Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Forecasting topic
No spam. Unsubscribe anytime.
Expert witness says Young Brothers forecasting omits price elasticities, cautions limits of models
Summary
At a PUC evidentiary hearing in Docket No. 2024-0255, economist Dr. Paul Brubaker testified that his cargo-volume forecasting uses a fixed-price, macro approach and does not incorporate price elasticities; he said data limits and structural changes since the 2010s make isolating price effects difficult.
Dr. Paul Brubaker, an economic forecasting witness who testified in Docket No. 2024-0255 at a Public Utilities Commission evidentiary hearing, told commissioners that his volume forecast model is a macro, fixed‑price approach and does not include estimated price‑elasticity effects. The testimony underscored gaps witnesses say regulators and parties must consider when weighing Young Brothers LLC’s rate proposals.
Brubaker said his model relies on aggregated “real” variables such as cargo tonnage, visitor counts and building permits, and that it has produced broadly consistent fits for historical tonnage. But he told the commission he has not modeled correlations between prices and cargo volumes and said he has never had access to firm‑level prices that would be needed to estimate customer price responses. “You can’t model pricing responses without the prices, and I’ve never seen the prices. To this day, I’ve never seen the prices,” Brubaker said on cross‑examination.
Why this matters: Regulators and intervenors use volume forecasts to estimate revenue and the effect of rate changes. If a forecast does not explicitly measure how customers respond to price changes, parties warn it can miss a channel that could reduce or shift projected volumes when rates rise.
Key testimony and context - Model type: Brubaker described his approach as a macro, fixed‑price model that operates on real (inflation‑adjusted) series rather than observed tariff or transaction prices. He said the approach was historically successful but that the pattern began breaking down in the mid‑2010s and again after the pandemic. - Data limitations: Brubaker noted annual cargo‑tonnage series are available back to 1996 and higher‑frequency (quarterly or monthly) data only since 02/2005, which constrains some analyses. He also testified that he has previously used published economic forecasts (UHERO, DBEDT) as inputs to judgmental components of his projections. - Price elasticities and micro vs. macro approaches: Asked whether he had searched his series to detect price‑elasticity effects, Brubaker answered that he had not and that separating price responses from other structural and macroeconomic effects would be difficult. He said he had collaborated with QSI (a volume‑forecasting team) and others and would “try” to merge bottom‑up (micro) and top‑down (macro) methods but expressed uncertainty about whether that integration could reliably separate price effects from non‑price influences.
Selected quotes - “I think I could try. I’m not sure I could actually do it,” Brubaker said about attempting to merge micro‑level price response analysis with his macro model. (Cross‑examination) - “My model is a fixed price model,” he said, describing his specification and how nominal variables are deflated to real terms for analysis.
What Brubaker said about the recent period Brubaker told the commission the forecasting model he has used performed well through the early 2010s but that the amplitude of drivers such as the investment cycle and tourism variation diminished thereafter, and that post‑pandemic patterns have not reverted to previous paths. He described the post‑pandemic period as one in which trend behavior dominated, reducing cross‑sectional variation that forecasts normally exploit.
Limitations and implications Brubaker repeatedly cautioned that, without access to price data, it is hard to assess whether and to what extent cargo volumes would respond to rate changes. He also warned that micro‑level heterogeneity (many different customers with differing behavior) complicates attempts to aggregate a price‑response estimate to the system level. He told the panel that he and the QSI team had discussed integrating flexible price considerations into the forecasting framework but did not claim that such integration had been completed.
Ending Brubaker’s testimony framed a technical limit in how volume risk is currently modeled in the proceeding: the absence of price data and the difficulty of isolating price effects from structural and macroeconomic changes. Commissioners and parties pressed him about what it would take to add price responsiveness to forecasts and whether doing so would materially change projected volumes, but Brubaker said those tasks would be difficult and uncertain without transaction‑level price series.
