White Paper Questions FCC’s Quantile Regression Analysis

"The Balhoff William’s Study filed today confirms NARUC’s original concerns,” said NARUC Telecom Committee Chair John Burke in a statement. “However the QRA is used, it is clear, in its current form, it does contain, what Ballhoff [sic] William’s characterizes as ‘major problems.’ At a minimum, the FCC should expeditiously incorporate the improvements suggested in section three of the white paper.” NARUC has asked the FCC to suspend the model until it works out the problems the state regulators see.

"The Connect America Fund is already working to expand broadband access to the 19 million rural Americans who lack it, while at the same time increasing fiscal responsibility in order to be fair to the consumers and small businesses nationwide who pay into the fund through universal service fees,” an FCC spokesman said in a statement when asked about the paper. “The FCC continues to work closely with all stakeholders as we make this once-in-a-generation transition to bring broadband and voice service to all of rural America."

Costs are improperly defined and 14 of the 16 benchmarks used in the analysis have “meaningful problems,” the paper’s authors said. There are also “poor procedures related to statistical data analysis” and conceptual issues creating problems, they said. A team of people has studied the model since last June: “The plan was to recreate the model, then test whether it generated results similar to those reported by the FCC, and, finally, disaggregate the variables and the data to understand the critical drivers of the QRA,” the authors said. “It was hoped that such a process would permit all stakeholders to gain visibility into the model in a way better than what was previously available.” The Wireline Bureau released last June two weeks before its implementation a description of how the model worked, said Alexicon’s Wiemer in a letter to the FCC with the paper (http://bit.ly/15yKfTx). He said that release prompted the work responsible for this paper.

The model can and should be revised, the paper recommends, suggesting the FCC end automatic cuts to funding and instead trigger reviews. “Specifically, improved data sources and variables should be identified and used; the opex and capex calculations should be combined into a single benchmark; and costs that cannot be controlled by the carrier -- because of exogenous factors that vary widely from one jurisdiction to the next including tax and depreciation rates set by governmental authorities -- should not be used to ‘identify’ excessive and unrecoverable expenses,” it said. It also recommended an advisory committee.

The model automatically disallows about $94 million in funding, the paper said: “The issue is that presumptive disallowance based on an approximate calculation process is likely to lead to harmful outcomes without ongoing consideration of whether the model has found or failed to find excesses.” It’s a “small portion of the fund but could have a significant effect on individual small carriers where the costs may actually be appropriate,” it said.

The model has attracted critics throughout the states and most recently by an Obama cabinet member (CD Feb 20 p3). “Earlier this month, Agriculture Department Secretary [Tom] Vilsack met with Chairman [Julius] Genachowski to share recommendations on how to keep broadband deployment sustainable in rural areas,” said Commissioner Jim Cawley of the Pennsylvania Public Utility Commission in a statement: “One recommendation that was not a surprise was his request for an adjustment in regression analysis-based support caps from the federal Universal Service Fund mechanism. The Alexicon Consulting/Balhoff & Williams paper only validates all the QRA criticisms the agency has received to date.” He said he hopes rural carriers and those lending them money have more certainty going forward. “I hope this time the FCC will pay careful attention to this report and its recommendations,” he said.