Modern RCBOs do have some electronics in them so do consume a bit constantly, even when unloaded.

The wylex ones for example  take about 2.5mA from 230V (though with a capacitive power factor so not even half a watt. ) .  Obviously the MCB part dissipates a bit in normal use in proportion to the square of the load current, but no more than any other breaker (edit I have just loaded one up on the bench and at 5A  it drops 180mV, so ~ 35 milliohms and ~ 1 watt  )
I'm at a loss to say why lighting should be much worse than anything else, unless they are much closer to full load or at the top of the box where it is hottest.

Mike.

It's strange isn't it.  We're seeing it more and more on thermal surveys.  In todays example the board and RCBO are newish (less than 2 years old) and the circuit is LED lighting, nowhere near enough for it to be anywhere close to full load.

I'm wondering if this potentially a faulty device given such an extreme variance in temperature from ambient.  

It's strange isn't it.  We're seeing it more and more on thermal surveys.  In todays example the board and RCBO are newish (less than 2 years old) and the circuit is LED lighting, nowhere near enough for it to be anywhere close to full load.

I'm wondering if this potentially a faulty device given such an extreme variance in temperature from ambient.  

Nicholas Moxey said:

and the circuit is LED lighting,

This might be part of the problem. The led lights are not a pure resistive load and therefore there may be harmonics messing up the heating effects. Just a thought.....