No, that will cause the reading to be too low rather than too high, and even if that input were zero, the numbers would still not add up.
It sounds as if you may have a batch of CTs of dubious origin. While SCT013-000 has been a workhorse and are usually pretty accurate, I have had some that read chronically low. Maybe a hairline crack in the brittle iron core, maybe a leak in the TVS diode, maybe a short in the winding. It’s probably a good idea to test them for consistency.
If you have an accessible length of mains cable or the outbuilding cable, you might just try clipping all of the CTs around it at the same time to see how close they read. If there are outliers, try swapping the input channel with another “good” one. If the problem moves, it’s the CT. If it stays on that channel, maybe it’s the IoTaWatt. Haven’t had any quality problems like that, but there’s always a first time.
Regardless of where the fault lies, it should be corrected and the inputs should add up very close to the mains. I say close to the mains because the SCT013-000 isn’t the best option for low wattage circuits. They have a lot of phase shift at low current and can also be a little off on linearity. The problem is compounded in 230V countries where the current for a given power is half. There are better choices for those circuits.