Thought id do a test.
I took a used Cyl head fitted a helicoil to the exhaust guide oil feed hole.
I used a new, dry, solid M7 8.8 bolt with similar thread engagement to that of a banjo bolt, fitted it to an oil union hex (with a steel washer for the correct depth) and torqued it to destruction using a deflecting beam torque wrench to gauge the strength.
I tested the thread in an inlet bolt hole in a similar way.
Result
The original thread torqued up to 10ftlb(13.5Nm) it then held that torque for 4 revolutions of the bolt as the thread progressively failed in the head. (The main force of the bolt is taken by the top few threads. As they shear the strain is passed on to the next coils and so on, it is a progressive failure untill there are no threads left). You can see below that the thread has unwound from the inside of the hole.
DSC00206 by
Sean602, on Flickr
The helicoiled hole took 14ftlb(18.9Nm) untill the torque remained stable despite further turning. At this point I took the joint apart and no damage could be seen to the thread. I reassembled and gave it a further few turns at which point the bolt sheared. There is slight damage just visible to the alloy at the start of the helicoil where the distorting bolt has pulled up the first coil of the insert, but as is claimed, the coil insert moves and the strain is distributed through out the full thread engagement. A much stronger joint.
DSC00208 by
Sean602, on Flickr
What is worrying about the results is that the original thread could be hiding damaged threads as each time the banjo bolt is tightened more of the thread is damaged untill there is none left. Also the Haynes manual gives the torque figures as 10ftlbs for the banjo bolts ensuring that damage almost certainly occurs each fitting,
especially if the threads are oiledSean