Introduction: narratives on the 2014 platinum strike
In 2014 South Africa's three largest platinum mining companies were partially shut down by a 70,000 person-strong, five-month long strike. Alongside closed-door negotiations between the companies and the Association of Mineworkers and Construction Union (AMCU), public debate revolved around competing narratives on the affordability of wage demands, the impact on the industry and the wider economic consequences.
This briefing uses financial data to assess the narratives employed by the three companies – Anglo American Platinum (Amplats), Impala Platinum (Implats) and Lonmin – for the mass media via their Platinum Wage Negotiations website (http://www.platinumwagenegotiations.co.za) and for the investment community via earnings releases and investor presentations. The exercise is not based on a binary opposition between narrative as fiction and numbers as fact, but rather follows Froud et al. (2006) in viewing corporate self-narratives and numbers as intertwined. Management stories are framed and corroborated (or undermined) by a process of selection and emphasis from financial data (Ibid., 133).
Communication strategy has become an important aspect of corporate management. Increased attention to shareholder value concerns has compelled executives to continually dialogue with major investors, analysts and the business press. Foregrounding of positive financial data corroborates narratives of success, while negative data must be explained, pushed to the background or drowned in positive news about future developments (Ibid., 122–136). If successful, this shapes wider narratives about an industry or company, influencing investor sentiment. However, levers to produce numbers corroborating narrative are limited, negative external developments cannot be controlled and disagreements over which ratios matter and what ‘shareholder value’ ultimately means are pervasive (Ibid., 36–48, 94–98).
For some companies the communicative task is more complicated still because, alongside appealing to the pecuniary interests of investors, they must address public concerns. In this instance, the platinum corporations created a picture of an industry in crisis to contend that wage demands were unaffordable, while simultaneously reassuring shareholders about the viability of the business (these questions are also taken up by Forslund in his forensic briefing on Lonmin's profit-shifting activities in this issue). This was set against the background of popular-political pressure over workforce poverty and high executive pay. The Platinum Wage Negotiations website, which was run by a public relations company on behalf of the three companies, addressed the media with accessible briefing papers emphasising the industry's acute difficulties: rising labour costs; headwinds in global platinum markets; and declining ore-grades on the Bushveld Igneous Complex (Baxter 2014; Platinum Wage Talks 2014a, 2014b, 2014c). Data were marshalled to demonstrate that ‘the reality is the platinum industry is in crisis’ – a crisis exacerbated by the strike (Platinum Wage Talks 2014d).
In contrast to this simple public narrative, the message to investors has been more complicated. The companies face a more sophisticated, interrogative audience where bad financial news – frequent for platinum in recent years – is seldom overlooked and management promises seldom forgotten. In these less public dialogues, company spokespersons offered reassurances: despite current difficulties the long-term prospects for global demand are improving; the strike was manageable given preparations in stockpiling and balance-sheet strengthening; and wage increases were in line with recent historic trends–albeit trends which were compelling a move toward industry restructuring away from more labour-intensive forms of production. Data underpinned a very different narrative of surmountable difficulties in a cyclical industry with better times ahead. Events since the strike took place have proven this to be unduly optimistic, at least in the short term. Accompanying the generalised rout in global commodity markets, the platinum price fell by a third to below $1000 per ounce in the space of the year to July 2015, making the pressures resulting from the wage increases even more acute and forcing producers to reassess their mining portfolios in an effort to cut costs.
This briefing explores these differing narratives of crisis and resilience. First, accounts data are used to overview value-added distributions from 2000–2013. Against the Platinum Wage Negotiations website presentations, which displayed company finances in a two-year time frame, this shows the current ‘crisis’ was preceded by a long boom where wages were manageable relative to shareholder rewards. In the following section, investor presentations on the strike and concurrent financial data illustrate a complex picture of corporate performance which, as a by-product of reassuring shareholders, disrupts the simplistic crisis narrative. The third section adds to this by analysing the wage settlement's costs, which are in line with recent historical trends.
Profits, distribution and labour's share of value-added
During the strike, the mining companies emphasised to the media the sector's aggregate cash losses in 2012 and 2013 and increasing wage bill (e.g. Baxter 2014; Platinum Producers 2014a).1 Wages have been increasing 8%–10% annually for many years (see third section) and the three companies have been under differing degrees of financial strain. A focus on losses (in 2012 and 2013) and on increasing costs is however a selective rendering of a cyclical industry, representative of a narrow time period. A longer-term view reveals a more complex picture of high profitability during the 2000–2008 commodities boom as platinum rose from $350 per ounce in mid 1999 to $2100 in mid 2008. As Figure 1 shows, the three companies enjoyed operating profit margins well above the Johannesburg Stock Exchange (JSE) top 40 average during these years. Shareholders benefited via large dividends, share buy-backs and rising stock prices.
Amplats, Lonmin and Impala operating profit margin (%) vs JSE 40 average. Source: Thomson Reuters Worldscope.
This section examines the recent financial history of the three companies through a value-added analysis of 2000–2013.2 Value-added is defined here as the monetary value left with a company after the cost of external goods and services used in production (Evraert and Riahi-Belkaoui 1998). It is therefore a measure of wealth created. This briefing uses a conventional additive calculation derived from the companies’ annual income statements, comprising labour costs (including pensions, medical costs and other employee benefits, alongside wages), plus pre-tax profit (comprising dividends, taxation and retained earnings), plus depreciation and amortisation, plus interest costs (Ibid., 2). Business performance, regulation and the bargaining power of the providers of capital, compared to labour and the state, determine relative distributions. This section first compares labour share of value-added to pre-tax profit share company by company, before examining cross-company aggregate distributions.
Figure 2 disaggregates Amplats’ 2000–2013 value-added. The company enjoyed booms from 2000–2003 and 2005–2008 when the profit share of value-added exceeded 50% and the labour share was around a quarter. The profit share slumped post-2009, with a 2012 operating loss. While labour costs rose from R10 billion in 2008 to R15.2 billion in 2013 (nominal), the dramatic increases in the labour share of value-added arose from a lower total value-added, which fell from R32.7 billion in 2008 to R21.3 billion in 2013, hitting a low of R10.7 billion in 2012 as a result of the aforementioned losses.
Anglo American Platinum, proportional composition of value-added. Source: company accounts, various years.
Implats (Figure 3), the world's second largest platinum producer after Amplats, displays a slightly different pattern. The period 2000–2013 also splits between a 2000–2008 boom and a subsequent slump in which the labour share of value-added grew towards and above 50%.
Impala platinum, proportional composition of value-added. Source: company accounts, various years.
However, Implats was more successful during the boom at maintaining a low labour share of value-added and a high profit share, while post-2008 total value-added plateaued rather than replicating Amplats’ saw-tooth fluctuations.
Lonmin (Figure 4) stands out as being under greater strain. The company suffered losses in both 2009 and 2012.
Nonetheless, for pre-2008, and in particular the early 2000s, Lonmin was highly profitable, with labour share maintained at around one-third and profit share around 50%. The stark reversal in fortunes post-2008 is in part due to the limitations of Lonmin's mining portfolio. With mines located entirely within South Africa and the Rustenburg area in particular, it has been more exposed to labour disputes and challenging geology. Amplats and Implats both own operations in Zimbabwe and other, more favourable, parts of the platinum belt.
Given that the time-series data for all three companies shows a split between two contrasting periods in the business cycle, it is useful to view 2000–2008 and 2009–2013 as aggregates to examine longer-term distributional issues. Figure 5 shows differing distributions of value-added aggregated from the three companies and adjusted for inflation, for 2000–2008, 2009–2013 and 2000–2013.3
Aggregate distribution of value-added from Anglo American Platinum, Impala Platinum and Lonmin, 2000–2008, 2009–2013, 2000–2013. Source: company accounts.
For both 2000–2008 and 2000–2013, the labour share of value added was relatively low, at 29% and 37% respectively. The average labour share of gross value-added for the South African economy was around 50% for 2000–2013 (OECD 2014). The post-crisis years have moved the tri-company aggregate closer to this national average.
These data on their own do not necessarily suggest the mining companies should have paid higher wages during the boom, even less that they can easily afford to do so currently. The value-added analysis here also does not take into account capital expenditure commitments which restrict free cash flow. However, they do complicate the platinum producers’ public narrative. First, if the industry's crisis is understood as profitability destroyed by higher labour costs, this is a recent development following a longer period of high profitability. The profit share of value-added for 2000–2008 amounts to over 60%, of which just under half – 30% of the total – was distributed to shareholders. This is significantly above the economy-wide average for dividends as a share of gross value-added, which was around 20% during this period. This raises political questions about whether a more equitable or socially beneficial distribution could have been achieved in the boom, or could still be achieved during any return to similar levels of profitability. Second, it raises questions about responsibility. Striking workers were widely criticised for irresponsible demands and actions, but given the well-documented poverty on the platinum belt the size of distributions to shareholders during the boom could also be questioned. While the companies can rightly argue wages increased over this period, as a share of total wealth produced distributions to labour in many years are low.
The historical value-added analysis represents the undisclosed back story to the strike. The platinum producers focused public attention on immediate financial difficulties and future obstacles to growth. The recent past prior to 2008 remained largely invisible and outside debate. Its inclusion may have raised important questions about how best to share risk and reward between owners, employees and the state during good as well as bad times.
The impact of the strike on the producers
If the past has remained largely undiscussed in debates on the strike, the present financial performance of the companies is unavoidably foregrounded and has posed a complex narrative management challenge.4 The crisis narrative has been maintained, but investors and bank analysts have simultaneously been reassured that the companies have weathered the strike and will recover. This reassurance comes in the context of a difficult period for the three mining companies in terms of their attractiveness to investors. As Figure 6 shows, having outperformed the JSE All Share Index for most of the past decade, delivering large capital gains for those selling their shares, the companies’ share prices have in recent years declined relative to the index, with management consequently coming under increased scrutiny.
Share price movements (weekly) for Anglo American Platinum, Impala Platinum and Lonmin (2000 = 100). Source: Thomson Reuters.
The contrary narratives on crisis and resilience are created from different acts of selection and emphasis from company data, and coexist owing to the relative esotericism of the discourse used in discussions with analysts and investors relative to the mass media. For example, as was widely reported, gross profits were dramatically reduced by the strike and much potential production was foregone. Revenue, however, for Amplats and Implats, has remained relatively stable owing to the sales of stockpiled metals (see Figures 7 and 8), built up partially in expectation of industrial action. This matter was frequently absent from public debate but explored in detail by stock market analysts.
Production, sales and cash flow: Amplats, Implats and Lonmin. Notes: Amplats estimates for production and sales are drawn from their respective third quarter production reports (Anglo American 2014a; Lonmin Plc 2014) and cash flow from an average of analyst report estimates (CIBC 2014; Deutsche Bank 2014; HSBC 2014; UBS 2014).
As of the end of 2013, Amplats held 430 koz (thousand ounces) of refined platinum and large pipeline stocks and was braced for the strike (Barclays 2014a). With refining mostly unaffected by the stoppages, this meant the company could fulfil contracted sales (80% of total sales) for the 2014 calendar year. Barclays’ analysts predicted spot market sales would be curtailed only at the end of June if the strike was ongoing (Ibid.). CEO Chris Griffith said by the end of the strike Amplats had drawn down 190 koz of refined inventory and 110 koz of pipeline inventory (Thomson Reuters 2014). This indicates substantial remaining room to manoeuvre for Amplats. Following 2012's six-week strike, Implats increased inventories by around 50% to 150 koz during financial years 2012 and 2013, while Lonmin ended the 2013 financial year with $449 million worth of inventories (over 80% of which were only partially processed), up from $260 million at the 2012 financial year-end.5
The sale of inventories supports revenues during a stoppage but brings a cost of sales charge onto the income statement, therefore depressing the operating profit margin. Amplats’ revenue increased 15% during January–June 2014 compared with the same period in 2013 (R24.1 billion to R27.9 billion) but the operating profit margin fell from 11.9% to 3.3% despite a large decline in on-mine cash operating costs (from R13 billion to R10.7 billion). This was largely due to a R4.7 billion debit from a decrease in metal inventories, and from a R2.8 billion credit the previous year. Implats’ revenue remained roughly stable for the financial year 2014 at R29 billion in comparison with R29.8 billion in 2013. However, the operating profit margin decreased from 15.8% to 11.2% (R4.7 billion to R3.2 billion). In contrast, Lonmin's revenue fell from $1.5 billion in the 2013 financial year to $965 million in 2014, with operating profit falling from $147 million in 2013 to −$255 million. Revenues were also buoyed by an approximate 7% rand–dollar exchange rate weakening between end June 2013 and end June 2014 (R10/$1 to R10.7/$1).
Another way this process can be shown is to contrast significantly decreased production with a smaller decline in sales and the resulting cash flow (sales and cash flow were bolstered by the drawing-down of stockpiles, indicating how revenues can be protected even if operating profit margins are lower). This is shown in Figure 8, where for all three companies 2014 sales exceed production. Lonmin stands out with the sharpest decrease in both production and sales owing to the geographic concentration of its operations in the Rustenburg area. With 100% of output volume impacted, it forewent approximately 348 koz of saleable platinum owing to the strike. Amplats and Implats lost 39% and 38% of output volume respectively owing to their more diversified operations, with Amplats foregoing approximately 440 koz of equivalent refined production and refined production at Implats 312 koz lower than projected (Barclays 2014a).6
Idle productive capacity during the strike resulted in what the companies call ‘cash burn’ – cash expenditure on operations not yielding output – on management expenses, non-AMCU workers, maintenance, power and water. Amplats estimated cash burn during the strike of R3.4 billion at the affected mines (Thomson Reuters 2014), pushing unit cost per ounce of production for the six months of the strike up to around R18,000 from R17,053 in 2013 – a smaller increase than between 2012 and 2013. Implats was the most effective in controlling cash costs, reportedly reducing them by 70% to 75% during the strike (Barclays 2014b); the company estimates a cash burn of R1.1 billion. This, together with other cost inflation, pushed unit cost of production up by 17.6% to R19,430 per platinum ounce in the financial year 2014 compared with R16,526 in 2013. Lonmin estimates the strike cost it $322 million in extraordinary costs and led to a dramatic 60% increase in unit cost per platinum group metal (PGM) ounce from R8832 in 2013 to an estimated R14,131.
The companies also emerged from the strike without a dramatic deterioration in balance sheet strength. Amplats’ net debt was R12.4 billion as of the end of June 2014, down from R13.2 billion at the end of June 2013 but up from R11.5 billion as of December 2013 (the end of the 2013 financial year). Implats’ 2014 net debt position of R3.5 billion remains in line with the R3.4 billion of 2013. Balance sheet concerns were most acute for Lonmin, which raised $767 million in a December 2012 rights issue in part in response to the damage done by the August 2012 strike. As a result of this balance sheet strengthening, Lonmin ended the 2013 financial year with $210 million net cash; the company ran down this cushion to $71 million by the end of March 2014 and zero by the end of June 2014 (Deutsche Bank 2014; Lonmin Plc 2014). The risk that Lonmin would need to raise new equity capital or breach debt covenants was perceived by market analysts to be small at the time of writing (CIBC 2014; Deutsche Bank 2014; UBS 2014). However, the subsequent crash of the platinum price during 2015 forced both Lonmin and Impala to turn to investors for cash, with the former in particular experiencing severe financial strains.
In sum, while Lonmin stands out for suffering most acutely from the strike, the immediate financial impact on producers was not catastrophic. Management suggested to analysts that the strike may even have had some positive side effects. Movements in the platinum price throughout the strike were moderate, and in July 2014 Amplats CEO Chris Griffith told investors:
Using data published from the market balances each year, including the 2014 estimate, shows that above ground [PGM] stocks are now substantially lower. As a consequence, the fundamentals for the platinum commodity have substantially improved compared to the previous years of an oversupplied market. (Thomson Reuters 2014, 5)
This, the company said, according to Barclays analysts, would mean that ‘price moves will be more rational going forward’ (Barclays 2014a, 8).
The other two companies and several industry analysts stressed that medium- to long-term market conditions for PGMs were likely to be good. The Johnson Matthey Platinum 2013 Interim Review predicted gross demand for platinum would hit a record high within the year, spurred by ‘a strong recovery in sales to industrial users and unprecedented offtake by investors’ with ‘little overall growth in recycling’ (Johnson Matthey 2013, 4). Amplats CEO Chris Griffith told an investor conference call in February 2014: ‘We see industrial growth driving short and medium-term deficits in platinum that should drive price recovery as the supply remains constrained’ (Thomson Reuters 2014, 5). Lonmin's Chairman Roger Phillimore expressed similar views (Lonmin Plc 2013). Thus, as one narrative of make-or-break crisis circulated in the public sphere, in the investment community management portrayed the strike more as a temporary obstacle, a storm which, as their financials attest, has to varying degrees been weathered. The following section examines the more specific impact of the wage settlement.
Impact of the wage settlement
On 24 June 2014 the three major platinum producers confirmed they had agreed to a three-year wage settlement with AMCU. Lonmin CEO Ben Magara grimly stated there were ‘no winners’ from the deal (SAPA 2014), with commentators suggesting the sector would now require restructuring as a result of irresponsible wage demands (but see also Forslund's Lonmin briefing on this notion of ‘irresponsible demands’). However, as this section shows, the wage increases were broadly in line with recent historical trends. If rising labour costs represent a crisis for the industry, it is a slow-burn crisis which producers have seen coming. The wage increases from the settlement, which decrease over time, are shown in Table 1.
| Y1 2013/14 | Y2 2014/15 | Y3 2015/16 | Average | |
|---|---|---|---|---|
| Amplats | 10.5% | 7.7% | 7.1% | 8.4% |
| Implats | 10.0% | 9.2% | 8.2% | 9.1% |
| Lonmin | 12.9% | 8.8% | 8.2% | 10.0% |
Notes: For Amplats and Lonmin, company estimates have been used (Anglo American Platinum 2014b; Lonmin Plc 2014), for Implats authors' estimates using data from Bowman and Isaacs (2014). For Amplats the company and authors' estimates are very similar, for Lonmin the authors estimate 10.2%, 8.5% and 7.8% increases over the three years, on average 8.9%, and on the basis of available data the Lonmin company estimates seem peculiarly high.
According to Barclays Capital's calculations (2014b), the average wage increase of 9% over the next three years is only marginally higher than the 8.7% average over the preceding 18 years (1996–2013) and lower than the 9.6% average for the preceding six years (2008–2013) (Figure 9).
This is not dramatically altered when accounting for inflation: the settlement's wage increases average 3.2% above the expected rate of inflation for 2014–2016, compared with a 2.7% average for 1996–2013 and 3.1% for 2008–2013 (Figure 10).7
Barclays Capital analysts (2014b, 2) were among several noting that:
Very little has actually changed in the dynamics of the South African Platinum industry over the last 5 years. Wage inflation remains at c.10% per annum … underlying mining inflation on the Western Limb is likely to remain at c.10% pa through to 2019.
This conclusion is at odds with the public narratives of the producers during the strike, who maintained they could not budge from earlier offers of average increases of 7.2% (21 January 2014), 7.4% (29 January 2014) and 8.4% (27 April 2014). Given that the deal sustains historic trends, one may also question AMCU's victory claims. However, the settlement was above the original offer and, unlike some previous settlements, skewed towards the lowest earners. Figure 11 shows the average increases for all three companies combined on the basic and total packages of A, B and C level workers, the three tiers covered by the agreement with A being the lowest paid.
That the wage settlement aligns with historic averages does not make it any more affordable for the companies. Rather it highlights that the difficulties producers face did not begin with the strike. They are caught between shareholder value pressures and the erosion of their traditional low-wage business model. This business model has been pursued by South African mining not only because of the historic availability of cheap migrant labour, but also because it avoids the investment risks involved in mechanisation as mass redundancies serve as a safety valve during downturns. Of the three companies, only Amplats seems capable of radically restructuring its operations away from this model (see also Stewart's briefing on the variable constraints of mechanisation in the platinum industry). For Amplats, mining assets more amenable to open-cast mechanisation on the northern limb of the Bushveld Igneous Complex are expected to provide higher returns in future, meaning its more labour-intensive Rustenburg and Union mines (possibility also Bokoni and Pandora) can be sold. Notably, Amplats’ motivation for the sales is not an absolute inability to achieve profitability but inadequate shareholder value maximisation. The company has said it will only pursue projects providing a return on capital employed of 18% or more (Barclays 2014a). For Implats and Lonmin, whose mines are less easily mechanised, these historic problems will be harder to resolve.
Conclusion
This briefing shows how a selective emphasis on underlying financial data enabled the platinum companies to present different narrations of the strike to different audiences. To the national media, the companies presented themselves as in a state of acute financial crisis. This was backed up by data reflecting short-term pressures on their cost base from labour. Meanwhile, the general sentiment expressed by analysts and company executives in less public forums was that the strike was challenging but manageable while the demand outlook for PGMs was positive. This dissonance highlights the conflicting pressures placed upon the modern public limited company in general and South Africa's platinum mining companies in particular. The companies presented their rejection of the ‘reckless’ wage demands as the responsible position and employed a narrative of a sector in crisis in order to do so. Simultaneously, they needed to reassure investors of their commitment to shareholder value maximisation.
No one is projecting a return to the boom conditions of 2000–2008, but future periods of increased profitability are clearly deemed possible by the industry given expected rises in PGM prices in the long run and the development of new low production cost mining operations. Examination of the longer-term financial history of the sector and the distributional choices made by the companies during the last boom raises questions about how exceptional profits are divided between the state, labour and shareholders.