Evolution of prices for mobile spectrum & possible explanations

The adoption of 4G and 5G technologies has led to a remarkable surge in mobile data consumption, prompting regulators worldwide to make substantial amounts of new spectrum available to mobile operators through spectrum auctions. This paper aims to examine the trends in spectrum pricing and to address the question of how much spectrum mobile networks will need in the long term. Spectrum auction prices should reveal the market value of spectrum which will provide an indication of its scarcity. Therefore, to address this question, the paper utilizes a global dataset of more than 400 spectrum auction prices from 2007 to 2023 to demonstrate that spectrum prices have generally declined over the past two decades. This decline suggests that while mobile data traffic has expanded, the release of new spectrum has generally kept pace and helped avoid capacity issues on mobile networks. In Section 2, we1 examine the trends for different types of spectrum, including low band, lower mid-band, and upper mid-band spectrum, which serve different purposes in mobile networks. The results indicate an overall decline in spectrum prices, with the steepest decline observed in low-band spectrum. This suggests that incremental coverage spectrum is no longer essential, as successive releases of this spectrum have adequately met market demands. Lower mid-band spectrum prices have also decreased, while prices for upper mid-band capacity bands have remained relatively stable, indicating their continued importance in providing capacity. In Section 3, we turn to mobile data consumption. Industry statistics generally show an accelerating growth rate in the early days of 4G, but a decline in the later 5G era. A possible implication is that data consumption is following an S-shaped curve typical for technological innovation. We fit a generalized logistics function to OECD data using least squares and then use it to forecast mobile data demand in the future. These forecasts are then compared to other projections which assume exponential growth. We find that the S-curve provides a more plausible path for mobile data growth, based on historical data, implying that exponential forecasts significantly overestimate future data consumption. Higher growth forecasts may only be justified if there is a significant future change in how we use phones, such as widespread adoption of data-intensive extended reality (XR)2 technology, This finding has significant implications for regulators planning future spectrum releases. Even with slowing data growth, the absolute expansion in data traffic carried by mobile networks over the next ten years will be substantial. Nevertheless, it may be possible to accommodate this with relatively modest increases in spectrum allocation for mobile, as opposed to the huge increases that would be required to support exponential growth in mobile traffic.