Education Policy
This hub is the education-finance practice layer behind Too Early To Say. It shows how applied economists use an AI agent to move from a federal funding rule to a reproducible district-budget claim. Each step is one phase of an IDEA and California school-finance analysis. The agent handles source retrieval, table extraction, and first-pass checks. The design, denominator, and interpretation remain ours.
This is for economists, policy analysts, and education-finance researchers who already know their methods and want an agent in the workflow without losing control of the analysis. The problem it solves is simple. Hours go to tracing statutes and reconciling tables, while identification and interpretation are squeezed into what is left.
One idea runs through all three steps. A funding rule becomes policy evidence only when the denominator, comparison, and public output match. When the specification names each quantity, the agent can check it; when a quantity is missing, the claim stays under review.
Step 1
The named case starts with IDEA's 40% of national average per-pupil expenditure benchmark. The agent retrieves the statute and formula language. We decide which expenditure denominator belongs in the comparison.
Outcome: A statutory benchmark with an explicit denominator.
Step 2
The article reports an 8.7% aggregate decline in the final ARRA-funded year as support was winding down. The agent recomputes the displayed table and maps each funding category. We decide what the descriptive comparison can and cannot establish.
Outcome: A descriptive comparison with a stated counterfactual limit.
Step 3
The article begins with special education at 16.7% of the displayed total. The agent checks row arithmetic and traces the source of each series. We decide whether the reconciled evidence supports a budget-shifting mechanism.
Outcome: A reconciled budget-mechanism test.
Defines the federal benchmark, district contribution, and denominator problem.
Separates an aggregate funding comparison from a causal stimulus claim.
Documents the arithmetic and source reconciliation needed before testing a budget mechanism.
This sequence organizes those lessons into one path. It is an entry point for economists who want an agent in the loop and still want every number to stand up when someone else reruns the script.