Of course, in my games unit size is the primary cost for units, about 50%.
I think I like this.
While 'elite' units will obviously be payed more than regulars ... I think that just because their armor costs 10x the amount of someone else, doesn't mean that their wages will be 10x as well (although I guess you could consider the costs to upkeep their equipment?)
While the US government pays millions of dollars for the production and maintenance for a single jet, the pilot is only paid about 60,000 dollars or so (depending on rank) ... which is only about double the wage of the average soldier (iirc ... either double or triple)
meanwhile his "equipment" costs millions of dollars while the logistics cost for a single rifleman is likely only a few thousand (an extra seat on the plane, an extra set of weapons, extra ammunition)
While you could get into continued costs of food/barracks/etc, this is the same for both and thus can be left out of the equation.
This is a 1000x increase in equipment cost with only a 2x or 3x increase on wages.
Meanwhile a regiment of 1000 riflemen would cost about 20 million or 30 million dollars, while the single jet pilot still only costs about 60 thousand (in wages).
Total costs?
Regular Jet: 20 million in equipment, 60,000 in wages
F-22 Raptor: 150 million in equipment, 100,000+ in wages (high rank)
1000 riflemen: 2 million in equipment, 25 million in wages
--> This doesn't factor in the 'maintenance' for all that equipment though, or the logistical costs of all soldiers involved.
The maintenance costs for the jets themselves would likely be astronomical ...
yet the logistical costs for each soldier would be X for the jet pilot and 1000X for the riflemen.
However, in a medieval setting I would think that any armor piece would require extensive care by either its user or by a local blacksmith as part of the retinue.
I think the wage should 'in part' be a part of that maintenance of equipment, yet should primarily be focused on unit numbers.
I think the 50/50 that Sean suggests sounds fair ... (although I don't see what equation he used).