全球金属与矿业：贵还是便宜？百年矿业周期估值信号分析

Global Metals & Mining: Cheap or expensive? What 100 years ofmining CAPE signals (Metals&Mining:Mining versus theMarket..Dimensioning theexpectedoutperformance)viaShillerCAPE(Cyclically AdjustedP/E).Sixplusyearslater westill havea fewclients askingus to update the analysis. In this note we return to the analysis. Specifically,we combine thisanalysiswithourreversionmasteranalysis. +19173448422bob.brackett@bernsteinsg.com +44 20 76766825andrianto.guntoro@bernsteinsg.com the cycle (Global Metals & Mining: Our reversion master implies things are as good asthey get),and hence the attractiveness of the sector. In essence, long-term returns isinversely correlated to sector EBITDAmargins (Exhibit 1).Filtering our Reversion MasterframeworkthroughaCAPElensyieldstwokeyobservations. First,for strategic/long-term holding period.The best setup has historically been whenminingCAPEtradesbelowtheS&P5O0CAPEand industrymargins arebelowmid-cyclelevels. In these periods, 10-year forward returns have been heavily skewed to the upside(Exhibit 7).Since December1994,142of 259monthlyobservations (55%of the sample)fell into this category.Of those 142 observations, 135 delivered positive 10-year forwardreturns, with an average CAGR of 9.4%. Second,for short(er)-term holding period.Whenmining CAPE trades above the S&P500 CAPE and industry margins are above mid-cyclelevels,1-year forward returns tend tobe skewedpositively (Exhibit9).Since December 1994,83 of367monthly observations(23%of the sample)metthese criteria.Of those, 59 observations (71%)generatedpositivereturns,averaging33%.The remaining 29%producedaverage losses of 38%,reducingthe overall expected return to c.13% Beyondthesetwoscenarios,othercombinations of CAPEand industrymarginsdonotshowaparticularlystrong or consistent returnprofile. Today,sector EBITDA margins stand at 49.8%, well above the LT average of 33.2%.Meanwhile, mining CAPE trades at 17x versus 41x for the S&P 500, implying a relativeCAPE of o.41x. Historically, when relative CAPEis below 1x but industry margins remainabovelong-term averages, the probability of a positivelong-term return is high (77%),but the payoff is limited, with average 10-year returns of just c.2% CAGR (Exhibit 12)The short-term picture is only marginally better: Exhibit 15 shows a 53/47 split betweenpositive and negative outcomes, with average 1-year returns of c.6%. Hence,in current environment, we believe commodityand stock selection will playan increasingly important role ingenerating alpha as sector-level return potentialratings on Barrick and Newmont, supported by our constructiveview on gold,and onRio Tinto, which we view as a high-quality, defensive mining exposure. RiO's c.3% CuEqproduction CAGR through2030 should helpto lower unit cost and expand EBITDA margin. INVESTMENTIMPLICATIONS Wemaintain our Outperform ratings on Rio Tinto, Barrick and Newmont.We maintain our Market-Perform ratings on AngloAmerican,Antofagasta,BHP,Boliden,Freeport,Glencore,andVale get)to determine where we are in the cycle,and hence the attractiveness ofthe sector.In essence,long-termreturns isinversely correlated to sector EBITDA margins (Exhibit 1).The model is available to clients -please let us (or your favourite salespeople) know if you'd like a regular update. Inthis note,we overlay our ReversionMasterframework withCAPE (CyclicallyAdjusted PE)analysis,which wasarrive at relative CAPE (Exhibit 4).Mining CAPE has been between O.2x and O.5x of SPX CAPE in the past decade (Exhibit 5).Wediscuss the CAPEmethodology in the next section. decade? In the 2oo0s, demand shock (hint: China) caused mining to re-rated.Supply shock next? First,for strategic/long-term holdingperiod.Thebest setup has historically been whenmining CAPEtradesbelowthe S&P 500 CAPE AND industry margins are below mid-cycle levels. In these periods, 10-year forward returns have been heavilyskewedtotheupside(Exhibit7).SinceDecember1994,142of259monthlyobservations(55%ofthesample)fellintothiscategory.Of those 142 observations, 135 delivered positive 10-year forward returns, with an average CAGR of 9.4%. are above mid-cycle levels, 1-year forward returns tend to be skewed positively (Exhibit 9). Since December 1994, 83 of 367monthlyobservations (23% of the sample)metthese criteria. Of those, 59 observations (71%) generated positive returns,averaging 33%.The remaining29%produced averagelosses of38%,reducingtheoverall expected return toc.13%. consistent return profile. at 17x,below the S&P500's 41x, implying a relative CAPEof 0.41x.Historically, when relative CAPE is below 1x butlimited, with average 10-year returns of just c.2% CAGR (Exhibit 12). The short-term picture is only marginally better: Exhibit 16shows a 53/47split between positive and negative outcomes, with average 1-year returns of c.6%. EXHIBIT 6: The best time to gain/overweight exposure to the mining sectoris when industry margins are belowlong-term averages and mining valuations trade at a discount tothebroadermar