Employment Data Inconsistencies

Hello, When comparing IMPLAN-based employment estimates with other secondary sources, such as BLS or CBP, it appears that there is a consistent pattern of IMPLAN’s estimates being significantly higher than the other single source-based estimates. We understand that IMPLAN's employment estimates are based on data from multiple sources, including the BLS Covered Employment and Wages (CEW) data, BEA Regional Economic Accounts (REA) data, and the County Business Patterns (CBP) data. However, we are confused as to why IMPLAN's total would be consistently higher than any one of those sources. This trend seems to hold for individual state-sector combinations (for the states we looked at), as well as at the national level (for individual sectors and the US total private non-farm employment) across multiple years. It seems that IMPLAN’s consistently higher employment estimates for individual states and the national total points to some kind of definitional differences between what IMPLAN defines as a job versus what these secondary sources might use. Has anyone else seen this pattern? Is there a specific reason(s) for this consistent trend?