The job base cost is based on the value of the item produced, with jobcost = itemvalue* sqrt(fraction of global job hours) thing. This is modified by a crapload of things; stations in the system, quality of the stations, number of stations, FW status. Now, two factors are relevant:

Basecost: How will the basecost differ between highsec where everyone has access to all factories vs. nullsec where factory access can be controled?
This depends on how the distribution of working hours will stabilize. In the snapshot, the basecost in a "semi-active" highsec system was 5%, while the worst nullsec system started at 4%. Further statistical data is rare; however, I'd expect the average nullsec basecost to end up anywhere between 50% and 10% of the highsec basecost. Assume 50% and the math pans out friendlier than in my initial example, assume 10% and it ends up worse.

Modifiers: By how much can the base cost be reduced by multipliers, including the mentioned NPC tax?
Building in an amarr factory outpost brings the same jobcost reduction as building in the "best of the best" highsec system Nonni, (.5 times costmultiplier for the amarr outpost, .48 (nonni) + 1.1 (npc tax) = .52). The salespitch example ends up with a 0.825 multiplier (.75 system * 1.1 NPC tax).

Summing effects up
Summing nullsec effects and comparing to the "average example", nullsec pays between 30% and 6% of the highsec job costs. The highsec costs for the salespitch example ends up at 3.8% of the itemvalue (7.6mio for a 200mio abadon). For the sake of the argument I'll stick with that, though I'd expect that value to end up closer to 5% due to the resulting new metagame. Nullsec can save between 70% and 94% of that, resulting in a cost difference of 2.5%-3.5% of the itemvalue.

Approximate itemvalue = (1+profit%) * mat value and for items with a profitmargin of 20%-50% you end up with additional costs of 3-5% mat value. This cost difference scales with the profit margin of the items produced.

The most important thing in this new system is how the basecost distribution ends up; that on its own is something I'd see like to see in action without the various other changes thrown in. I really doubt anyone, including CCP knows how the meta will in up in that regard, so adding even more new stuff (e.g. Teams) on top of that seems unreasonably risky to me.