Plus at the end of the year when the decision to sign Gibson comes up, the could look at innings pitched and say we can't give that much money to a pitcher that averages X innings or pitched Y innings, or ERA is Z when he might not have been given a chance to increase/decrease those numbers.
I absolutely think that money is a big reason for lessening starter workload. The "opener" is a way to game the arbitration system - it should be no surprise that the Rays introduced the opener, that the A's used one in the playoffs, and that the Twins were one of the teams that used the opener most prominently last September; they're all teams that have historically pinched pennies. I read an article about the Rays' pitcher (Yarbrough I think) who vultured a bunch of wins due to the fortune of being the ostensible "long man" after the opener, and he had an arbitration battle over the winter because his stat line was all goofy - a lot of wins, but very few starts and not enough innings to qualify as a starter.
It's frustrating how seriously it's being taken as a strategy. It's less of an analytics innovation than it is a way to keep your payroll down. The more 5 innings (two trips through the lineup) becomes the norm for starters, the more likely that a full season for a starter will approach closer to 150 innings (under the qualifying threshold), thus muddying the arbitration game and ultimately keeping salaries down. And we wonder why Kuechel is still sitting at home? Why pay $18M for one year when you can just have three bullpen pitchers combine for his workload for a tenth of that contract? And as a bonus, by doing so you can be viewed as "smart" and "innovative" by the spreadsheet crowd? Cal Griffith was just born too early.