There are a huge number of factors which will influence the performance of a system, and _might_ influence it in a significant way.
A great number of the factors which influence a design are so subtle as to have a maximum range of effects which are smaller than unknowns inherent in the design. In this case, it is wasted effort to focus on the subtle effects, because effort spent there will give benefits totally swamped by random noise.
A good engineer will have judgement about which design aspects need to be concentrated upon, and which can simply be ignored as too small to matter. But as a cautionary tail, sometimes issues dismissed as insignificantly small can rear up and bite one's donkey.
A good engineer will also judge some things to be potentially significant, but decide that it is cheaper to overdesign the system rather than do an exact design.
To the OP, the huge unknown is the power factor of the lamp system in question, and weather or not the 1.5KW number is power _to the lamp_, or power to the lamp system (including ballast).
Once you know the VA of each lamp system (lamp, ballast, controls, etc) then it is a simple matter of adding up the VA numbers to get the total VA consumed by a set of 12 lamps (on a single post) or a set of lamp posts.
Now, these lamps systems are going to have a tolerance on their power consumption. It is very unlikely that a 1.5KW lamp is going to consume 1500.000W of power. Similarly the VA going into the ballast will not be an exact number, but somewhere in a range of values that depend upon things like applied voltage, ambient temperature, age of the lamp, burn time, etc. So even though you can simply add up the VA of each lamp system to get a good idea of total consumption, you _must_ expect that the consumption will be some % off from your calculated value.
Similarly, the transformer will almost certainly have a tolerance on its power rating, and the power rating itself will be based upon assumptions that leave lots of room for 'fudge'.
You could start doing a detailed tolerance stackup, to make sure that the maximum _actual_ KVA of your lamp system is less than the minimum _actual_ KVA of your transformer. Or you could do the easy thing and make your design load some fraction of the transformer rating. A bunch of the guys here have suggested using a 200KVA transformer at 180KVA...but note that power companies often use transformers at well _above_ their nominal ratings.
As to the relevance of power loss in the wiring to the lamps, I believe that is something that could potentially (and counterintuitively) be a problem. In particular, some modern lamp and ballast combinations are constant power devices, consuming more current as the supply voltage goes down. The lamp continues to perform adequately even with horrible voltage drop on the supply conductors. At the same time, outdoor lighting applications often involve large distances where the cost of conductors is a major factor, and some amount of voltage drop almost needs to be designed in to optimize the economics of the system. In this case (and the OP never specified what sort of lamps were being used), voltage drop in the supply conductors could significantly increase transformer loading.
-Jon