Risk Assessment in Spectrum Policy

Abstract

“Everybody wants more radio,” conference organizer Pierre de Vries told participants by way of introduction to the Risk Assessment in Spectrum Policy conference hosted by the Silicon Flatirons Center on October 23, 2015 at the University of Colorado Law School. Spectrum access is hotly contested, De Vries explained, since there are practical limits on the number of radio devices that can operate concurrently in close proximity in time, place and frequency. The main question confronting a spectrum regulator, therefore, is whether to allow new wireless entrants to operate in frequency bands with incumbents. To answer that question the regulator must balance the potential gains to be realized from new services against possible harm that those services would cause to incumbents.Interference analysis is important in determining the costs and benefits of allowing new entrants into the spectrum field. However, spectrum engineering risk analyses have traditionally been deterministic, which is to say that they calculate interference by using single values for all the relevant parameters, often with many or all being extreme values – a worst case analysis. “It tends to be conservative as a method,” De Vries told the audience. “It easily leads to rules that provide incumbents with more protection than they need, while at the same time not allowing the new services to realize their full value.”Worst-case analysis made sense when spectrum rights were not in such great demand, De Vries explained; but as wireless services are packed ever more tightly together, over-conservative protection rules will not maximize the economic and societal value of the spectrum use. For example, worst-case analysis may lead to the establishment of unnecessarily wide guard bands for a particular incumbent service.Risk-informed analysis – also known as Quantitative Risk Assessment (QRA) or Probabilistic Risk Assessment (PRA) – is gaining traction in the spectrum community as a complement to the deterministic, often worst-case, analysis described above. Probabilistic risk assessment takes a range of values as inputs, rather than single worst-case values, and attempts to give a more complete picture of risk by addressing three questions, the “risk triplet”: 1. What can happen? (That is, what can go wrong?); 2. How likely is it to happen?; 3. What are the consequences of it happening?The Silicon Flatirons Conference on Risk Assessment in Spectrum Policy gathered experts in the currently distinct fields of risk assessment and spectrum policy to discuss how engineering risk analysis is being applied to public policy, how risk assessments inform government regulation generally, and what the major challenges and opportunities are for applying quantitative risk assessment to spectrum regulation specifically.This conference report is organized thematically. Each of the four sections that follows identifies a major topic of discussion from the conference, summarizes the views expressed on each topic, and records selected quotes from panelists.