Cannabis is a C3 plant so it looses a considerable amount of carbon to respiration under current ambient atmospherical CO2 conditions. Photosynthesis happens mostly within the leaves, inside the thylakoid lumen of the chloroplasts. The other things that happens there is a processing of NO3-, NH4+ and S. It’s the xylem that distributes most water up from the roots to the leaves, most of it evaporates as a means to cool leaves and do gas exchange via open stomatas to let CO2 in. Once done these, and the photosynthates, get loaded into the phloem, which can distribute what’s needed bidirectionally, though it depends on specific gradients caused by the aforementioned source/sink-mechanism(s).
But unlike the xylem, leaves consists of living cells, these need energy to stay alive, so their mitochondrias consume some of the produced energy to stay alive.
Now there’s a point in a low-light scenario for a leaf where the carbon-fixation and its loss by photo- & darkrespiration cancel each other mutually out - this is called “the light compensation point”. My only source, which doesn’t include a scientific reference, tells me for Cannabis this would be 150umol/m^2/s. Usually for C3 plants this would be lower by factor approx. 5-6. Does somebody here have a better source maybe, that would be nice to know.
This could mean that in an indoor setting, where PPFD greatly diminishes with distance (in comparison with an outside grow) these lower defoliation could be meaningful, but we shouldn’t forget that a quantummeters sensors is actually not able to monitor all photons arriving at a measuring point, not at all. It basically just tells you how strong the stream from one direction is, and totally neglects the other ambient diffuse light coming from the other directions, where the sensor is “blind”.
Now outside as good as each and any leaf will be able to positively generate carbohydrates for the plants to acquire more biomass.
However, there is not always a linear relationship with photosynthesis and yield, although there is no other more significant relationship with yield than actually photosynthesis. The thing is that plants usually have a surplus of energy at their hands and so it may not always matter that much if they loose a fraction of it. This is especially true in some artificial setups like hydroponics where they will still loose quite much to feed the microbes, which may just be wasted in a bucket of water instead.
Also, leaves do get old and then gradually loose their ability to do proper photosynthesis. Chloroplasts have several stages, when the leaves are very new and still yellow, it’s because they are still filled with etioplasts which differentiate at the first day of being in contact with PAR light, later these chloroplasts turn into dysfunctional gerontoplasts. All these cell organelles can be broken apart by the plant and then contribute a good amount of energy, and essential nutrition, to the plant. So a heavy defoliation also deprives a plant of its “emergency” nutes. It’s stands to reason that the so-called “nitrogen use efficiency” would drop by doing so.
It is also noteworthy that damaged, dysfunctional leaves can still harvest light but the energy creates dangerous reactive-oxygen species instead, which can bleach a leaf from the inside out (it’s essentially the same process as if too much light was applied and non-photochemical quenching/fluorscence couldn’t regulate enough anymore…). So this is the main reason why perennial plants need to get rid of old leaves (they do drop from alone) once most major content has been retranslocated.
Then there is also the spectral discrimination, a dense canopy leaves usually just green and darkred, but red and especially blue, is quickly deplete. Now there is some longgoing controversy about this but one cannot deny the effects of UV/blue light on plant phototropism and photomorphogenesis mediated by some bluelight receptors. These may also be responsible for a gene expression to increase some Cannabinoids, predominantly THC, to a tad higher levels. Maybe it is this why defoliation can help bring a better quality over sheer dry harvest mass? (We really should tell these scientists they better be focusing on the netto amount of cannabinoids and terpenes grown and not only the raw MJ to get a better comparison).
. No offence intended, just a remark … 
