Increasing Government Effectiveness: Evidence, Implementation, and AI

This module builds the practical apparatus for reading research critically: the four ways a correlation can arise, the strategies researchers use to construct a credible counterfactual, and how to judge whether a result travels to your context.

Participants work through the seven identification strategies a reader needs to know, learn how to select the estimate a specific decision requires, and apply a four-question generalizability framework to research of their own choosing.

No regressions are required and no prior statistics are assumed, but the course is designed with different levels for those of increasing sophistication with both statistics and coding.

Most reforms do not fail because the evidence behind them was wrong. They fail in the distance between a decision and a delivered service. This module is about that distance, and how to lead your teams across it.

Participants diagnose why reforms stall, isolate workable problems, map political economy and coalition dynamics, and learn to sequence action by where there is room to move rather than where the problem is largest. The module also covers the latest from personnel economics: who to hire, what incentives improve and what they degrade, and how much autonomy to grant.

Effective monitoring is not about collecting more data. It is about knowing what you already have, what it can answer, and how to use it to learn rather than to report.

This module maps the data sources already available inside most government institutions, teaches participants how to interrogate any administrative dataset for its source, coverage, and gaps, and covers the ethics of monitoring from the design stage.

Participants leave with a monitoring plan for their own first test. AI and text classification on administrative records are covered with deliberate attention to verification.

What works in a pilot does not automatically work at national scale, and the gap between the two is where most large-scale reforms fail. This module covers the specific reasons effects shrink as programs grow, how to design evidence built for a scale-up decision, and why implementation detail is often what decides whether a good policy reaches anyone.

This module will feature cases that are drawn from real government programs to illustrate the full range of outcomes — successes, failures, and programs that caused harm no one anticipated. Guest practitioners who have navigated these transitions inside government join for individual sessions, bringing firsthand experience that goes beyond what the research alone can offer.

Participants arrive at the University of Chicago with a real policy problem from their own institution and work through it across five days with faculty, peers, and government practitioners. The challenge is structured around the operational problems that most commonly derail reform at scale: whole-of-government coordination failures, implementation sequencing, cross-agency KPI design, and procurement and budget cycles as policy constraints.

Participants will diagnose the problem, identify the key actors and reform levers, develop a causal logic model linking intervention design to expected outcomes, and write a policy brief detailing key recommendations for a senior decision-maker audience.

In line with the program’s broader mission, faculty presence during the Policy Challenge extends beyond classroom interaction. Faculty will be available for conversations with participants, informal roundtables, and advising sessions to maximize the institutional value of the visit.