Seems strange to me as a layman that 16amp mcbs are used on cable that can only carry 14amps if routed in particular ways.
As above, this would not be permitted if the load were in excess of 14A. There is a simple principle:
Ib<= In<=Iz,
where Ib is the design load, In is the protective device rating and Iz is the cable rating after all factors have been applied for its particular situation.
In words, the MCB must be sufficiently large to supply the load, and the cable, as installed, must be sufficiently large to carry anything the MCB will let flow through it. The exception is where the MCB is only providing short-circuit protection to a fixed load, in which case the cable must still be sufficiently large for the load but the MCB might be larger than the cable's thermal rating. Note that although it can take 1.45In to actually trip the MCB, i.e. a 16A MCB will carry 23A, this is already taken into account in the cable rating.
When sizing cables near the limit, watch out for grouping where many cables exit the DB, especially on heating loads where many cables might run at full current simultaneously. Two cables grouped together for the first few feet will knock the rating back by 20% (Cg=0.8)
I guess I could just turn it on and monitor the small length of exposed cable for heat
You won't learn much from this unless you have spent a lifetime feeling cables. Yes, it will get warm, but how warm is too warm? And how much warmer is the length that is not exposed where there is less air circulation? Published ratings allow for the cable insulation to last its full design lifespan. A cable that runs hot might have its life curtailed by half, but you still won't know about that for a quarter of a century. OTOH it takes 80A or more to actually destroy 2.5mm² (often rated around 28A in typical installation environments) which would mean a heat dissipation (80/28)² = 8 times what it should be.
