Event-Study Project

The pathway has separated the earnings date, the windows, the benchmark, the residual, the sum, the overlap problem, and the covariance. This note writes one design memo that uses all of them on a single announcement, then computes a three-day CAR from stated inputs so the arithmetic can be checked.

The memo also lists the observations that would falsify a claim that this CAR is an announcement effect. A second, equally legitimate event — an index inclusion — is used only as a contrast in the dependence paragraph, because inclusion is totally clustered by construction.

• write a one-page design for one earnings date: event, windows, benchmark, dependence, falsifiers;
• compute the three-day market-model CAR and its large-$L_1$ standard deviation from the fixture;
• say what the ratio of those two numbers is, and what it is not;
• name two features of the design that would make the ratio incomparable with a standard normal.

1. Design memo
2. Fixture
3. Audit on the running announcement