It's a vicious cycle: good schools -> increase demand -> high home prices -> affluent residents -> more tax revenue -> better schools -> repeat

The reverse happens with bad schools.

In this case I was only including the schools in the town's unified school district, so the tax-base is identical. There are a couple of school districts in affluent areas just outside of town that each have only 1 elementary school. One school is funded to double that of the unified district on a per-student basis, and the other triple. On top of that, a lot of activities at schools are funded from donations from parents, which are obviously higher at the more affluent schools.

[edit] The two schools I mention on the outskirts of town are 6% and 8% hispanic, respectively.

This is why schools should be funded evenly per student by the state or federal government. Not municipal.

This is a very important policy reform. Readers outside the US may not realize that the school funding system has this inbuilt feedback loop: good schools get more funding from property taxes and keep nearby property values high, bad schools get less funding from property taxes and keep nearby property values low. School quality is often key in determining home buyer location preferences.

School quality also affects home buyer location preference significantly in Canada as well. But Canada doesn't feature this particular financial feedback loop (though there's another feedback loop of well-learned students attracting more parents with well-learned children. But that is not institutionalized and far less severe a problem).

Is school funding distributed in such a small district that tax revenue in the district itself affects the school budget?

That is often the case in the US, yes.

A recent court ruling in Connecticut may change that, at least in that state. Some links:

https://www.google.com/amp/mobile.nytimes.com/2016/09/08/nyr...

https://www.google.com/amp/mobile.nytimes.com/2016/09/12/nyr...

http://mobile.nytimes.com/2016/09/16/nyregion/connecticut-ed...

And yet in Connecticut there are many well regarded school systems in affluent towns that spend about the same per student than the failing schools in Connecticut cities.

http://www.courant.com/data-desk/hc-in-perpupil-spending-con... https://k12.niche.com/rankings/public-school-districts/best-...

Cities: Waterbury ($15,859), Hartford ($18,397), Bridgeport ($15,228), Stamford (18,188), and New Haven ($17,084) spend on average $16,951 per student.

Some well-regarded suburban school districts: Simsbury ($15,969), Ridgefield ($17,315), Glastonbury ($15,481), West Hartford ($15,720), for an average of $16,121.

Clearly there is something going on beyond the dollars spent per student.

In many places in the US, this is the case, but not for the sample I was choosing (see my sibling reply).

The above post was with somewhat old data; I downloaded the relevant database from california, and using the most recent numbers (2014) updated things[1]. There are fewer schools because I originally included schools that were not geographically based, as we are talking about geographic segregation. The blue dots correspond to % hispanic students, the pink is % qualifying for free/reduced school lunch, and yellow is % English learners. The Y axis is state standardized test scores.

There are now not any schools with less than 20% hispanic student populations (there is still a charter school with under 20% hispanic, but it was excluded per above), but there is a huge hole between 45% and 75% hispanic, and 6 out of 10 are 80% or more hispanic. If children were assigned randomly to these schools, one would expect a normal distribution around the population mean.

And to clarify, all of these schools are within a single school district and pull from the same funding source.

1: http://imgur.com/a/8iLt9

I'm guessing either Riverside or MoVal?