Low-carbon hydrogen could be an important component of a net-zero carbon economy, helping to mitigate emissions in a number of hard-to-abate sectors. The United States recently introduced an escalating production tax credit (PTC) to incentivize production of hydrogen meeting increasingly stringent embodied emissions thresholds. Hydrogen produced via electrolysis can qualify for the full subsidy if the input electricity is generated by carbon-free resources, but may fail to do so if emitting resources are present in the generation mix. While use of behind-the-meter clean resources can guarantee compliance with emissions thresholds, the PTC could also be structured to allow producers using grid electricity to qualify subject to certain clean energy procurement requirements. In this work we model the evolution of the power sector in the western United States through 2030 to assess the emissions impact of the clean hydrogen PTC under multiple possible implementations. We find that with no requirements for grid-connected hydrogen producers to procure clean electricity, embodied emissions from hydrogen produced via electrolysis in California are worse than those from hydrogen produced via conventional, unabated fossil pathways. By contrast, requiring producers to match 100% of their electricity consumption on an hourly basis with locally-procured, 'additional' clean generation ensures embodied emissions rates equivalent to those of behind-the-meter installations. Failure to meet requirements for hourly matching, locality, or additionality of procured clean generation can result in significant excess emissions. Added hydrogen production costs from enforcing an effective hourly matching requirement rather than no requirements are less than $1/kg, and can be near zero if clean, firm electricity resources are available for procurement. Raw input and results data used in this work are available at: https://doi.org/10.5281/zenodo.7141069