The dimensional stability of paper with regards to moisture content is
obviously an important issue, and always was to judge by old adverts.
Dry wax paper is good in this respect because the wax prevents moisture
take-up and gives very stable dimensions in the sauna-to-furnace
conditions some users subject it to.

Storage conditions are fairly obvious: the paper needs to stabilise
and be cut in the conditions that the roll will be used in.  Then even
fairly hygroscopic paper will be the right width in normal conditions.
Aeolian (in London, anyway) kept ("matured") paper in special
controlled storage for several months before cutting it, to ensure best
results.  Given their production volumes, the scale of the storage was
vast, showing how important they considered this to be.

Many commercial printing web sites describe paper storage requirements,
normally suggesting the paper is kept in its sealed delivery wrappers
until use.  They especially say that laser printer paper should be kept
well wrapped, because it is delivered especially dry (about 20%
humidity) to prevent it curling in the fuser unit.  Inkjet paper is
delivered at more like 40% R.H. and, as with so many modern high-tech
products, users almost certainly don't understand just how sophisticated
the materials specification is.  I'm not quite sure what questions need
to be asked of a paper vendor about these things, but I suspect that all
this type of printing paper has to be fairly stable dimensionally to
avoid permanently jammed machines.  Wrapping paper presumably has a lesser
requirement in this regard unless explicitly waterproofed.

Roll brands that are consistently underwidth in normal conditions
almost certainly indicate damp storage conditions (as well as poor
paper choice), leading the roll to shrink after use.  The 1950s-1970s
British Artona brand is a particularly gross offender in this regard,
but other post-war makes do consistently show the same trend in lesser
degree.  Maintaining the paper store and perforator room at a middling
humidity range would seem to be sensible, depending on how much leeway
the chosen paper gives.

Julian Dyer