I am working on a quasi-experimental study with a large unbalanced panel dataset. There are N=300,000 and T=20, where roughly 50,000 individuals receive treatment since several different periods.

I use a DiD estimator to estimate the ATT:

$y_{i,t} = \beta \times D_{i,t} + \alpha_{i} + \delta_{t} + \epsilon_{i,t},$

where $D_{i,t}=1$ iff individual $i$ has received treatment at period $t$. Besides, $\alpha_i$ and $\delta_t$ are individual FEs and time FEs, respectively.

Here comes my concern. When $t=1$, I have only 80,000 individuals in sample, and the number gradually increases to N=300,000 over time. The periods with treatment events are $t=6, 8, 10, 12$. That is to say, many individuals only have observations in later periods of the sample.

Is there an issue if I use all observations to estimate the equation? Is it legitimate to keep only individuals that exist from the beginning?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.