I'm using a dataset ranging from 1960 to 2019, some of the years have missing variables. I want to run FE and perform EHA while maintaining the same observations (N and n). So, for the EHA, I have created a variable (gdppercapita) as a quotient gdp and population. I want to have event (where gdppercapita is below 3000 as 0 and above 3001 as 1) and spell where gdppercapita is below 3000 as 1 and above 3001 as 0. How do I go about this? (the N should not drop or change between models)