Negative lags

Project managers sometimes like to use a negative lag to model that a succeeding task can start just prior to completion of the preceding task.

The best practice is not to use lags in Finish-Start relationships (https://www.pmi.org/learning/library/schedule-101-basic-best-practices-6701) 

In the CCPM world the consequence of this will be that if both tasks are part of the critical chain the buffer will be calculated over the duration of both activities. And if combined with task time reduction, the total length of the project after task time reduction and buffer creation will be more than it was initially. This can lead to a project that needs to start earlier (JIT) or will finish later (ASAP) than originally anticipated.

If you make the conscience choice to use negative lags, you can consider to adjust the length of the buffer for the negative lag.