You have data on individual contributions over time to pre-tax college savings plans (529 plans) in two different US states. Suppose one of the states implemented a policy seeking to increase the amount of such savings (called Treated in the below). You have data both before and after this policy change (so in the below, After-1 is after that policy change). You run a difference-in-difference model and obtain the following results: log (contributions) 20+12Treated, -2After, +5 (Treated; x After) it (there were also other variables included but their coefficients are not reported). The interpretation of the coefficient on the interaction Treated*After is: O O O O The negative impact of the policy change on both states. The 5% increase in contributions in the periods after the policy change for the treated state. The 5% difference in contributions in the periods after the policy change in the state that made no policy change. Part of the 35% higher contribution (20+12-2+5) that occured in both states.