Analyzing the Anticipation of Treatments with Data on Notification Dates

When treatments may occur at di®erent points in time, most evaluation methodsassume - implicitly or explicitly - that all the information used by subjects about theoccurrence of a future treatment is available to the researcher. This is often called the\no anticipation" assumption. In reality, subjects may receive private signals aboutthe date when a treatment may start. We provide a methodological and empiricalanalysis of this issue in a setting where the outcome of interest as well as the moment ofinformation arrival (noti¯cation) and the start of the treatment can all be characterizedby duration variables. Building on the \Timing of Events" approach, we show thatthe causal e®ects of noti¯cation and of the treatment on the outcome are identi¯ed.We estimate the model on an administrative data set of unemployed workers in Francewhich provides the date when job seekers receive information from caseworkers abouttheir future treatment status. We ¯nd that noti¯cation has a signi¯cant and positivee®ect on unemployment duration. This result violates the standard \no anticipation"assumption and rules out a \threat e®ect" of training programs in France.