Why is the number of dead tuples causing the performance of deferred triggers to degrading so rapidly (exponentionally)?

Obviously,

this part of tr_f_def():

******************************
-- delete the contents
-- delete from f;
IF EXISTS (SELECT 1 FROM f) THEN
DELETE FROM F;
VACUUM F;
END IF;
******************************

should simply read:

******************************
-- delete the contents
delete from f;
******************************

--
Best,

Frank.