold Pacific Basin Shipping

Sold Pacific Basin Shipping (HK:2343) at HKD 1.60 on 8th June, for loss of SGD 1703.

I've changed my mind.  Not confident on this one, as new ships can be built in 28-24 months, limiting any sustained rise in prices.  We can't just look at a 10-year chart and say the BDI is gonna go back to 2008 levels. I think the time to buy shipping stocks is when even the best players have been losing money for 1 or 2 quarters - when things are so bad they can only get better.  At that point, you may get 50% upside.  It would be a small trade - trying to catch a falling knife - and I would be dribbling in slowly... maybe 1% after one bad quarter and another 1% after a second.

If I want to play at all.  Shipping is a tough sector.  There's no way to value a shipping company - earnings and vessel values are cyclical.

Tuesday, May 9, 2017

Resona Bank (TYO:8308)

Japanese banks are trading at single digit PE ratios and below book value.  After 20 years of near-zero interest rates they are still profitable.

I'm looking at Resona, the fifth largest bank in Japan, because its purely a domestic play and makes most of its money from deposits and loans.

Market Share

Japan is 'over banked'.  The top four banks have around half the market share, and theres a 'long tail' of smaller banks due to each province historically having their own bank.  
     (Source: Credit Suisse, Midtier banks, Jan 2015)

The Japanese banking market is more competitive than SingaporeAustralia or the UK, where the top 3-5 banks have 60-75% of market share.  Consolidation has been occurring for years, but still has a long way to go.

To escape the competition and ever lowering interest rates, the larger banks (MUFJ, Mizuho Financial Group, Sumitomo Mitsui Financial Group and Mitsumi Sumitomo Trust Holdings) have expanded overseas and have built up their investment banking.  For example, MUFJ owns 20% of Morgan Stanley, plus banks in the US and Thailand.  Resona is the largest Japanese bank that is still purely domestic.

Resona's market share in certain prefectures:

     (Source: Feb 2017 Company Investor Presentation)

Income Breakdown

The vast majority of their operating income is from interest:

Interest income seems to track interest rates, with a spike in 08/09, amidst a general downtrend (the time period in the red box below):

This 'single focus' on deposits and loans makes it easier to analyse than the other large Japanese banks:
  • They have no investment banking or overseas operations.
  • And very little trading.  Most of their trading profits are from derivatives.
  • Fee and commission income is the second largest contributor, and seems to track the economy1:


Loan losses2 are slightly favourable compared to other Japanese large banks:


Resona's CET1 CAR is 8%, far lower than the other 4 big banks which are in the low teens.  This is because of their low capital base.  In 2015 they finished paying off public funds which were injected in 2003.  They now expect to increase their capital and aim for a 9% CET1 CAR in 2019.

As a domestic Japanese bank, their required CET1 CAR is only 4.5%. 


Resona's ROE is artificially higher than its peers, due to its low capital base:

As they increase their capital in future years, ROE will go down.

ROA is comparable to the other large Japanese banks:

This is far less than other countries.  Singapore banks for example, have ROAs of between 0.5 and 3%.


David Eirnhorn bought a stake in 2014 at 547 yen, and probably still has it.

At 620 Yen, Resona yields 2.7%.  Withholding tax for Singapore residents is 15.315%.


Reasons to buy:

  • Pretty Cheap.  At 620 yen, its trading at 9X earnings and price/book of 0.87  (From year-ending March 2016).
  • It is still profitable, despite operating in a fiercely competitive environment with interest rate headwinds.


  • Japanese banking market is more competitive than other countries.
  • Japan's well know demographic problems.
  • Risk of a sudden Yen devaluation if Japan loses control of its bond and currency markets.  May not happen as Japan's debt is domestically funded (1) (2).
  • Cyclical risk.  Japan is now in a boom, with record corporate profits and full employment.  The boom may end if it was solely due to money printing.

Overall, I think its an OK company trading at a cheap price.  Not a 'Buffet buy-and-hold-forever' stock.  The risks are priced in.

[Edit: Bought 2900 shares at 612.4 Yen on Friday 12th May.  Total cost in SGD is 29,505]

I like their mascot too:


1 There is a breakdown of their fees on p16 here, but it doesn't say much

2 From JGAAP financial results. "Loan losses" calculated as: 
     (Provisions for loan losses - Reversals in provision for loan losses) + (Write off loans - Recovery of write-off loans)

Saturday, April 1, 2017

Pacific Basin Shipping (HK:2343)

Pacific Basin shipping is a bulk carrier, operating Handysize and Supramax ships.  These ships carry minor bulk cargo (e.g.: logs, steel products, cement, grains), are smaller, and often have their own cranes, allowing them to dock at small ports that do not have loading or unloading facilities.  So they operate in a different market from the major bulk carriers (iron ore, coal, and sometimes grains), which form the majority of bulk shipped materials.  Despite being in different markets, the Baltic Dry Index (BDI) and Baltic Handysize Index (BHSI) follow each other:


Pacific Basin has the largest market share in the handysize segment, but is still a price taker in a fragmented market:

Pacific Basin has 111 Handymax ships (75 owned) and 34 Supramax ships (20 owned).  To get an idea of the market, some competitors are:
  • Western Bulk: 3rd largest owner of Supramaxes in the world.  Private company, releases short annual reports.  126 vessels in 3Q16.
  • Navibulgar: Privately owned Bulgarian shipper, expanding rapidly (1), (2).  20 Handymaxes and 9 Supramaxes.
  • Maybulk (KLSE: MAYBULK): 7 Supramax and 7 Handymax.
  • The rest seem to be small, private companies.  eg: Falcon Marine (17 Handymax),  Apex Marine (5 Handysize), MT Marine (13 Handysize).
Dry bulk shipping is highly cyclical.  In February last year, the BDI & BHSI dropped to there lowest levels in 15 years.  Every player in the industry was losing money.  

The key for future profitability is in these numbers.  All are from Pacific Basin's annual report:

  • New supply of ships.  New supply is still coming online as orders places several years ago are fulfilled.  The bulk of these orders are expected to be delivered this year (9.2% fleet growth), then dropping off sharply next year:

  • No new dry bulk ship orders have been placed this year (except for 31 Valemax iron ore carriers, which serve a different market).  A five year old second-hand Handysize is benchmarked at USD 13.5m, far below the estimated USD 19m for a new one.  This excess of second hand vessels on the market means new ones are unlikely to be ordered soon.
  • Scrapping: Last year, 3.6% of global dry bulk capacity and 3.1% of Handysize capacity was scrapped. New ballast water treatment regulations from September this year (requiring the costly retrofit of ballast water treatment systems) may encourage ship scrapping.
  • Demand: Demand is unpredictable, and depends on global growth and industry specific factors.  No point trying.
A bet in this company is a bet on the industry.  I'm hoping that the industry will continue to improve, and stock prices have not yet taken it into account.  Bought 62000 shares on 29th Mar at HKD 1.69, for a cost of SGD 19,076.

Of course, it would have been better to buy in Feb last year.

Wednesday, March 22, 2017

Boustead Projects

BP has 2 businesses:
  • Design-and-Build: Undertake construction projects to design and build industrial properties for clients.  This is project based, so is lumpy.
  • Leasing: Lease industrial properties to clients.  This is recurring.
Breakdown of profit segments:

Theses two ways of looking at this company: as an asset play (buying $1 for 50c), or as a cyclical play (buying near the bottom of the industrial property market).

As an Asset Play


BP holds its assets on the balance sheet at less than market value.  Considerably less, in fact.   Going through the items:
  • Properties held for sale (4 in SG, 3 in China) are held on the 3Q17 balance sheet at 30.5m.  However they have a 2016 valuation of $103m ($107m for 2015).  This is based on "income and comparable sales".
  • Investment properties (all in SG) are held at 141m, but their 2016 valuation is 259m (265m in 2015).  
  • Since the 3Q17 results have been released, a stake in Triple-One Somerset has been sold, with an 8m profit (before tax).   This is recorded on BS as Available for Sale Financial Asset at 17.8m.  Assuming a PAT of 6.5m, this gives 24.3m.
  • There is another financial asset for sale held at 22m, but this is a share of a China project.  Impossible to value, so ignore it.
  • Also ignore investments in Joint Ventures (15m).
  • Looking at current assets & liabilities (working capital, receivables/payables, WIP, deferred tax) gives -10m.  Small enough to ignore.
  • No borrowings.  Net cash 9m.  Small enough to ignore too.
With 320m shares outstanding, that gives $1.13 per share.

As a Cyclical Play


  • Subtract the Triple-One Sommerset stake, sold for 7.5c/share from the market cap.
  • Assume they sell their 'Properties-held-for-sale' at 50% of the market value.  Thats 50m, or 15.6c per share
  • Assume their leasing revenue then drops proportional to their market values, as those properties sold were previously being leased out.  Now the leasing business is earning 11m a year (before tax), or 3.4c/share.
  • At an 80c share price, minus the first two items above, this makes it 17X before-tax-earnings, just for the leasing business.  These are recurring earnings.  And we get the highly cyclical design-and-Build business for free.

The Industrial Property Cycle

The industrial property market is cyclical, and past downturns have been long and painful:

(Source: Singstat)

One reason for the downturn is excessive building bought about by endless QE.  This will peak this year:

The other reason for property market downturns is lack of demand.  The chart below overlays recessionary periods onto the industrial property index:

So recessions are not the sole cause of property cycle downturns.

DBS expects "further weakness till the end of 2017 before bottoming out from 2018 onwards."  Unless there's a recession, in which case all bets are off.

My expectations are:
  • No recession soon, probably a mild global recovery in the next 6 months.
  • But property supply has exploded due to 7 long years of low interest rates.  This is now reversing, so we can expect the downturn to be long and deep.
Though its tempting to buy due to the discount to asset values, I'll wait till the end of the year at least.   This is my gut feeling, or judgement call.  Or I'll wait until the price drops enough that its irresistible.  And I need to remind myself that markets look forward, and we want to buy before things get better - we want to buy when they are getting less worse.

Saturday, March 18, 2017

US Growth Expected

Last month, ECRI stated that they see a genuine cyclical upturn in growth which will last for at least six months, maybe longer.    But growth will be lower than past cycles, at 1-2%.  3% is unlikely.

This doesn't mean the US stock market will go up - high expectations may already be priced in.  But it means lower unemployment, rising inflation, and rising interest rates.  Probably globally.  Including gently increasing growth in Singapore.  And higher USD1 with lower gold.  I don't know about commodity prices, since demand (from China) and supply are bigger factors.

I'll still be holding a lot of cash though, as stock markets are expensive and its hard to find things to buy.

ERCI's predictions have been pretty good, though they made a bad recession call in 2011.

1 Higher USD vs other global currencies, not necessarily against SGD.  But a lot depends on the Euro.   No way to know if the EU will breakup till after the French and Italian elections.  

Saturday, January 14, 2017

Coach (NYSE:COH)

Coach's sales and profits have dropped for the past 3 years, but the company is now supposed to be in a turnaround.

Before looking at the numbers, the most important thing for luxury goods is branding: a combination of consistent pricing, retail environment and advertising.  Styles come and go.  But consistent branding means a company can survive product flops and style disasters:

Here in Asia, Coach is mid-range luxury, compared to brands such as LV.  All bags are sold through their own stores, or their 'store-within-a-store' in upscale department stores.  Prices start around SGD 500 for a leather handbag.  But occasionally in some places, you could still see piles of the old cheap 'CC' bags on wooden racks, where some enterprising person has imported a few and set up a temporary shop.

And this is the problem - Coach is actually two brands:

  • An "Accessible luxury" brand, around USD 400 to 1000 for a nice bag.  Still expensive, but affordable:

  • A cheap knockoff brand available at "Outlets" and "Factory" stores.  Which has cheaper (MFF - made for factory) versions of the same products:

The factory outlets sell different versions of the same models under the same brand name.  You can tell the two apart by looking at the bag's serial numbers.

Coach has destroyed their brand by opening outlets and selling cheap versions of their products.   Some comments from the purseblog forum about the brand's history:

It used to be a luxury brand about twenty/twenty-five years ago. Coach's quality took a huge nose-dive in the 00's, for sure. The advent of the Coach outlet brought it down. I remember when Coach bags used to be displayed behind the counter and in the glass cases at department stores (late 80s and 90s for me). The bags were all thicker, sturdier leather back then. (link)

It'll be very interesting to see if Coach will be able to backtrack from the endless turnover/scarcity marketing/quick markdown schedules that has characterized their brand for the past 10 or so years. LV still makes Speedy, Chanel still makes GSTs. Does Coach (outside their classics line) make anything from even 2 years ago? That's what keeps them from being considered high end, IMO - they have a lower-end selling scheme. (link)

And about the relentless markdown or copying of boutique products:

Phoebe's ... not valid on PCE ... then they were valid ... since then SAS and outlets ... now they actually *are* a MFF bag. (I literally :lol: out loud when I type that.)

Legacy from five years ago ... some were not valid on PCE ... then were valid on PCE ... then "they'll never go to outlet" ... then they did ... then MFF came up with some very similar versions.

Boroughs ... not valid on PCE ... then they were valid ... then on SAS and outlets (and I got a few through both routes and love them dearly) ... then MFF actually made their *own* Boroughs.  (Feb 2016 - link)

This is another reason I have yet to pull the trigger on the oxblood rogue (plus no boutiques near me carry the rogue to see if it's really worth the $). My boutique swears it will never hit the outlet, go to SAS, or allow PCE but we have all heard that story before. 

...I rarely buy anything full price because I know there is always a department sale or pce coming up. I would be comfortable paying $795 for a rogue if I knew it wasn't going to be deeply discounted at Macy's or hit the next SAS at 40% off. (Feb 2016 - link)

Coach presented a turnaround plan in 2014, which involved:

  • Closing 20% of their boutique stores, and 'consolidating' some of they factory outlets
  • Upgrading their factory outlet stores
  • Selling new designs from Stuart Vevers
  • Limit discounting (PCEs - Preferred Customer Events).
Have they done this?  Yes, and sales/profits have recently improved.

But fundamentally, Coach still sells 2 different priced brands under the same name.  The company does not give a breakdown of revenue/profits from outlets, which is estimated at 70% (1) (2) (3).  But they give the numbers and areas for different store types.  The trend is clear:

Looks like Coach will never get rid of its outlet stores.  And until they fix this, I won't consider the shares no matter how good the finances are.  If you're brand is no good, then you have to compete on fashion/design/fads/whatever, which is unpredictable.

Selling the same brand at two different price points is misleading.  Making cheap knockoff copies of your own products and selling them a discount is lying.  My wife has seen bags that she likes, but won't buy them because they are Coach.

What would make me change my mind?  If they were to put their outlet stores under a separate brand, maybe.

Nice bag, but not buying it.

Saturday, January 7, 2017

Bought UUP and URA

On the 4th & 5th Jan, bought 1507 units of the Global X Uranium ETF (URA), for total cost of USD 20611.24 (average price USD 13.67).  Thats the maximum I'm willing to risk, given operating leverage involved, and the fact that that the uranium price could be low for years.

Bought UUP (US Dollar Index ETF) on December 5th 2016.  800 units at USD 25.97, for a total cost of 20,781.59.  This fund replicates being long USD against other developed countries' currencies, mostly the Euro (Click on 'Portfolio' tab here).  Synthetic fund with a 0.75% management fee.

This is a small bet on the USD rising.  Whether its because of Trump, the falling Yuan, European elections, or whatever.  The main risk is that this is a consensus trade - google around, and you can't find anyone bearish on the USD or bullish on the Euro.

My portfolio is still 71% cash:

Just taking small bets.  Stocks in general are still expensive, especially the US.  I'm only really comfortable buying in a recession or crisis.

Tuesday, January 3, 2017



The price of uranium has been going down forever.  For 5 years, since the 2011 Fukushima disaster.  Or for 10 years on a longer term chart:

Source: Cameco

Although uranium use has been falling since 2011, 60 new reactors are now under construction, mostly in Asia:

Source: IAEA.  See World Nuclear Association for an updated & detailed table.

This is a 13% increase in the current 450 operational reactors.  The bull case for uranium is that an overreaction to Fukushima and the multi year slump in prices has undermined sentiment in the industry, halting exploration and curtailing mining.  And prices should see a massive jump when new demand comes online.

Economics of Nuclear Plants

Nuclear fission makes up 11% of the world's electricity.

Nuclear plants have high fixed costs, but have low operational costs and run for a minimum of 30 years.  It is hard to vary their energy output, so they are best suited to base-load power plants.

To startup (or restart) a reactor, you need twice as much uranium in the first year.  Most reactors will stockpile 7 years of fuel before starting.

Nuclear plants need water for cooling, so can only be operated in costal areas.

In the US, cheap natural gas may make nuclear plants uneconomical.  Exelon came close to closing 2 Illinois plants which were losing cash on an operating basis.  Even though the 2 plants were saved, I think its unlikely many new plants will be build in North America, due to the high upfront cost.  This won't affect nuclear plants in Asia - cheap natural gas cannot be exported from the US to Asia - once you do its no longer cheap1.

Solar and wind power are getting cheaper and may already be a parity.  But they don't provide electricity throughout the day unless we get improvements in battery storage.  So the alternatives for base-load production are nuclear (expensive, risky), coal (cheap, dirty) or natural gas (clean, expensive in Asia) or oil (expensive).  So I'd expect that nuclear plants will continue to be used for base-load power generation where cheap piped natural gas is unavailable, water is plentiful, and air pollution is a concern.

Uranium Demand, Supply and Stockpiles

Demand is straightforward as the only commercial use of uranium is for fuel.  The number of reactors operating, under construction and planned is known.  Forecast uranium demand is from up 10% over five years to 26% over 10 years.

Mines supplied 60,469 tonnes of Uranium Oxide concentrate in 2015.  The amount required was estimated at 63,404 tonnes in 20162.    The difference was made up by drawing down stockpiles.

No one knows how much is stockpiled.  Early uranium production first went into military stockpiles, then later on in to civil stockpiles.  Since the 80's, these stockpiles have made up the difference between demand and mine output:

Source: World Nuclear Association

Even the size of civilian stockpiles is uncertain.  It is suggested that China has stockpiled more than one worldwide year's supply of uranium.  Japan has been selling off its stockpile since 2011, and nobody knows how much they have.  Global inventory estimates are all over the place.  Nobody knows.

Cost Curve

The latest cost curve I can find is here, but its not labelled.  The article says that most mines were cashflow positive in 2015, due the falling currencies of commodity producing countries.  Long-term contract prices fell by around 30% in 2016, so some may be losing cash now.

The lowest cost producers are ISL mines in Kazakhstan, and Cameco's mines in Canada.

Cameco (NYSE:CCJ)

The textbook strategy while awaiting a commodity price turnaround is to buy the lowest cost producer.  Cameco is the lowest cost (listed) producer - its two biggest mines, McArthur River and Cigar Lake have ore grades of 16-17%.  Most other mines have grades of less than 1%.

Some quick back-of-the-envelope numbers for Cameco:
  • Profits in 2014, 2015 and 9-months 2016 were CAD 183m, 63m and 85m respectively.
  • You could add another 40m to 9-month 2016 profit, due to one-off costs in winding down Rabbit Lake 3
  • Debt is around 1.5bn.  Long term notes, mostly due between 2019 and 2025.

The trouble with Cameco is their massive tax dispute with the Canada Revenue Agency (CRA).  They are alleged to have engaged in transfer-pricing from 2003 to 2015, by selling to Swiss subsidiary at below market prices.  They may receive tax expenses of up to 1.7bn (maybe more 4), plus interest and penalties.  The case for years 2003, 2005, and 2006 is under trial now with a result is expected in 2H17 - the company says the amounts claimed for these 3 years are 'modest' and can be covered by cash.  But the results may be later applied by the court to the other years.  Cameco says they have not broken the law, and have only recorded a provision of $54 million (as of 3Q16).  The case is too complex for a layman to understand (1) (2).

The possibility of such a large payment adds an unknown binary element to investing in Cameco.  There's a small possibility the company is screwed.  In the worst case for example, having to issue 1.7bn in bonds at a 6% interest gives an expense of 100m, raising doubts about their ability to survive when uranium prices are so low.  Or issuing more shares, which would dilute shareholders, and come close to nationalising the company.

Global X Uranium ETF (URA)

Due to Cameco's potential tax problems, its may be better to buy the URA ETF instead.  It holds:
  • 22%: Cameco
  • 39%: Other Uranium E&P companies, that are currently producing.
  • 24%: Uranium exploration companies, not currently producing - more speculative.
  • 8%: Nuclear companies (involved in mining, processing and building/running reactors).
  • 7%: Uranium ETF (holding actual uranium)
Around 50% of their holdings operate primarily in North America, 10% in China/Kazakistan/Mongolia, and 8% in Europe.


Risks to the bull case are:
  • China's nuclear plans do not work out, perhaps due to economic problems.
  • Advances in battery technology make solar feasible for base-load generation.
  • We may simply still be in the downward part of the cycle - people have been saying that uranium will recover for years.  There still may be years more to go, especially since the size of stockpiles is unknown.
  • Cheap supply from Kazakhstan.

1 Majority of LNG price is from liquefaction - see the third slide here.
2 This was not actual demand, as it excludes some outages, but it was potential demand.
3 See question by Greg Barnes in 3Q16 Transcript.  Care and maintenance for shutting down the mine is immediately expensed from COGS, not capitalised over time.
4 Its unclear, see page 11