There have been a number of stories recently about litigation (there are a few pending cases), or threats of litigation, premised on the notion that fraud or mismanagement must account for discrepancies between the number of entries on registration rolls and Census estimates of citizen population. There are plenty of reasons to be skeptical about such claims: even apart from the fact that the measures cover different time periods, involve comparisons of survey estimates to individual counts, and are often inflated by pending entries slated for impending deletion, the measures fundamentally involve apples and oranges. Census depends on usual physical presence and registration (usually) depends on domicile. Those are often, but not always, the same. Members of the military stationed domestically, for example, are counted by the Census where they are stationed, which may or may not be their voting residence: those still domiciled “back home” are often validly registered there even if they’re counted elsewhere by the Census.
PILF recently issued another in a series of letters claiming mismanagement of 248 counties’ voter rolls based on comparisons like this; the claims have drawn extra attention recently because the organization’s President and General Counsel is also a member of Sec. Kobach’s Advisory Commission on Election Integrity.