For the avoidance of doubt, I'm talking about the main bonds between the service earth terminal and other metallic utility services, structural steel work and lightning conductors. In addition to anything the Regs may require, for a PME supply the network operator insists on this bond, and specifies its size, to ensure among other things that equipotential conditions would be maintained under a lost neutral condition. Without it, in such circumstances the supply neutral (and thus the electrical earth) could rise in potential while other metallic services could remain at or about earth potential, creating a hazardous potential difference. In most domestic situations such bonds would connect only to gas and water (telecomms being specifcally excluded).
As for fault current, as most gas and water utility services are now plastic, where is the earth fault current going to go? If we're talking about a fault between a live conductor and a metallic water pipe (say) somewhere in the house, then the fault current would have multiple return paths through all the supplementary bonding and there certainly wouldn't be a need for a 10mm cable at the main intake to carry it back into the mains neutral, bearing in mind the sub-circuit feeding the fault will probably be no more than 4mm.
If we're talking about fault current flowing back into the main earth terminal, this will take the lowest impedance path back to the network operator's distribution transformer, which will be through the service/mains neutral conductor if the service is PME; any other path (assuming there are continuous metallic services connected to the cross-bonds) will be only secondary and of relatively high impedance. So the size of the bond is not related to this scenario.
Finally, what I said is not a theory, it's a fact, albeit a little appreciated one outside of the engineering departments in the electricity supply industry.