A team of students from Carnegie Mellon University’s Heinz College of Information Systems and Public Policy helped the U.S. Government Accountability Office (GAO) build the first known comprehensive inventory of state and local semiconductor incentives in the post-CHIPS and Science Act era. Their work follows a $50 billion federal investment to revitalize the U.S. semiconductor industry.
Found in most electronic devices, semiconductors serve as the "brains" for phones, cars, medical devices and computers. Master’s students Henry Albright, Shawn Lu, Ana Rowley, Isabella Veneris and Albert Yang worked with the nonpartisan GAO on a project spanning policy analysis, data engineering and spatial mapping. Their work gives federal analysts a clearer foundation for evaluating how public investments are shaping advanced manufacturing, workforce development and regional economic growth across the country.
A national question without a clear answer
The CHIPS and Science Act requires the GAO to assess how federal investments are unfolding over time. To do that effectively, analysts first needed a clearer understanding of the policy landscape beyond Washington, particularly the incentives states and local governments were offering alongside federal funding.
For the GAO, the challenge included both scope and lens. "While we’ve been charged to look at the federal angle, we were also aware that there were some non-federal elements," explained Candice Wright , the GAO’s director for science, technology assessment and analytics. "Recognizing that the states can play a really important role here, we wanted to get that perspective."
"For the GAO to have a better understanding of how the federal legislation is working, the analysts need to understand how state and local policies are impacting the industry," said Albright, a second-year student in the Public Policy and Management master’s program at Heinz College.
That information exists, but not all’in one place. Data related to incentives is scattered across legislation, economic development agencies and public records that vary widely by state. Some programs are highly targeted, while others are embedded in broader economic policy. Comparing approaches to understand how they interact with or are influenced by federal investment is difficult without a consolidated baseline.
The GAO asked the student team to help close that gap.
Building a foundation for federal oversight
Rather than producing a one-time report, the team focused on creating a tool the GAO could continue to use.
"What we built was essentially a database that could serve as a baseline for the GAO," said Rowley. "It’s a structured database that they can continue to add to and then analyze that data in different ways." The database included 133 state and local incentives across 37 states; the students coded each policy using consistent categories that allow for future analysis.
To do that, students translated dense legislative language into standardized variables. This approach allows analysts to test different questions over time, using employment data, investment patterns or other outcomes without rebuilding the dataset from scratch.
For GAO analysts, the value wasn’t just the data itself, but what it made possible. "The final deliverable provides us a really great resource," Wright said. "If we have questions from congressional staff down the road about state legislation, we can go to the tools the team provided rather than starting from scratch."
Faculty adviser Jordan Pallitto, a Heinz College adjunct professor, emphasized that
"This project sits right in Heinz College’s wheelhouse," he said. "It’s a blend of qualitative and quantitative thinking applied to a real-world problem."
When the data complicates the story
As the team began analyzing the data, some findings challenged common assumptions. One example came from case studies they created to compare states with different incentive strategies. Arizona, for instance, attracted significant semiconductor investment despite having relatively few state-specific incentives, while other states with more targeted programs saw different results.
"The amount of money that states spend on incentives doesn’t necessarily impact the number of semiconductor facilities in their state," Albright said.
Rather than pointing to a single explanatory factor, the data suggested a more complex picture shaped by existing industry presence, federal investment timing, workforce capacity and historical development patterns. Importantly, the team avoided drawing causal conclusions that the data could not yet support.
"That complexity is an example of how capstone projects help students transition from classroom theory to real-world application," Pallitto said. "The team had to decide what mattered and whether something would actually be useful to GAO. They had to think like policy analysts."
In some cases, the students’ analysis revealed gaps the GAO hadn’t yet identified. "We had a lot of unknown unknowns," said Kelsey Kennedy, senior analyst at the GAO. "This project helped surface things that weren’t on my radar before. If I had just reviewed the database quickly, I may not have come away with those conclusions. The case studies helped pull out what mattered and why it mattered to policymakers."
Learning to work with imperfect information and uncertainty
The project required skills that went beyond technical analysis. Unlike most classroom coursework, there was no predefined dataset and no guarantee that the information they needed would be available or consistent across states.
"We had to decide what data and which variables we should include," said Yang. "The experience helped me a lot."
Students also had to learn when to stop searching for data that simply wasn’t accessible. "With school assignments, you’re used to knowing there’s an answer," Albright said. "Here, knowing when to stop researching was part of the job."
The experience reshaped how several students think about public policy analysis - particularly the need to balance rigor with practicality when working under real-world constraints.
Midway through the project, a federal government shutdown limited communication with GAO project sponsors, forcing the team to continue their work with minimal client feedback. The experience mirrored conditions analysts often face in professional settings.
"It was weird working on a project when nobody could legally respond to us," said Veneris. "But we still had deliverables due."
Their initiative did not go unnoticed by the project’s federal partners.
"It was incredibly impressive the amount of work the students did, especially given the government shutdown. The fact that they stayed resilient, focused and on track was really impressive," Wright said.
Work That Matters
Researchers at CMU are working on real world solutions to the biggest challenges.
Read more about the latest discoveries.
Lessons in Real-World Impact
The students are careful not to overstate the project’s immediate impact. Their work does not make policy decisions or predict outcomes. Instead, it provides infrastructure - data, documentation and analytical groundwork - that enables oversight to happen more effectively over time.
"Oversight work isn’t flashy," Rowley said. "But it’s essential."
"This work gives us a sense of the array of policy levers that are out there. Understanding that range is important when we start looking at policy options in the future," Kennedy explained.
And for the students involved, it marked the moment when academic training became something more tangible: work that extends beyond the classroom and into the systems that shape public life.
