The concept of parallel confocal microscopy with aperture de-correlation has been developed in order to reduce acquisition time and to increase light efficiency over conventional scanning confocal microscopy. It consists in simultaneously observing numerous confocal volumes of a specimen, in lieu of scanning one confocal spot over the region of interest. The cross-talk between neighbouring parallel confocal systems is eliminated by capturing images using a sequence of linearly independent illumination and detection patterns which permit to de-correlate the confocal from the non-confocal image. In this paper, we compare the relative efficiency of parallel and conventional confocal microscopy, using a simple analytical approach, and show that this depends strongly upon the illumination configuration. In particular it is shown that despite its drawbacks in other applications, conventional confocal microscopy still is more efficient than parallel confocal microscopy when the dose, i.e. the total energy onto the sample, must be limited. This may represent a limitation to the advent of parallel confocal microscopy for the imaging of photo-sensitive tissues.