Puerto Rico’s EV Adoption Trails Planning Targets by Up to 19 Years, Model Shows

A new study of Puerto Rico’s electric-vehicle transition finds a substantial gap between the electrification pathways assumed in long-term planning documents and the adoption trajectory implied by the island’s own vehicle-registration records. Using a generalized-logistic diffusion model calibrated on registered EV-stock data from 2014 through 2023, researchers compared three planning-reference electrification envelopes against the calibrated trajectory and quantified the differences with two metrics: an adoption-level discrepancy, ΔP(t), measuring gaps in electrification level at given years, and a transition-time discrepancy, ΔT_x, measuring delays in reaching specific electrification milestones.

The calibrated model reproduced the observed registered EV-stock series closely, with a coefficient of determination of 0.996 on cumulative stock and 0.959 on annual increments. Yet when projected forward under a common planning endpoint of full electrification of the reference vehicle fleet by 2050, the calibrated trajectory remained below all three planning-reference envelopes—front-loaded, linear, and delayed—at every representative planning year evaluated. Relative to the 2050 benchmark, the adoption-level discrepancy reached approximately 25 percentage points, with a 95 percent interval of 22 to 30 percentage points obtained by propagating parameter uncertainty.

The timing gaps were often more striking than the level gaps. Representative electrification milestones were attained approximately 2 to 19 years later under the calibrated trajectory than under the corresponding envelopes, with delays of roughly 11 to 16 years at the 90 percent electrification milestone. Under the 100 percent reference endpoint, the base-case discrepancies for the three primary envelopes remained positive across all bootstrap replicates evaluated, indicating that the direction of the gap was robust to parameter uncertainty within the chosen model.

The study’s authors are careful to frame what the numbers do and do not mean. The calibrated trajectory is a conditional, model-based reference derived from registration records, not a deterministic forecast of future adoption. Likewise, the planning envelopes are transparent reference pathways constructed for consistency assessment within Puerto Rico’s policy and decarbonization context; they are not legally mandated fleet targets or predictions of future fleet composition. Puerto Rico’s legislative and regulatory instruments—including Law 81-2014 promoting electric-vehicle use, Laws 17-2019 and 33-2019 on energy and climate policy, the federal NEVI charging-corridor program, and the PR100 grid study—support transportation electrification through mid-century but prescribe no unique temporal pathway for fleet electrification.

The calibration relied on registered stock of electrified light-duty vehicles from the Puerto Rico Department of Transportation and Public Works, in collaboration with the Department of Treasury’s Excise Tax Bureau. The saturation parameter was fixed at 1,985,222 vehicles, the DTOP-reported light-duty fleet total, under the hypothetical assumption that every vehicle in that fleet is ultimately replaced by an electric one. Because the calibration window covers only the early diffusion stage—with registered EV stock below 1 percent of the saturation reference—the long-horizon extrapolation carries substantial model-form uncertainty that the reported statistical intervals do not capture.

Sensitivity analyses revealed which assumptions drive the reported discrepancy. Varying the 2050 reference endpoint across 70, 80, 90, and 100 percent showed that the 2050 discrepancy vanishes at a break-even endpoint of approximately 74.8 percent and reverses sign below it. For the delayed-transition envelope, the calibrated trajectory actually exceeds the planning pathway at 2040 once the endpoint falls below roughly 81.9 percent. Measured one dimension at a time, envelope shape dominated the discrepancy at 2040, contributing roughly 33 percentage points of variation, followed by the endpoint at roughly 22 points, while the saturation reference contributed less than one point.

A continuous sweep of the power-law envelope exponent confirmed that the main-analysis conclusions were not artifacts of the three chosen shapes. The directional result held for exponents up to a critical value between roughly 2.6 and 3.0 depending on the evaluation year, well above the most delayed envelope exponent of 2 used in the main analysis. A sigmoid-family robustness check with independently specified parameters likewise left the discrepancy intact, with milestone delays of approximately 7 to 19 years in the reference case. The saturation reference, by contrast, had a comparatively small effect on the normalized metrics—a contracting-fleet sensitivity changed ΔP(2050) from 25.2 to 24.6 percentage points—but materially affected absolute charging-energy estimates.

The planning implications follow directly from the adoption-level discrepancies. Using a plug-in-equivalent scaling factor of 15 kilowatt-hours per vehicle per day, the difference between the linear envelope and the calibrated trajectory corresponds to roughly 7.5 to 10.6 gigawatt-hours per day of aggregate charging energy across the representative planning years, with scaling sensitivities spanning about 5.0 to 14.1 gigawatt-hours per day depending on year and factor. The authors stress these are order-of-magnitude screening figures, not charging-profile estimates, and that feeder loading, hosting capacity, and peak impacts require spatially and temporally resolved studies beyond the paper’s scope.

The authors also flag a definitional nuance in the data: the registered series aggregates vehicles in electrified model lines, including roughly 38 percent battery-electric and 30 percent plug-in hybrid vehicles in the 2024 cross-section, alongside non-plug-in variants. Narrower drivetrain definitions increased the discrepancy—raising ΔP(2050) to 31.4 percentage points under a plug-in-capable definition and 40.9 points under a battery-electric-only definition—so the aggregate definition was retained as the conservative choice.

Ultimately, the framework is offered as a reproducible screening step for planners: a way to check whether the electrification trajectories fed into charging-infrastructure and power-system models remain aligned with observed diffusion evidence, and to reassess that alignment periodically as registration data accumulate. The authors emphasize that the findings do not establish that long-term electrification objectives are unattainable; rather, they demonstrate the practical importance of distinguishing a long-term planning endpoint from the temporal pathway assumed to reach it—a distinction with direct consequences for load forecasting, infrastructure-deployment schedules, and network-investment timing in islanded power systems like Puerto Rico’s.

Subject of Research: Consistency assessment of EV electrification planning pathways against a calibrated diffusion trajectory in Puerto Rico

Article Title: Assessing planning-reference electrification envelopes against a calibrated EV-diffusion trajectory in Puerto Rico

Article References: Feo-Cediel, G. A., Luna, A. C., Rivera, A. I., Andrade, F., & Campo-Ossa, D. D. (2026). Assessing planning-reference electrification envelopes against a calibrated EV-diffusion trajectory in Puerto Rico. Discover Vehicles, 2(1), Article 33. https://doi.org/10.1007/s44465-026-00039-8

Image Credits: AI Generated

DOI: 10.1007/s44465-026-00039-8

Keywords: electric vehicles, Puerto Rico, EV diffusion model, generalized logistic, transportation electrification, power-system planning, charging infrastructure, adoption forecasting, planning envelopes, milestone delays, parameter uncertainty, decarbonization