What are the "objectively harmful externalities" you're referring to?
More informally, smartphone usage by children promotes a short attention span, a lack of any sense of presence where actually situated in the physical world, as well as less and lower quality interaction with others, leading to poor social skills, anxiety, social isolation, and a focus on superficial social signaling over meaningful human interactions, ultimately producing the mental illnesses referenced above.
See e.g. https://kagi.com/search?q=summary+of+mental+health+outcomes+... -- there are far too many sources to even list here.
I'm not even talking about methodologies or replication or evidence or p-hacking (all of which are huge challenges to this sort of research).
On a much more fundamental level, the statement "unauthorized smartphone use in a classroom setting is objectively harmful" is a defensible statement. I don't need a study to tell me that, nor should anyone. The extraordinary claim in this case would be the opposite, for which I would have to see tremendous evidence, and which even then I would likely not believe.