Pan African Resources — Analysing H224 and looking forward to FY25

PAF’s FY24 results, announced on 11 September, were closely in line with our forecasts for headline EPS (HEPS) and EPS and also guidance, as provided in its announcements of 29 July and 5 September.

Production in H224 was confirmed at 87,581oz, representing a 5.6% increase compared with H223, while guidance for FY25 was reiterated at 215–225koz at an all-in sustaining cost (AISC) of US$1,350–1,400/oz – not more than 3.4% higher than during FY24 – albeit with the proviso that ‘the delay in the commissioning of Evander Mines’ subvertical shaft, scheduled to be completed during September 2024, could impact guidance by approximately 5,000oz’. Exhibit 1, below, shows Pan African’s group production in H224 compared to H223 and H124, and also our forecast for a further 12.1% output growth in FY25.

Output for each of PAF’s four producing operations in H224 was exactly in line with guidance, with the exception of a tiny amount of production (183oz) from surface sources at Evander. However, with the exception of the Barberton Tailings Retreatment Project (BTRP) that output was generally achieved via processing slightly higher volumes of tonnage at slightly lower grades than we had anticipated. In the case of the BTRP, both tonnage and grade were lower than our prior expectations, but were offset by the highest metallurgical recoveries since H116, as shown below:

Relative to a 2.8% year-on-year increase in tonnes milled/processed and a South African inflation rate of 5.1% (June-June), Pan African’s electricity costs (14.1% of the total) increased by 66.4% during the year, while processing and metallurgical costs (21.7%) increased by 59.9% and security costs (3.2% of the total) increased by 27.7%. As a result, unit costs per tonne of ore milled/processed were higher than we had anticipated at Barberton, Elikhulu and the BTRP and, while we still expect these to moderate over time, we now expect them to moderate from a higher base. By contrast, Evander milled 14.2% more ore in H224 compared to H124 for 7.0% less cost in rand terms to result in an 18.6% lower unit cost, enviably demonstrating both economy and the benefits of incremental production on a relatively high fixed cost base. Within this context, the overall increase in the group’s AISC, from US$1,309/oz in FY23 (restated) to US$1,354/oz in FY24 – an increase of only 3.4% – was a creditable achievement.

The other noteworthy feature of the group’s results was the price of gold that PAF received for its output, which was reported as US$2,015/oz for the full year (a 3.0% discount to the average gold price of US$2,078/oz during the period) and calculated by Edison as US$2,077/oz for H224 (a 5.8% discount to the average gold price of US$2,206/oz). These discounts arise from the accounting treatment of the contract liability related to the fixed-price forward sales associated with Pan African’s ZAR400m Mintails financing facility, which is different to that relating to ‘normal’ derivatives. Derivatives are subject to mark-to-market fair value accounting, with any adjustment going through the ‘other income/expense’ line in the P&L. By contrast, fixed-price forward gold sales result in the fixed gold price for these ounces being recognised in revenue – there being no equivalent mark-to-market fair value accounting. However, synthetic forward sales of gold, which are composed of a series of options (puts and calls at the same price) are also accounted for as a ‘plain vanilla’ forward sale (and not a derivative instrument), with the fixed gold price recognised in revenue as the ounces are delivered since, despite being a derivative instrument, they are, in substance, identical to a ‘plain vanilla’ fixed price forward sale.

By contrast, for ease of comparison, Edison forecasts revenue based on the prevailing price of gold with all derivative-type profits/losses being booked through the ‘other’ line. Although this is not strictly in accordance with accounting standards, it allows the underlying performance of the operating company to be distinguished from the volatility created by derivative-type profits and losses, which are then presented as a single line item. In the case of H224, Pan African sold 1,154 fewer ounces than it produced in H224, which cost it c US$2.5m in revenue at the average gold price during the period (albeit we would expect this figure to be recouped in coming quarters). In addition, we estimate that its Mintails financing arrangements depressed revenue by a further c US$11.1m (86,427oz sold at US$2,077/oz in H224 rather than US$2,206/oz). Adding back US$11.1m to reported revenue implies like-for-like revenue of US$191.0m in H224, compared to US$196.8m forecast and US$179.8m actually reported. Note that this adjustment, with revenue presented on a ‘plain vanilla’ basis and derivative-type losses all shown in the ‘other’ line, is shown in the column denoted ‘H224 (adjusted)’ in Exhibit 4, below. On this like-for-like basis, revenue was US$5.8m (or 1.5%) less than we had forecast for both H224 and FY24 (of which c US$2.5m could be accounted for by unsold gold), while costs were US$12.1m (or 5.8%) higher for the full year, leading to a US$17.2m (or 10.8%) negative variance in mining profits. However, this was essentially offset by an US$8.2m positive variance in ‘other’ expenses (the difference between a US$22.5m loss forecast and an underlying US$14.3m loss reported), a US$2.0m positive variance in royalty costs, a US$4.1m positive variance in net finance income and a US$1.9m positive variance in the overall tax charge to result in actual H224 and FY24 attributable profit within US$0.6m, or 1.0%, of our prior forecast.

A complete analysis of PAF’s H224 and FY24 results on both an ‘as reported’ and an ‘adjusted’ basis (with all derivative-type losses taken through the ‘other’ line) is provided in Exhibit 4, below:

Note that the effect of the Mintails contract liability on prior periods was small to the point of being negligible given that the contracted price of gold was close to the actual price of gold. The exclusion of US$4.1m in ‘other income’ from ‘H224 (adjusted)’ results would also have brought Edison’s forecast of a US$22.5m loss more closely into line with the US$14.3m loss imputed. Otherwise, on this comparative (ie ‘adjusted’) basis, normalised HEPS were 7.9% lower than our estimate for the full year. However, they were still nevertheless a record for Pan African for both a full-year and half-year period and well to the top end of the range of analysts’ expectations of 3.0–5.7c/share with a mean of 4.5c/share (source: LSEG Data & Analytics, 6 September 2024).

From the perspective of PAF’s four principal operations (ie excluding the negligible contribution from Evander surface sources), performance held steady at both Barberton and Elikhulu, but fell away at Evander (which is the operation most affected by the Mintails contract liability – see Exhibit 2), while the BTRP announced record adjusted EBITDA in both rand and US dollar terms:

PAF has two organic growth projects currently underway (namely the MTR project within the Mintails Soweto Cluster and the Evander 24 to 26 Level expansion project) and one more immediately in prospect (the Sheba Fault project). Beyond these, it has the Egoli and Fairview sub-vertical shaft projects at feasibility study stage followed by Rolspruit, Poplar and Evander South also available for eventual development.

On 1 August 2023, PAF announced that all conditions precedent for its ZAR1.3bn senior debt facility, designated for funding the group’s MTR project, had been fulfilled, thereby completing the full upfront funding package of ZAR2.5bn. Since then, PAF has reported that construction has been progressing on time and within budget, with commissioning and steady state production still on track for December 2024.

On 9 May, PAF reported that it had updated its financial model for Mogale (relative to the initial definitive feasibility study model) to reflect the latest operating cost updates, as well as a ZAR19.00/US$ forex rate and a US$2,200/oz gold price. The result was:

In addition to its update on the MTR, in May PAF also announced the results of an internal pre-feasibility study (PFS) for the Soweto cluster that was completed in March 2024, based on drill results from the 2L16 and 2L24 tailings storage facilities. The PFS considered a number of options, of which the most feasible was considered to be the processing of the Soweto cluster material at the MTR plant. In this case, the MTR plant’s capacity could be expanded to process 1Mtpm of feed material (cf the current design capacity of 800ktpm) to result in a mine life of 21 years for the combined Mogale and Soweto cluster resources. The resultant tailings could then be deposited into the expanded Mogale tailings storage facility at the West Wits pit and 1L23-25 footprint.

Other specific results of the PFS were as follows:

These changes have already been incorporated into our financial model since May. In the meantime, Pan African is proceeding with the necessary permitting and servitudes required for the re-mining and processing of the Soweto cluster, with a final investment decision anticipated ‘in due course’.

Progress at Evander’s 24 to 26 Level underground expansion project remains on track, with the following notable achievements during H224:

Following an internal process to consider feedstock sources for the BTRP, final drilling and metallurgical test work results were recently retrieved from Bramber’s dormant tailings storage facility (TSF), which will increase the life of the BTRP from two to seven years.

In the meantime, Pan African will focus on developing the Sheba Fault decline to access high grade mineral reserves to supplement Barberton Mines’ production in the medium term and longer term. Preliminary optimisation work at Sheba and Western Cross indicates:

The Egoli project at Evander Mines’ 7 Shaft is a standalone underground operation that will utilise existing mining and metallurgical infrastructure, including 7 Shaft’s hoisting systems and processing facilities at Kinross’s metallurgical plant.

In the light of these developments (including PAF’s unchanged guidance for FY25), we continue to forecast that group production at PAF will reach c 250koz per year in 2026 and push normalised headline EPS (HEPS) to beyond 7.00c per share (see Exhibit 10).

After three years of holding its dividend at ZAR0.18/share, PAF declared a dividend of ZAR0.22/share (c 1.2584 US cents or 0.9476 pence per share at prevailing forex rates) for FY24 for approval at its forthcoming annual general meeting (AGM) on 21 November. In addition, the company explicitly set out a dividend policy targeting a payout ratio of 40–50% of net cash generated from operating activities, after providing for the cash flow impact of capital expenditure (reduced by externally funded capital), contractual debt repayments and the cash flow impact of one-off items (discretionary rand cash flow). In proposing a dividend, the board will also take account of the company’s financial position, prospects, satisfactory solvency and liquidity assessments and other factors deemed by the board to be relevant at the time.

FY24’s dividend of ZAR0.22/share followed three years of a static dividend at ZAR0.18/share, during which time the company was either in the midst of, or anticipating, a period of material capital investment. Within this context, FY25 should be the last year in the foreseeable future of material capital outlay for PAF. Where before we had assumed that Pan African would pay out a flat dividend in FY25, followed by its maximum potential dividend from FY26 (for valuation purposes), in the light of the company’s dividend policy statement, we have now assumed that it will declare a dividend according to its restated policy in FY25–28 followed by its maximum potential dividend from FY29. In FY25 we believe that this will continue to be relatively conservative as capex remains elevated at the same time that PAF will be required to make debt capital repayments. For FY26 and beyond, however, we believe that the possibility exists for materially increased dividend payouts as production increases, operating costs flatten and capital expenditure falls away.

In deriving our longer-term estimates for PAF over the life of its operations, we have made a number of changes, the most material of which are summarised below:

In addition to changes to our immediate operational assumptions, we have adjusted our long-term foreign exchange rates (in real terms), to reflect the recent ‘above trend’ strength of the rand (and to a lesser extent sterling) against the US dollar:

In the aftermath of these changes, our absolute valuation of PAF (based on its existing four producing assets plus the 25 and 26 Level project and Mogale) has reverted to its pre-May 2024 level of 40.93c (cf 49.39c in July – see Exhibit 12), which is based on the present value of the estimated potential dividend stream payable to shareholders over the life of its mining operations (applying a 10% discount rate to US dollar dividends).

This may be compared with the equivalent graph from July 2024, as follows:

Readers’ attention is drawn to the flattened profile of dividends assumed in FY25–28 in our updated valuation relative to our July 2024 valuation, which have been calculated with due regard to PAF’s dividend policy, but which has resulted in a reduction in valuation of 3.02c/share as a result of the deferral of higher dividends.

Including its other growth projects and assets, our updated total valuation of PAF as a whole is provided in Exhibit 12, below.

An analysis of the change in valuation according to each separate adjustment to our assumptions is provided in the graph below:

Exhibit 14 below depicts PAF’s average share price in each of the financial years from FY10 to FY24 and compares this with HEPS in the same year. For FY25 and FY26, the current share price (33.65p) is compared with our forecast normalised HEPS for those years. As is apparent from the chart, PAF’s price to normalised HEPS ratios of 6.1x and 6.7x for FY25 and FY26, respectively, remain close to the bottom of its recent historical range of 4.1–14.8x for the period FY10–24 and consistent with its range of 4.1–7.0x for the period FY19–24:

If PAF’s average year one price to normalised EPS ratio of 8.2x for the period FY10–24 is applied to our normalised earnings forecasts, it implies a share price for PAF of 44.70p in FY25 followed by one of 41.17p in FY26. Stated alternatively, PAF’s current share price of 33.65p, at prevailing foreign exchange rates, appears to be discounting FY25 and/or FY26 normalised HEPS of 5.48c per share (cf our forecasts of 7.28c and 6.71c, respectively).

In the meantime, it may be seen that PAF remains cheaper than its London- and South African-listed gold mining peers on 80% of comparable common valuation measures (29 out of 36 individual measures in the table below) regardless of whether Edison or consensus forecasts are used.

Alternatively, applying PAF’s peer average year one P/E ratio of 11.1x to our normalised HEPS forecast of 7.28c per share for FY25 implies a share price for the company of 61.03p at prevailing foreign exchange rates. Applying its peer average year two P/E ratio of 8.6x to our normalised HEPS forecast of 6.71c per share for FY26 implies a share price of 43.33p.

Pan African reported net debt of US$104.4m on its balance sheet as at end-June 2024 (cf US$61.7m as at end-December 2023 and US$22.1m as at end-June 2023), which equated to a gearing ratio (net debt/equity) of 28.6% (cf 18.8% at end-December 2023 and 7.5% at end-June 2023) and a leverage ratio (net debt/[net debt+equity]) of just 22.2% (cf 15.8% at end-December 2023 and 7.0% at end-June 2023), after cash flow from operating activities of US$109.1m before dividends (cf US$45.5m in H124, US$88.5m in H223 and US$31.6m in H123). Capex guidance for FY25 is ZAR2.27bn (c US$130.0m at prevailing foreign exchange rates). Beyond that, we forecast that PAF will continue to generate cash from operations comfortably above the US$100m pa level (and potentially close to US$200m pa) into the foreseeable future, such that we calculate that net debt will peak in CY24 (in reality at end-December 2024), before being eliminated in FY26, by which time we assume that capex will once again have returned to near-sustaining levels.

Note that, for the purposes of these forecasts, we have assumed that FY24’s dividend will be paid in FY25 and that our newly increased FY25 dividend estimate will be paid in FY26.

Including all other components, total net debt as at end-June was US$106.4m (cf US$64.3m at end-December, US$22.0m at end-June and US$53.7m at end-December 2022), as shown below:

The US$2.0m difference between net debt of US$106.4 and net debt of US$104.4m as per Exhibit 19 is accounted for by the US$2.0m of ‘other’ items included in Exhibit 17.

In the meantime, the group remains very comfortably within its revolving credit facility debt covenants: