If you had a 5 volt battery rather than a 3.7 volt...the 5 volt would be more "efficient". However, if you take an e-cig that has a single 3.7 volt battery and make it VV WITH A BOOSTER you have to account for losses due to having to boost the voltage.
So...when this vv stuff started to be mainstream, I posed a question to the "mind trust" in the modder's area about single-battery vv implementations like the twist. What would be more efficient at the same wattage, high volts or low volts? We never got a good theoretical answer due to implementation details for each device.
Reports seem to indicate that people using lower voltages, on a twist for example, seem to get longer life. So...bucking the voltage (reducing it) seems more efficient than boosting it even when accounting for the theoretical gains obtained using higher voltage. This is PROBABLY true for the kick as well. Now, for a stacked battery mod where you only buck the voltage, higher voltage is often more efficient. Particularly if using a linear regulator (that dissipates excess voltage as heat).
That's pretty much what I have come to realize- that batt life has just as much to do with the device you use, and how it functions, as it does with the theoretical power consumption based on output voltage and coil resistance. IRL I find, for instance, that a V1 Lavatube (L-Rider) lasts twice as long- on the exact same battery- as a VV Chrome Tube (YJ version). Technically, the mods both attempt to do the same thing- in my case, lowering the voltage to 3v (supposedly), but in the case of the Chrome Tube, the PWM-ed waveform is exaggerated more-so than on the Lavatube, leading me to believe it is causing the batts to drain quicker. When comparing other devices with different batt configurations, etc, the differences start to compound on each other even more.