With any transport over certain distance, you always get some goods losses and transport cost. And as it use to be, lowering the losses means increasing the cost and vice versa, lowering the cost means increasing the losses. The key is to find a way to minimize these losses with minimum cost increase (and consequently allow to save the cost elsewhere) The electricity transport is no difference, the key to minimize losses without much extra costt (overrated wiring) is to transfer the power with as minimum line loading as possible. Now with a lot of mathematics you may prove the minimum wiring losses for a given transferred power and voltage source waveform shape (sinewave with regular AC) is, when the voltage and current waveform shapes exactly match. And that happens to be when the load is a pure resistor. So the desired key to low transport losses at low distribution cost is to make all loads behaving like such resistors. In order to quantify the success in that, a concept of "power factor" was invented, defined as a measure of a ratio of how much of real power is transfered versus how much power could hypothetically be transferred with the same wiring losses on the given supply system. So the hypothetical maximum is a unity, lower value means less real power is being transferred than hypothetically possible with the given wire loading so losses. The lower power factor does not directly means lower efficiency so higher power input of the device itself, but it means higher distribution loading, so higher losses in the distribution network.
So at the end the power factor is a measure, how much the voltage and current waveforms match. Nkw the mismatch in sinewave AC systems could be of two kind: Phase shift and shape (harmonic) distortion. The first in lighting is the dominant one in mains frequency ballast systems. Because practically all high efficacy light sources need the ballast to regulate the current flow for them, they need some impedance in series. That impedance could be a resistor, but that would mean high losses (although no phase shift). Or it could be some reactive (either inductor or capacitor) component, which limits the current without the losses. But that means a phase shifting component. So when used without anything else, the phase shift gets then transferred to the input terminals, so such system (ballast plus lamp) then exhibit power factor way below one. Because this is how the basic ballasted circuit normally behaves, it is called "normal power factor". Now this phase shift could be compensated by connecting another (opposite) reactive element parallel to the mains input (mains parallel capacitor in the series choke or HX transformer ballasts, parallel inductance in the capacitive lead type ballast like an air gap under the primary side of a CWA or so) Or in case of ballasting two lamps in one fixture where one lamp uses inductive ballasting and the other uses capacitive one, so the reactive component (phase shifts) cancel each other on the mains input. Because there is no phase shift on the system input, the power factor is nearly unity, hence the "high power factor" ballasts.
With electronis ballasts the thing is more complex: As practically any other, the electronic part itself runs only on DC and not directly AC. So when intended for a regular AC mainy, such thing needs to employ a rectifier on the input, plus some means of voltage smoothening. The basic formof this is a diode bridge rectifier plus a tank capacitor for smoothening. But thisbasic system draws current only around the peaks of the mains,causing the current to beco erather spikey. And that is quite huge shape mismatch towards the voltage (nice sinewave), hence the rather low power factor of such simple circuit and the designation "Normal power factor". By the way this type of load forma another danger, mainly in symetrical star connected 3phase systems (the 230V world): When operated at full capacity with resistive load, the Neutra currents cancels out, so no current in Neutral. But when such rectifiers are the loads, theNeutral current is nearly twice the current in each phase. Because the protection systems usually sense only the phase current and the wiring uses the same gauge for all 4 conductors, the Neutral gets overheated. This is thereason, why most regulations put very strict limits on the shape distortion, compare to way less strict limits for phase shift caused power factor lowering. So to tackle this, the rectifier has to be equipped with an additional power conversion stage, whose single task is to shape the input current so it follows the voltage waveform. This circuit is then called a Power factor corrector and such device (ballast,...) high power factor input.
