Every one I've measured (both old and use) used close to 5W.
With what? And if they all measured the same, did the bulbs all state
the same consumed wattage among the ones that stated that? If all with
wattage statements stated 7 watts and you have a wattmeter reading them
all as 5 watts, especially if the wattmeter is good for a thousand watts
or more, I suspect the explanation is that the wattmeter reads loads in
the ballpark of 7 watts low by 2 watts.
I suppose the made them like that once. People always want things that
save energy.
I don't see much in claims of "new improved more-energy-efficient
version" or anything similar in the area of colored C7's and less than
that in the area of C9's. I see that more for series strings and
especially LEDs.
I just checked that on one LED string, and got about .6
Maybe not much out of the ballpark - I figure that if the LEDs have any
conduction at all half the time then the power factor should be close to
..7. The relationship is roughly a square root one - power factor in a
resistive-current-limited LED string without capacitors is roughly square
root of fraction of a cycle that has current flowing. Accordingly, .6
translates to conducting 36% of the time - ROUGH BALLPARK! I consider
this roughly true.
However, I would rather allow for .5 or maybe .4 power factor when it
comes to not blowing fuses or overheating wires.
And I forgot about the loading not being a pure sine wave.
Main reason for RMS amps to exceed ratio of volts to watts in most LED
holiday light strings. Harmonic content in the current waveform does
mainly heat wiring and fuse elements and little else when the voltage
waveform is a sinewave or close to a sinewave.