A while ago I had some email with Noah Smith on this topic, now we are getting somewhere, this is from a new NBER working paper by Daniel M Hungerman, Kevin J. Rinz, and Jay Frymark:
We use a dataset of Catholic-parish finances from Milwaukee that includes information on both Catholic schools and the parishes that run them. We show that vouchers [funded by the government] are now a dominant source of funding for many churches; parishes in our sample running voucher-accepting schools get more revenue from vouchers than from worshipers. We also find that voucher expansion prevents church closures and mergers. Despite these results, we fail to find evidence that vouchers promote religious behavior: voucher expansion causes significant declines in church donations and church spending on non-educational religious purposes. The meteoric growth of vouchers appears to offer financial stability for congregations while at the same time diminishing their religious activities.
I’ve long maintained that the fiscal effects of vouchers, if they were implemented on a much larger scale, are the elephant in the room. For better or worse.