Catholic schools, once a mainstay for the Irish, Italian, and Polish communities in American cities, are struggling. With shrinking numbers of nuns as a source of free labor, and fewer parishioners passing the donation baskets on Sunday and enrolling their kids in parochial schools, many simply cannot afford to keep their doors open. Just last week, the Archdiocese of New York announced the closure of five more schools for financial reasons; that’s on top of dozens that were shuttered in 2011 and 2013.

At least in Milwaukee, Catholic churches have kept their schools alive with the help of vouchers—public money given to parents to spend for their children’s education at the private school of their choice. The economists Daniel M. Hungerman and Kevin J. Rinz and the church administrator Jay Frymark spent three years poring over the financial records of 71 parishes in Milwaukee—information that is rarely shared with researchers—between 1999 and 2013 to understand the impact of school vouchers on churches. In a National Bureau of Economic Research working paper, they explain that vouchers actually staved off imminent school closures in Milwaukee, though they did not improve the church’s overall finances.

Secretary of Education Betsy DeVos is a strong advocate for school choice. Donald Trump pledged to create a $20 billion block grant to the states to expand charter schools and vouchers. Given the new administration’s embrace of school choice, it is important to understand who benefits from school voucher money, as well as its impact on children and the community at large. The vast majority of students who attend private schools through vouchers are enrolled at religiously affiliated institutions.

* * *

There are about 6,500 Catholic schools in the country, educating roughly 2 million students in 2014. At their peak in 1965, parochial schools educated 5.6 million students.

In an interview, Daniel Hungerman, an economics professor at the University of Notre Dame, explained that Catholic schools have been undermined by an overall decline in Americans who are religiously affiliated in the past decade. The Catholic Church, in addition, has been rocked by the high-profile cases of child abuse by parish priests. The changing demographics of cities like Milwaukee, Cleveland, and Washington, D.C., have also had a large impact on church attendance. Hungerman and his co-authors found that the number of baptisms in Milwaukee dropped by 25 percent between 1999 and 2011.

Fewer people in the pews mean less money in the basket on Sunday, and less money for the schools; typically local parishes, not the archdiocese, support these schools. With fewer worshippers donating dollars and sending their kids in ties and kilts to St. Ann’s or Our Lady of Peace, Catholic schools have been hit hard. Charter schools have also emerged as competitors for students.

While many Catholic schools have been forced to raise tuition to pay lay teachers in the place of nuns who formerly provided that labor for free, their tuition is still remarkably modest. According to the study, the average Catholic school tuition nationwide was $4,200 during the 2007-08 school year, which is far less than the per-pupil expenditures of public schools.

Milwaukee launched its voucher initiative, Milwaukee Parental Choice Program (MPCP), in 1990. The program has expanded its eligibility criteria over the years, most recently in 2013; vouchers today are restricted to families—if comprised of four people—that make $72,900 per year or less. Currently, 28,188 students take advantage of the program, attending 121 participating schools and, on average, each receiving—as of last school year—a voucher worth $7,384.In 2012, according to Hungerman, the state spent $154.6 million on vouchers for Milwaukee residents.

The clear beneficiaries of Milwaukee’s voucher program are religious schools, including parochial ones. Nearly 90 percent of the voucher recipients in Milwaukee attend a religious school, which is close to the national average; from 2007-09, 85 percent of all students attending a voucher accepting school attended a religious school. A plurality, if not a majority, of these religious schools are Catholic.

The U.S. Supreme Court has ruled in two cases—Zelman v. Simmons-Harris (2002) and Arizona Christian School Tuition Organization v. Winn (2011)—that school-voucher programs that indirectly support religious schools do not violate the Establishment Clause in the First Amendment. In the Zelman case, the justices found that Cleveland’s voucher program was constitutional on the grounds that it was “neutral with respect to religion.” The city gave vouchers directly to the parents, who had the ability to freely choose between a number of school options for their children; they could opt to use that money for, say, a Buddhist school or a secular one.

Much of the research on vouchers to date has focused on education outcomes, showing only modest, if any, gains in academic achievement for participating students, although vouchers might boost educational achievement in neighboring public schools. Hungerman, Rinz, and Frymark’s work focused on the economics of vouchers—specifically, its impact on the church and the parish.

Using the financial records of the 71 parishes, they compared parishes with voucher-accepting schools to two control groups—one including parishes without schools and one including schools that don’t accept vouchers. After creating a model that predicted the likelihood that a school would close, they found a very strong association between school vouchers and an educational institution’s financial well-being. Vouchers typically increased the revenue for the parishes and prevented school closures or mergers. Catholic-school closures were fairly common in Milwaukee, they found, until the first major voucher expansion took effect in 2006, when the city expanded voucher-eligibility criteria and the number of participating students grew.

Vouchers are now a dominant source of funding for many churches. The average voucher-accepting Catholic school in Milwaukee received about $1 million per year from the program in 2012, according to Hungerman, which is a serious help for the cash-strapped parochial schools. “Parishes in our sample running voucher-accepting schools got more revenue from vouchers than from their worshippers,” Hungerman said.

But does this million-dollar influx to the parochial schools free up donations to be spent on other items—say, organs and stained glass? Not quite. According to Hungerman, they found that voucher-accepting churches also saw a significant decrease in donations and, thus, had to reduce spending on items such as staff salaries, mission support, and church maintenance. In fact, the Milwaukee voucher program may have contributed to a decline in non-educational church revenue of $60 million over time. “There are no new church organs in Milwaukee churches,” Hungerman said.

Hungerman and his fellow researchers are puzzled about why school vouchers would lead to a decline in Sunday offerings from parishioners. Perhaps church members were less generous when they saw that financial support came from other sources or maybe members decided to move to a different church when new students entered the school. Their research could not answer this question.

Hungerman, Rinz, and Frymark’s research ultimately shows that vouchers can both help and hurt Catholic schools, allowing them to keep the doors open but also, at least in the short term, driving away loyal parishioners and diminishing other religious activity. Regardless, the fact that these findings reiterate how much religious schools benefit from voucher funding is especially relevant given the policy preferences and ideological track record of the current administration. If vouchers really do jump start the sputtering engine of Catholic schools, then it may be time to examine how those school run and evaluate whether the old jalopy shall be traded in for a new model.