I ran and set the emc energy of all truth-identified klong clusters to
0. The result is 

Batch  Description                          BRBR
  o12  KL EMC clusters = 0                  0.0203027 +- 0.00250567
  o01  with KL scaling                      0.0201532 +- 0.00251257
  o00  default                              0.0202685 +- 0.00250473

At  least it's  in  the  other direction  compared  to increasing  the
deposited energy with Alessio's scaling.

In  case  you  check  in  the  logfiles (at  the  usual  places):  The
printout's claim  of scaling  with a number  is wrong, because  I only
updated the code but not the printout...