This is why officials in many states, even ones that were gung-ho social distancing like California and Washington, were reluctant to close the schools. It was predicted that a large number of kids would instead just be congregating together on streets, defeating much of the point of closing the schools. But they were caught between a rock and a hard place. If they didn't close the schools, the kids become an infection vector that undoes much of the sacrifice of having people wfh, avoid social gatherings, and cancel large events.
It's refreshing that the attitude you're describing in parts of the USA aligns with what we saw and heard in AU -- almost every comment is about the health and wellbeing of the children, and rarely included any consideration of the risks to the administration and teaching staff at the schools, the public transport workers moving the children to and from the schools, or the risks (mostly from) the parents doing pick-ups and drop-offs.
Any anyone who's worked in a school knows how quickly diseases spread there. Like, my district, and all the districts near us, routinely close down every February come flu season as kids come to school with it (or other diseases) and they spread like wildfire. We take 3 days off, give the school a deep cleaning and let the kids pass their infectious period or get to the doctor, and then come back. Diseases spread like wildfire in schools.
This paper[1] on a Lancet sister publication disagrees, on the basis of literature review, that in this specific case the schools are a major part of SARS-CoV-2 infections. Of course there are many caveats in the discussion section.
From your source: "Recent modelling studies of COVID-19 predict that school closures alone would prevent only 2–4% of deaths" a 2-4% reduction in death rate implies quite heavily that children would be a disease vector.
