Aug 9, 2019
 in 
Tech

Time to Rethink the Coasean Legacy

W

e owe much of our telecommunications progress to Ronald Coase, a renowned British economist who moved to the United States in the early 1950s, and made an important contribution to the Federal Communications Commission (FCC) in 1959.

On the subject of Radio Spectrum allocation, Coase saw that regulators were assigning spectrum based on usage restrictions, and proposed to use a price system instead. He was mocked by his counterparts, who argued this would lead to chaotic conflicts of radio waves saturation.

Yet that was exactly what the FCC ended up doing in 1994, when the understanding of spectrum changed and the commercial opportunities found solace in both market and normative regulation. As a result, more than 31,000 licenses were sold in the United States and more than $32 billion were paid to the US Treasury.

The fundamental insight was that analytical symmetry was demanded, accounting for the net benefits of both regulation and markets. Coase augured that the price system would outperform administrative allocation, a conclusion mocked by communications policy experts. Source: JSTOR

Services and innovation proliferated, and the world followed that model. But it was something that worked great in a time when telephony ruled and AM/FM airwaves were clearly structured, not presenting any exponential growth or new uses.

The proposition of bidding competitively for spectrum licenses was an efficient one when market maturity was a given.

It meant that companies could obtain the license, make the necessary investments to install and operate a network, and then have somewhat predictable and stable returns to recoup the amount invested.

This has worked well until recently. Technology behaves in different cycles now, as we have seen with the semiconductor industry, reflected in the smartphone and personal computer segments. In one period you have incremental upgrades, and at a subsequent one, dramatic, disruptive shifts.

Intel's Tick-Tock model of innovation. Source: Intel

It is called the tick-tock model, and it was coined by Intel in 2007. In it, the "tick" part of the clock analogy would represent a small swing, in the form of just making the processors smaller. While the "tock" would mean providing a whole new architecture, with substantial speed and reliability changes.

In a similar fashion, the spectrum used for frequencies that provided us with 3G connectivity could be considered a "tock" cycle. It was certainly perceived to be so by most regulators around the world, which priced licenses at a significant premium and elicited a bidding euphoria in most relevant markets.

The result was 77 licenses issued by the auction method (the vast majority of 3G licenses in the world), representing $101 Billion US Dollars, with most of the licenses being active by 2002. While this was great for the issuing countries, it left network operators with a lot less resources to deploy, expand and improve their service offerings.

We can observe this by noting that 15 years later, the world features a similar amount of 3G networks as it does with 4G. Partly because everyone curbed their expectations, and partly because 4G could be considered a "tick" cycle, providing ROI efficiency and better speeds but largely the same use cases.

With most of the 4G licenses issued in 2009, it took 8 years to attain 29% of global deployments, market maturity and cheaper hardware for consumers and operators played a part, but it was aided by reasonably priced spectrum auctions.

The 2017 global deployments from 2G to 4G vs the projected numbers for 2025 including 5G. Source: GSMA Intelligence.

But now let us go fast forward into 2019. The excitement and market euphoria for 5G, perceived as the ultimate "tock" cycle, promises to bring just about everything from driverless cars to remote, AI-assisted surgery.

This is why staying in the Coasean legacy of auctioning for spectrum can be dangerous, because it will take time to make the investment pay off. There are more forces at play. An auction now can make way for a bidding war that makes conditions far worse for everyone.

Regardless of the price of spectrum, there is hardware that has yet to be invented, made accessible and later diffused enough for 5G to hit its stride. Making 5G spectrum expensive would be disastrous for global rollouts and the price of that connectivity.

Indulging in the temptation to generate cash with the spectrum can stifle innovation for as much as a decade, at a time when network operators are integrating vertically and engaging in many efforts to provide aggregated services in connection to connectivity.

It is less obvious, but governments should be able to see that they can transfer the excitement away from spectrum bidding and into the taxable profits generated by 5G-enabled services. Make the spectrum affordable, and reduce the time-to-market for new solutions.

Maybe capping the bidding amounts? Giving it away under certain investment conditions, where a network that does not get substantial and continued investment would make the operator lose the 5G license?

There is already a need for change. This last June in Berlin, the spectrum auction for 5G was closed raking in $6.5 billion Euros from 4 companies, which doubled the initial estimate from regulators. The bidding war lasted 3 months and about 200 rounds.

One-fourth of the spectrum is dedicated to industrial uses to encourage new manufacturing methods. It remains to be seen how much this will really cost the companies once they get into the execution phase.

-Mariano Malisani