Obligatory disclaimer: This is merely speculation. The Rays haven't said they are actively shopping Matt Garza or any of these players, and it's entirely possible that we do something like a 6-man rotation or moving one of these to the bullpen in case of injury.
With six starters and five rotation spots, it doesn't make sense for the Rays to keep all of their starting pitchers. Similar to Edwin Jackson, Jason Hammel, and Scott Kazmir, we can expect one of our current pitchers to be on their way out in exchange for prospects. Wade Davis, Jeff Niemann and Jeremy Hellickson will be making the minimum, and Price will also be making a tiny amount of money (for a baseballer). That leaves Matt Garza or James Shields.
As the title would suggest, I believe Garza will be the odd one out between those two. Shields's contract has loads of flexibility and is team friendly, and additionally there's reason (his xFIP) to believe that his value is artificially low right now. Garza, on the other hand, appears to have artificially high value for the exact opposite reason: his xFIP is much lower higher than his ERA. The decision of which to trade, if indeed a trade would occur, is a simple economics rule. You buy low and sell high; right now Garza is at a high and Shields is at a low.
The obvious question here is, if Garza does get traded, what can be expected in return?
For this we turn to Sky Kalkman's beautiful trade value calculator. In examining Garza's trade value, the two things that must be considered are Garza's performance and his salary.
We'll begin with a salary projection because that's way easier. Players get raises in arbitration and Matt has three years of team control left. History says that players generally get 40% of their value in arb1, 60% in arb2, and 80% in arb3. Garza's already established a baseline of 3.35 million since he was a super 2 this past off-season, and so we can expect Garza to enjoy gradual raises going forward. Although these numbers are largely estimates, Garza will probably make in the neighborhood of 5 million, 6.5 million, and 8 million the next three years.
Predicting Garza's production is a little more difficult. The innings pitched total is fairly predictable; Garza's pitched 203 and 204 innings the last year, and so we can probably expect him to pitch roughly that many innings again over the course of a contract. CHONE projects that Garza has a true talent level of a 4.23 nERA and with this many innings, Garza would be worth 3.2 wins.
The more difficult thing from here is projecting Garza's aging. MGL's study here suggests that pitchers of Garza's age tend to decay at a rate of .2 runs of ERA a year. This converts to about .45 wins a year. Plugging in Garza's expected win values of 3.2, 2.8 and 2.3, we can calculate Matt's surplus value using the trade value calculator above. It totals out to be about ~$21 million.
Given that a trade for Garza will probably be centered around hitting in return, we can turn to Victor Wang's prospect evaluation research to determine to see what that $21 million converts to. Wang's work suggests that the amount of surplus value that Garza presents is equal to a top 11-50 hitting prospect, or perhaps an A- hitting prospect. Additionally, Garza could command a bevy of smaller prospects in exchange, something like 40 grade C hitting prospects above the age of 23? In all seriousness though, packages are generally not so neat as the Edwin Jackson package, and more like the Scott Kazmir package.
That being said, Garza's value might appear higher than this because of a 15-10 record and a 3.91 ERA. Additionally, the plan might be to keep Garza until mid-season when we can enjoy half a season of his performance, get a better idea of our contention chances, and also have teams far more likely to make the mistake of overpaying. It's difficult to say what exactly will happen or can happen with a trade, and there's so many possibilities, but if Matt Garza is on the way out, a top 50 or so hitting prospect is roughly an equal value in exchange.