Spectrum has long been a sector curse. History is littered with examples of auctions that have either cost substantially more than expected, or facilitated new entrant competition. Recent newsflow only reinforces these themes. The US C-band auction has soared to $96bn (3x consensus), and the Portuguese auction will foster a new challenger. With our analysis identifying 28 additional global auctions on the horizon, and €52bn in spend, investors would be forgiven for questioning our contrarian 2021 sector OW. But we stay the course. Note: (1) 2021-22E global spend at €134bn (€30bn ex US), translates to “just” €33bn (€14bn ex US) for the Euro incumbents; (2) Helpfully, several EM regulators have altered policy objectives away from maximizing auction proceeds; (3) In Europe we expect auctions in Spain and the Nordics to prove benign, with only the UK competitive; (4) At a stock level, we alter 2020-22E spectrum spend assumptions for DT (+€10bn to €22.5bn due to the US), Telenor (+NOK8bn to NOK17.5bn due to DTAC), Vodafone (-€0.8bn to €4.7bn due to SA/Turkey), and Telefonica (-€0.7bn to €3.0bn due to Latam, though offset by higher capex).

 US C-band (3.5GHz): Auction results are expected imminently with spend set to hit $96bn (3x consensus). We believe the “over-spend” is market specific, and should not be extrapolated. We have long argued consensus incorrectly forecasts US auctions with half an eye to European Price/Mhzpop benchmarks, failing to adequately adjust for differences in license duration (US perpetual vs Europe 15 years) or industry profitability (US sector EV is 4x Europe). For DT, C-band pricing serves to underscore the value of the 150Mhz of 2.5Ghz spectrum it gained through the $59bn Sprint acquisition (worth $31bn even if we assume a 40% discount to C-band).

 Portuguese new entrant: 563MHz of spectrum is being sold across various bands. Frustratingly 40MHz is reserved for a confirmed new entrant (at odds with the depressed nature of industry returns). Whilst disappointing, note attempts to promote a new entrant in Czech failed, and in Germany states are proposing free licenses in return for coverage commitments.

 €134bn in spend across 31 auctions: Our detailed bottom up, country by country modelling, leads us to forecast €134bn in 2021-22E spectrum spend across 31 global auctions. 78% (€104bn) of this is attributable to the US. By contrast European spend at €7bn (2021E) is only in-line with the 2019 level, and even this is only due to auction delayed from 2020E due to COVID-19.

 Stock implications: We forecast €33bn of spectrum spend across the Euro Telco incumbents. We alter 2020-22E assumptions for DT (+€10bn to €23.5bn due to 2022E US auctions), Telenor (+NOK8bn to NOK17.5bn due to Thailand), Vodafone (-€0.8bn to €4.7bn due to more benign expectations across SA/Turkey) and Telefonica (-€0.7bn to €3.0bn due to Brazil where cash payments will be small, and the rest paid in capex commitments).