Allocation Timing in Satellite
Market
Allocation Timing in Satellite
Market
Authors
Federico Bobbio (Northwestern University), Michael Honig (Northwestern University), Randall Berry (Northwestern University), Rakesh Vohra (University of Pennsylvania), Thanh Nguyen (Purdue University), Vijay Subramanian (University of Michigan)
Abstract
The FCC allocates spectrum for non-geostationary satellite broadband through sequential “processing rounds.” Operators authorized in earlier rounds such as SpaceX’s Starlink enjoy regulatory protection from later entrants such as Amazon’s Kuiper who must demonstrate that their systems will not cause harmful interference to incumbents. This inter-round protection has evolved from an undefined case-by-case standard (2017), through a codified regime with a ten-year sunset clause (2023), to quantified throughput-degradation criteria (2024). The protection is temporary: after sunset, all operators share spectrum on equal footing under the FCC’s default 1/n band-splitting rule. We abstract this institutional structure into a two-period Cournot model with linear demand, quadratic launch costs, and reciprocal interference. A batch mechanism has both firms deploy simultaneously; a sequential mechanism gives the incumbent a first-mover advantage and imposes an asymmetric regulatory cost on the entrant. The inter-round protection maps to a single parameter œà that penalizes the entrant’s deployment, and the sunset corresponds to œà dropping to zero after a fixed horizon. We establish five sets of results. First, a complete threshold hierarchy: the entrant’s indifference point is nested inside the total-profit roots, but profit and consumer-surplus favorable regions do not nest opposite curvatures make the welfare ranking parameter-dependent, with three sub-cases governing when consumer surplus and welfare agree or disagree on the preferred mechanism. Second, the incumbent’s Stackelberg optimum always exceeds the batch quantity, so the equilibrium always falls in the region where the entrant is strictly worse off; the “aligned preferences” region where both firms gain exists in theory but is never reached at the equilibrium. Third, when the incumbent can expand its constellation via license modifications, the entrant is worse off under sequential allocation for every level of incumbent deployment, and the aligned-preferences region vanishes entirely the key mechanism is strategic substitution between early and late deployment. Fourth, introducing a second frequency band with shared demand restores what expansion destroys: the entrant uses the new band as a regulatory bypass, both indifference thresholds shift outward, and uniquely among all model variants the equilibrium can fall in the aligned-preferences region, providing the strongest theoretical case for the FCC releasing new bandwidth alongside sequential allocation. Fifth, a saturation theorem: the welfare gap between simultaneous and sequential band release is strictly increasing in the incumbent’s installed base, so the FCC should release new bandwidth early rather than waiting for existing bands to saturate result consistent with the FCC’s 2017 decision to release V-band just one year after the Ku/Ka round.