Stocks for discussion:
Mining (Kumba)
“All over the place”
Share price: 12700 cents
Net shares in issue: 320,8 million
Market cap: R40,8 billion
Fair value up to R87 at $50/ton iron ore
Trading Sell (from Buy) and Portfolio Buy
Iron ore nudged beyond $70/ton this past week compared with a low of $40/ton earlier this year. There has been lower production than envisaged, included Roy Hill in Australia, higher Chinese steel production for construction, and an increase in the margin requirement to 8% in China on speculative retail trades on iron ore.
One order of magnitude difference in profitability judging by the gyrations of the Kumba share price.
We are going from say R1 per share in EPS to R10 per share in EPS in a matter of weeks if the share price is any guide to earnings forecasting. The share price, closing at a low of 2540 cents on 18 January 2016, closed at 12700 cents on Friday 22 April. This a stock that was briefly being bought and sold for over 60000 cents per share in January and February 2013.
I’ve mentioned before that programme or algorithmic trading has partly been a contributor to wildly exaggerated variations in commodity shares. Spot valuation assumes the variables present at that time prevail for an extended period.
Company management though cannot run a mining business like the markets price the stock. Cash flow realities in mining mean you have to bite the bullet in a cyclical trough until demand and supply rebalances to improve pricing.
Kumba had an average cash breakeven price of $49/ton for the year ended December 2015 with $41/ton achieved by year end, which compares with $63/ton in 2014. At prevailing freight rates and currency management targets between $32/ton to $40/ton.
Cash breakeven has a sensitivity of $3/ton for every R1/$ move. The stronger the rand the higher the threshold for breakeven. A $1/ton move in iron ore has a R500 million impact on the P&L, a ten cents move in the rand versus the dollar impacts profits by R250 million and a 1000 ton change in volume affects profits by R350 million. Sensitivities are only applicable for a full year, all else equal.
These sensitivities partly explain manic depressive market behaviour.
62% Fe iron ore average $56/ton in 2015, $61/ton in the first six months and $50/ton in the second half. This compares with $97/ton in 2014, $135/ton in 2013 and $130/ton in 2012. World iron ore exports grew by 3% in 2015 with 7% growth in the second half, with Anglo’s Minas Rio kicking in with 8,5 million tons in sales.
For the year to 22 April 2016, iron ore has averaged $50,83/ton – in line with my $50/ton forecast average. For Q1 2016, the average is $48,65 compared with $62,48/ton in Q1 2015, a decline of 22%.
The average for January was $41,89/ton, the average for February was $47,10/ton, the average for March was $56,14/ton and the average for the first three weeks of April 2016 was $59,15/ton.
The share price performance earlier this year would suggest bankruptcy and yet even on January prices Kumba would have been cash flow positive (just) over a full year. Moreover, the balance sheet is today much stronger than a year back and getting progressively so.
Up until a year ago, there was a close correlation between the trend in the iron ore price and the Kumba share price. Between April 2014 and April 2015 the goodness of fit in relation to what was predictable is high between the two data sets, giving an R² of 0.98 and a Pearson product momentum correlation of almost 0.90.
Between July 2015 and 22 April 2016 the correlation is far less pronounced. And, from the beginning of December 2015 until the low point of the share price on 18 January 2016, the Pearson correlation is only 0.028 with the share falling 45% but iron ore broadly stable.
KIO share price and Fe 62% iron ore price from 22 April 2015 to 22 April 2016
There are other disconnects going back over longer time series.
Between the beginning of August 2011 and the end of February 2013 the share price deviated from the trend in iron ore, exhibiting a strengthening bias relative to a weakening bias for iron ore. In fact, in mid-September 2012 iron ore was over 40% lower than in August 2011 but the Kumba share price had barely changed.
The rand was relatively strong too, averaging R8,16 to the US dollar between August 2011 and February 2013 with little volatility compared to what we have seen in the past year or two.
Breakeven cash costs were closer to $70/ton in that period or 75% more than today.
The goodness of fit between the two data sets over this period is very poor with the correlation coefficient R² 0,019 and a Pearson product momentum correlation that is also 0,019.
However, there was sufficient cushion in the 2011 through 2013 period between the iron ore price and cost of production for Kumba to be comfortably profitable. Adjusted earnings of R4,2 billion in 2015 compare with R11 billion in 2014, R15,4 billion in 2013, R12,5 billion in 2012 and R17 billion in 2011.
The results for 2015 were a tale of two halves but the largest downward adjustment came in 2014 not 2015. Cash generated from operations fell by 43.4% in H1 of 2015 and by 19.6% in H2 but was down by 47.1% in H2 of 2014. Operating cash flow for the year was down 36.4%.
Net debt, helped by cessation of the divided, fell by 42% and I expect a net cash position in 2016.
Earnings in 2015, excluding the impairment of the Sishen mine and derecognition of a deferred tax asset, were R4 177 million or 1302 cents, a decrease of 62%. Adjusted earnings declined 51.8% in the first half and 76.9% in the second half. The H1 and H2 EPS split was 978 cents and 324 cents.
However, capitalisation of waste stripping costs boosted earnings by R1 billion or around 300 cents per share. There is a limit to necessary capex and other expenses without compromising long term viability. Deferred stripping capex at Sishen was R2,5 billion alone in 2015, which was 37% of total capex, but this will fall to R500 million in 2016 before rising to R1 billion in 2017 and R1,6 billion in 2018.
Sishen is transitioning to a smaller mine plant with additional exposed ore and less waste stripping of 150 million tons versus 230 million tons at a production rate of 27milion tons rather than 36 million tons. Life of mine remains up to fifteen years.
Kolomela is performing to plan with life of mine at 21 years. A new mine plan will see ramp-up of production to 13 million tons by 2017.
In Q1 2016, Sishen mine produced 34% less tonnage and Kolomela 9% less tonnage but guidance is unchanged for the year at 27 million tons and 12 million tons respectively with export sales guidance unchanged at 40 million tons.
Saleable grade from Sishen is 65.4% Fe and at Kolomela it is 64.4% Fe. Proven group grade is 65,5% and probable grade is 64,2% for a mix of 65,1% Fe. Kumba has a high iron or Fe content. Fortescue in Australia for example is 57,3% Fe blended across all mines on reserves and 56,3% on resource.
Kumba is profitable in a $50/ton iron ore market given the cost cutting and production optimisation. Down to $40/ton is a tough ask but not impossible to sustain for say a year. I’ve kept my estimates based on $50/ton in F2016.
I continue to assume no dividend payments until the market turns sustainably upward.
Kumba has high operational gearing to small movements in variables. At R16/$ and $45/ton iron ore then EPS is approximately 900 cents in 2016 and 750 cents in 2017. If iron ore is $40/ton then EPS is about 350 cents and 170 cents respectively. At $39/ton, Kumba is roughly cash breakeven in 2017.
I am estimating earnings of R3,5 billion or R11,06 per share in 2016 on the assumption of a $50/ton iron ore price and a R15/$ currency. If the rand averages R14/$, not unrealistic, and iron ore stays at $50/ton then earnings could be roughly R1,3 billion lower taking EPS to R6,90. If I assume a $45/ton iron ore price in 2017 and a R14/$ then EPS is around R5,38.
On 9 February 2016, in my synopsis of the F2015 result, I observed that the outlook for Kumba is not as gloomy as the share price was suggesting. I further mentioned that I thought it unlikely Kumba will run to losses and that with the dividend suspended then the balance sheet should have the strength to ride through the trough. I maintain that view.
Meantime, Kumba is up for disposal by Anglo America. Anglos other iron ore asset Minas Rio stays – for now – because it is unsaleable.
This erstwhile cash ATM for Anglo, having served its purpose, will arguably be better off as a stand-alone. An unbundling of the 69,71% stake in Kumba would be my preferred route but that would not realise cash for Anglo. A partial or total sell-down in a trade sale is possible too. The IDC and PIC collectively hold 15,45% of the shares and are interested observers.
We’ll see what happens. Kumba is a reasonable asset in a scenario of >$50/ton pricing for iron ore and this downturn is making it a more cost efficient one too.
At 12700 cents, the stock is pricing in a bit too much good news. Bi-polar speculators have done a 180º shift since January.
In my DCF, I have a fair value of 8700 cents per share in a $50/ton scenario – not that much different to the value I’ve had for a few months now. At $60/ton this goes to 14900 cents per share and at $70/ton it is 21100 cents per share.
Bottom line. You pays your money, takes your choice. The result could be anything on a continuum of variables. Potential corporate action will also be a prop for the share price.
The share has motored northwards from hugely oversold levels. My long standing Trading Buy is now a Trading Sell. If iron ore tumbles back to $60/ton, let alone $50/ton or even $40/ton, you don’t want to run the risk of being the lone manic when all others are suddenly depressive again.
Abbreviated estimates follow: