The age-old debate over whether principals or district offices should hold the reins in schools has gotten a fresh empirical test, courtesy of a new NBER working paper from Northwestern University economist Kirabo Jackson. The question seems simple: Should principals be empowered as the CEOs of their schools, or should they mainly implement directives crafted by their districts?
How that question is answered could have implications for school cultures, teacher morale, resource allocation, and ultimately, how much students learn in school.
The new study uses data from Chicago Public Schools (CPS), where, in 2016, the Independent School Principals (ISP) program was launched. It gave select school leaders more authority over budgets, staffing, and instructional priorities. Only principals with strong past performance and favorable evaluations were invited and, even then, joining was voluntary. Once on board, these leaders received a salary supplement, agreed to stay at their schools for at least two years, and were empowered to tailor school policies to their students’ needs. For example, they might reassign specialist teachers into core classrooms, shrink class sizes, or redirect funds toward bilingual programs, counseling, or special education.
Jackson’s study leverages a decade of detailed CPS data (2009–2019) and compares the performance of ISP schools against carefully matched peer schools whose principals didn’t join the program. Sixty-eight elementary schools ended up in the treatment group, and the results they achieved are revealing when compared to control group schools. On average, student test scores rose by about 8 percent of a standard deviation, a modest but real effect. And in ISP schools, teachers and students also reported better school climates, class sizes fell by more than one student per class, and principals were less likely to bolt.
These improvements, however, weren’t evenly distributed. The strongest leaders, including the principals with the highest past ratings and results, drove the most impressive school gains. By contrast, principals with relatively weaker records sometimes presided over stagnation or even declines. This pattern underscores an uncomfortable truth: Autonomy amplifies both strengths and weaknesses. For skilled leaders, it can unleash a virtuous cycle of stability, morale, and achievement. For less capable ones, it risks compounding problems.
Another important finding of the study is how schools with atypical student populations benefited disproportionately from greater principal autonomy. Schools with many English learners, high concentrations of students with disabilities, or demographics far outside the district norm saw especially strong gains under ISP. That makes sense: Central-office policies are inherently one-size-fits-most. The more a school deviates from the “average,” the more useful it is to have a principal who can make adjustments. In such settings, local discretion can be transformative.
The study also looks at the mechanisms behind these effects. Principal stability was perhaps the most straightforward. Turnover plummeted after leaders joined ISP and remained lower even after the two-year commitments expired. This alone has educational value, since frequent leadership changes can destabilize schools. Jackson says that principals choosing to alter their schools’ allocation of resources is also probably part of the story. Class sizes dropped, schedules were reworked, and in schools with atypical student needs, funding flowed toward bilingual or special education services.
Most impressively, all this occurred without additional resources devoted to the ISP schools, apart from the small bonuses to the principals. These changes are a result of different management and resource allocation, not increased spending. That these positive effects were realized without new funds, splashy technology purchases, or costly retraining efforts implies that the return on investment from this program is off the charts, even if the effects are modest.
Still, caveats abound. This wasn’t a blanket district policy but a selective intervention aimed at proven leaders. The results, then, tell us what happens when already capable principals get more autonomy, not what happens when every leader does. The fact that, even within this selective group, weaker principals sometimes stumbled may suggest that the eligibility bar for the ISP was set too low. Expanding autonomy indiscriminately could backfire.
Still, Chicago’s experience shows that when schools have effective leaders and distinctive needs, giving them more latitude can yield real dividends for kids—all without draining the public purse. That’s a lesson worth heeding for reformers nationwide.
SOURCE: C. Kirabo Jackson, When Decentralization Works: Leadership, Local Needs, and Student Achievement, NBER Working Paper No. 34216, National Bureau of Economic Research, September 2025.
