Wheaton Precious Metals — Honing forecasts

WPM’s Q121 results were characterised by record quarterly revenue, the repayment of the group’s revolving credit facility in its entirety and a third successive increase in the quarterly dividend (to US$0.14/share). In a departure from recent trends, it was the silver division that led the way in terms of increased production and sales as the gold division lagged, principally as a result of scheduled and unscheduled maintenance at Salobo that took longer to undertake as COVID-19 limited Vale’s ability to mobilise contractors and that also restricted mine movement. Otherwise, production and sales were closely aligned and ounces of gold and silver produced but not yet delivered to WPM reduced in both cases, providing evidence of a return to more normal operating conditions in the wake of the worst depredations of the COVID-19 pandemic in FY20.

From an operational perspective, production from Penasquito and San Dimas was affected by lower grades (in the latter’s case on account of increased ore development activities), while that at Salobo and Sudbury was affected by lower plant throughputs – at Salobo on account of the scheduled and unscheduled maintenance disruptions, as detailed above. Antamina benefited from higher grades, while Voisey’s Bay made a maiden contribution of cobalt to WPM. Elsewhere, Constancia continued to be affected by the delay in accessing the high-grade Pampacancha pit. However, on 7 April, its operator Hudbay announced that the final land user agreement for Pampacancha had been completed and that consequently it had full access to the site and had begun pit development activities. The following mines outperformed our expectations during the quarter in terms of production: Penasquito, Antamina, Constancia (silver), Los Filos, Yauliyacu, Neves-Corvo, Aljustrel, Cozamin, Minto and 777. In addition, all six of WPM’s partners’ mines that were directly affected by shutdowns and suspensions in 2020 (namely Penasquito, San Dimas, Antamina, Constancia, Yauliyacu and Los Filos) were operating at, or close to, full capacity.

Since our last quarterly results note on Wheaton (see A solid start, published on 10 May 2021), there have been a number of developments pertaining to the mines with which it has streaming agreements. A summary of these is as follows:

Each of these developments and Edison’s treatment of them is considered below, in order of priority. In addition, we have adjusted our Q221 forecasts for actual, rather than forecast, metals prices, as shown below:

On 1 June, a union representing approximately 2,500 maintenance and production workers at Sudbury went on strike after rejecting a pay offer. Talks have resumed between the union and company after a second offer was also rejected later in June, with Vale saying that it needs changes to the contract to justify future investment at Sudbury. However, the union has responded by saying that Vale’s contract offers contain concessions that are not needed since the Sudbury operation is so profitable.

On 19 July, Vale announced that production of nickel and copper at Sudbury were 9.0kt and 13.0kt respectively – consistent with an approximately one-third cut in output relative to prior quarters:

The average of the four declines in Exhibit 2 is 33.0%, which is close to the pro rata decline expected for production over two, rather than three, months. While these comparative figures are not definitive proof of a similar decline in the output of gold, they are strongly indicative of it. Gold produced at Sudbury attributable to Wheaton has averaged 6,953oz per quarter over the past eight quarters (albeit within a wide range from 3,798–9,360oz per quarter). A one-third decline in Q221 would imply production of c 4,635oz in Q221. However, this may be tested against two further pieces of Vale information:

In the light of the strike therefore, Edison has reduced its estimate of gold produced and delivered to Wheaton from Sudbury in Q221 to the most conservative of the above calculations, ie 4,635oz (cf 10,000oz previously). In addition, while we are hopeful of a near-term resolution of the dispute between the Sudbury union and Vale, we note that a similar strike in Canada in 2009–10 lasted for a full year. As a result, we have reduced our Q321 estimate of deliveries of gold from Sudbury to Wheaton to zero (cf 5,900oz previously). For the moment however, we are retaining our forecast of 10,000oz gold for Q421, although we will keep this under review as the situation develops.

Operations at Salobo were adversely affected by a backlog of maintenance and repairs in Q221. On 19 July, Vale announced that production of copper at Salobo in Q221 had been 38.7kt, or 13.5% higher than the 34.1kt produced in Q121. Pro rata with the amount of gold delivered to Wheaton in Q121, this increase would imply the delivery of 52,916oz from Salobo to Wheaton in Q221.

Alternatively, Vale announced that it had sold 84,000oz of gold as a by-product of its copper operations in Q221. While this aggregates the gold from both Salobo and Sossego, an analysis of gold delivered to Wheaton from Salobo in recent quarters suggests that between 84.7% and 94.2% (average 89.5%) of the gold sold from the two mines is produced at Salobo. That being the case, it suggests that 75,138oz of the 84,000oz sold from the two mines in Q221 could be attributed to production at Salobo, of which 75%, or 56,353oz, could be attributable to Wheaton under the terms of their streaming agreement. This is within 6.5% of the figure calculated in the paragraph above and compares with our prior forecast of 59,250oz attributable to Wheaton.

As a consequence, Edison has reduced its estimate of gold produced at Salobo attributable to Wheaton from 59,250oz to 52,916oz and its estimate of gold sold from 59,250oz to 56,353oz.

On 22 July, Newmont announced, among other things, that it had produced 7,428koz of silver at Penasquito in Q221 and that it had sold 7,615koz of silver. Pro rata with its 25% streaming interest, these figures imply production attributable to Wheaton of 1,857koz silver (Ag) and sales of 1,904koz Ag during the quarter, which were a modest 206koz and 159koz below our prior forecasts of 2,063koz Ag for production and sales, respectively.

On 22 June, Equinox Gold announced that mining activities at its Los Filos Mine in Mexico had been temporarily suspended as the result of illegal blockades by a group of unionised employees and members of the local community, both of which were demanding payments in excess of their contractual agreements. As such, the blockades appeared to be a carbon copy of the blockade that caused a temporary suspension of mining activities in Q420, but which was raised on 23 December, after which normal operations resumed. Subsequently, on 26 July, Equinox announced that the union blockade had been removed and that access to the mine and mining operations had been restored. However, it also noted that, ‘Certain members of the Xochipala community remain illegally camped near the Guadalupe and Bermejal open pits, disrupting mining activities in this area of the operation only’, and that, ‘Los Filos representatives continue to engage with government and Xochipala representatives to achieve resolution.’

As a result of the blockades, we have reduced our production estimate for Los Filos by one week, or 2koz, in Q221, although we have left our sales forecast unchanged (on the basis that sales tend to lag production chronologically). By contrast, we have reduced our production estimate by one month (or 16koz) in Q321, as well as assuming that the over-sale of silver relative to production in Q221 will be redressed in Q321 (ie silver sales will be reduced by 18koz in Q321 as the inventory of ounces produced but not yet delivered is built back up).

On 19 and 28 July, Vale announced, among other things, that it had produced 463kt of cobalt at Voisey’s Bay in Q221 and that it had sold 568kt. Pro rata with its 42.4% streaming interest, these figures imply production attributable to Wheaton of 433klb cobalt and sales of 531klbs cobalt during the quarter (cf our prior forecast of 525klb cobalt each).

As noted above, Constancia was adversely affected in Q121 by the delay in accessing the high-grade Pampacancha pit. However, on 7 April, its operator Hudbay announced that the final land user agreement for this deposit had been completed and that it had full access to the site and had begun pit development activities. Hudbay is scheduled to release its Q221 results on 9 August. In the meantime, while we believe that production of gold and silver attributable to Wheaton from Constancia will increase in Q221, we expect that it will be another quarter before we see this increased production translate meaningfully into sales.

In the light of these recent developments, Edison’s updated forecasts for WPM for FY21 are as shown in Exhibit 3, below. The forecasts assume that operations will continue throughout the remainder of the year without major COVID-19 induced interruptions. Apart from precious metals prices, the principal remaining risk to our forecasts relates to the extent to which sales differ from production and therefore the extent to which inventory (in the form of ounces produced but not yet delivered to WPM) either increases or decreases during the course of the year.

Readers should note that, consistent with past practice, for the purposes of FY21 we are assuming production and sales are closely aligned and that there is little or no change in the level of ounces produced but not yet delivered. Within this context, our basic EPS forecast of US$1.49/share for FY21 is closely in line with the consensus forecast of US$1.51/share (source: Refinitiv, 30 July 2021) and towards the middle of the range of analysts’ expectations of US$1.36–1.66 per share for the period:

Our basic EPS forecast of US$2.10/share for FY22 (see Exhibit 10) compares with a consensus of US$1.62/share within a range of US$1.17–2.10/share (source: Refinitiv, 30 July 2021). In this case, our estimate is, once again, predicated on an average (nominal) gold price during the year of US$1,892/oz and an average silver price of US$30.78/oz, which assumes, among other things, the silver price will revert to the long-term correlation that it has exhibited with gold since the latter was demonetised in 1971. If both metals instead remain at current levels, however (US$25.54/oz Ag and US$1,810/oz gold (Au) at the time of writing), our forecast for WPM’s EPS in FY22 would then moderate to US$1.75 per share and our forecast for its DPS to US$0.71/share.

At the time of its Q420/FY20 results, WPM provided production guidance of 720–780koz gold equivalent (AuE) for FY21 as well as five-year production guidance of 810,000oz AuE per annum and maiden 10-year guidance of 830,000oz AuE per annum. This compares with Edison’s updated forecasts as follows:

WPM’s updated five-year guidance and its 10-year guidance are now based on standardised pricing assumptions of US$1,800/oz Au, US$25.00/oz Ag, US$2,300/oz palladium (Pd) and US$17.75/lb cobalt (Co). Of note in this context is an implied gold/silver ratio of 72.0x, which is actually very close to the current ratio of 70.9x, but slightly at odds with the long-term average of 61.5x (as implied by silver’s correlation with gold since the latter was demonetised in 1971).

Readers will note that Edison’s FY21 silver production forecast remains just above the top end of WPM’s guidance range. After producing 6.8Moz Ag in Q121 however, WPM’s mines will only be required to produce at a rate of 5.7Moz Ag per quarter for the remainder of the year in order to achieve the top end of WPM’s guidance range of 22.5–24.0Moz Ag for FY21. This compares with a long-term average quarterly production rate of 6.6Moz per quarter since Q112. Conversely, Edison’s gold production forecast remains slightly below the bottom end of WPM’s guidance range. After producing 77.7koz Au in Q121, WPM’s mines would have to produce at a rate of 97.4koz Au per quarter for the remainder of the year in order to achieve the bottom end of WPM’s guidance range of 370–400koz Au for FY21. While this is certainly possible (WPM’s gold mines produced at an average rate of 102.4koz per quarter in the period Q318–Q419), we think that it may prove demanding, given recent developments at Salobo and Sudbury in particular. In this respect, Edison’s financial forecasts for FY21 may prove conservative. Self-evidently, at the standardised prices indicated, Edison’s gold equivalent production forecast of 737koz AuE remains within the range of Wheaton’s guidance of 720–780koz, albeit fractionally towards the bottom end of the range.

Otherwise, readers will note that Edison’s (updated) production forecasts for FY22–25 and for FY26–30 remain, to all intents and purposes, unchanged and slightly conservative when compared with Wheaton’s guidance. However, this difference is largely negated in its translation into financial results by the fact that sales have historically been recorded at a level (on average) c 8.5% below production (at prevailing prices) whereas Edison’s financial forecasts typically assume that the two are closely aligned. Moreover, our forecasts exclude (for the moment) any contribution from Santo Domingo (see our note Q121 results preview, published on 31 March 2021), which we expect to be in the order of 42,350oz Au per annum on average in the period FY24–25 and 25,520oz Au per annum on average in the period FY26–30.

WPM provided guidance for non-stock general and administrative (G&A) expenses of US$42–45m (or US$10.5–11.25m per quarter) in FY21, compared to a guided range of US$40–43m in FY20 and an actual outcome of US$38.7m (ie 3.1% below the bottom of the range), including all employee-related expenses, charitable contributions, etc, but excluding performance share units (PSUs) and equity settled stock-based compensation. In the event, non-stock G&A expenses in Q121 were below the pro rata quarterly rate implied by WPM’s full-year guidance, while total G&A costs were well below Edison’s forecast for the quarter on account of relatively small PSU accrual and equity settled stock-based compensation expenses:

Compared with non-stock G&A expenses, total G&A expenses are relatively difficult to forecast, given their dependence on the price of WPM’s shares. However, a simple analysis of stock-based G&A expenses over the past six quarters against WPM’s share price (also in US dollars) exhibits a relatively close Pearson product moment (correlation) coefficient between the two of 0.79, which is just statistically significant at the 5% level for a directional hypothesis (ie there is less than a 5% probability that this relationship occurred by random chance). The graph relating to the analysis is shown below, including its linear trendline:

From end-March to end-June 2021, WPM’s share price has risen by US$5.86/share. On the basis of the above analysis, we would predict that such a move would give rise to a stock-based G&A expense of US$7,460k to add to guidance for non-stock G&A expenses of US$10,875k to give a total of US$18,335k for the quarter. This is almost exactly the same as our forecast figure of US$18,329k in Exhibit 3, which we are thus leaving unchanged.

Excluding FY04 (part-year), WPM’s shares have historically traded on an average P/E multiple of 30.0x current year basic underlying EPS, excluding impairments (cf 30.9x Edison or 30.5x Refinitiv consensus FY21e, currently – see Exhibit 9).

Applying this 30.0x multiple to our EPS forecast of US$2.11 in FY23 (previously US$2.07) would ordinarily imply a potential value per share for WPM of US$63.28 in FY23 or C$79.38 in that year. However, the graph above suggests that the current year multiple has been on a broadly upward trend between FY12 and FY19, on which basis we would argue that a multiple in excess of 40x (as evidenced by FY18 and FY19) could be supported, not least given the fact that these years were not subject to the extraordinary tribulations experienced in FY20. In this case, applying a 40.7x earnings multiple (the average of FY18, FY19 and FY20) to our updated EPS forecast of US$2.11 in FY23 implies a potential value per share for WPM of US$85.82 in FY23 or C$108.44 in that year and/or for as long as precious metals prices remain strongly supported and/or the current coronavirus crisis persists. Even at such share price levels, however, a multiple of over 40.7x would still leave WPM’s shares at a 4.9% valuation discount compared to Franco-Nevada (see Exhibit 9).

In the meantime, from a relative perspective, it is notable that WPM has a lower valuation than the average of its royalty/streaming ‘peers’ on all of nine valuation measures in Exhibit 9 if Edison forecasts are used or seven out of nine measures if consensus forecasts are used. On an individual basis, it is cheaper than its peers on 83% (30 out of 36) of the valuation measures used in Exhibit 9 if our estimates are adopted or 72% (26 out of 36) of the same valuation measures if consensus forecasts are adopted. Among other things, this could possibly indicate the market has more conservative precious metal pricing expectations than we do (especially in FY22 and FY23).

As at 31 March, WPM had US$191.2m in cash (cf US$192.7m in Q420) and no debt outstanding (cf US$195.0m in Q420) under its US$2bn revolving credit facility (which attracts an interest rate of Libor plus 120–220bp and now matures in February 2025), such that (including a modest US$3.5m in leases) it had US$187.7m in net cash overall (cf US$6.0m in Q420) after US$232.2m of cash generated by operating activities during the quarter (cf US$208.0m in Q420). All other things being equal, we estimate that WPM will have US$257.8m on its balance sheet at the end of Q221 after net cash inflows of US$191.5m during the quarter and after the payment of US$121.4m in combined dividends for Q121 and Q221.