1. ask important questions. it's helpful if people care about the summary stats in your paper
2. consider what we already know about the subject and how your paper will improve our understanding of the world relative to existing lit/ data/approaches/models.

3..Some areas are much more crowded than others. The 100th paper on unemployment insurance better be super insightful/ interesting, and you may have better luck revisiting classic questions on capital taxation/ spatial public finance where there hasn't been as much empirical work

4. Think about the policy debates that we will likely be having in the coming years and what types of evidence/ parameter values do we need to inform these questions