If it would be correctly implemented, the account deletion should cause the replacement of the Global unique Boinc user ID (unique identifier) with a new pseudonymised ID as well as the WCG account ID with an alternate ID. Obviously WCG should not keep the cross information, making possible to re-identify the original user from the pseudonym.
The statistics data should remain available, including for quantifying the contribution to the team, but the data are not related anymore to a specific individual.
Obviously, deleting the account does not impact the science results.

I don't know how the WCG's lawyers did plan the final implementation.
Cheers,
Yves

Clearly we don't (yet) know exactly how this will be implemented, but I think Yves' idea is about right. Another option might be to have a "deleted" bucket that the stats get added into, perhaps one at global level and another at team level.
But I think the OP was worried about the scientific data. That is removed from the system fairly smartish and sent to the scientists, and I'm sure that there is no chance that deleting a user would delete any of the associated completed science -- I do not believe that that is associated with users in any way when it is returned to the scientists.

Yves is correct that if a member decides to exercise their right to erasure, or their account is flagged for deletion under our data retention policy, then all the aggregate level statistics are not impacted (e.g. each team would still include in its total contribution stats the contributions of all deleted members), but all the member-level data, including cross-reference IDs, is deleted.

Also, the science data isn't impacted by account deletions because once processed, that data holds no ties back to individual members.

Thanks,
Juan