What Might James Shields Earn In Free Agency?

We heard earlier today that the Royals have no intentions of making a run at extending ace hurler James Shields. That would seem to confirm what has long been expected: namely, barring a surprising change of circumstances, Shields will hit the market next year looking to sell his services for age-33 and beyond.

As Andy McCullough of the Kansas City Star noted earlier today, Shields and his camp have rejected rumors that he is seeking to land a deal in the realm of Zack Greinke's six-year, $147MM pact. (Note that Greinke did not require draft compensation to sign, as Shields almost certainly will unless he too is dealt mid-season.) 

But, as McCullough notes, Shields compares well to Greinke — indeed, he compares well to just about anyone not named Clayton Kershaw — in terms of recent production. He has pitched at least 200 innings (and often quite a bit more) for each of the last seven seasons, and has logged ERA totals of between 2.82 and 3.52 over the last three. Shields has been among the 20 or so most valuable pitchers since he cracked the league, and has looked more like a top-ten guy more recently.

So, can Shields earn Greinke money? Though his numbers make that contract look like a solid comp, another major factor speaks firmly against it: namely, age. At the time Greinke inked with the Dodgers, he was just over 29 years old, and threw at that age for the entire season in the first year of his deal. Shields will be four years older when he faces the open market for the first time.

This series of observations led MLBTR's Tim Dierkes to suggest that, perhaps, it would be worthwhile to look at the historical results of top-level, slightly older hurlers. The results of that research, and my discussions on the topic with both Tim and fellow MLBTR writer Steve Adams, make up what you'll read below. 

Unsurprisingly, it turns out that slightly older pitchers have not readily landed huge contracts. That makes sense, for at least two major reasons. First is simply the factor of selection bias. Most really good pitchers accrue service time early in their careers and thus hit the open market slightly earlier than will Shields, who did not crack the bigs until age 24 and then gave up two years of free agency in an early-career extension. Likewise, as age increases, the likelihood of significant injury goes up, further attriting the possibly worthy arms. The second reason is even easier: older pitchers have less near-peak years left to sell.

Let's take a look at all pitchers to have signed deals with at least a $60MM guarantee at a $15MM or greater average annual value — after having turned 31 years of age. (The last column represents the player's age at Opening Day of the first season of their deal).

Among these potential comps, we can reject several out of hand. First, Brown and Lowe were both significantly older and signed deals in quite a different market. (Frankly, I only kept them on the chart to show how remarkable their contracts were.) Wainwright signed an under-market extension, while Halladay was also probably on a different performance/perception level and signed his own deal under somewhat odd circumstances.

That leaves us with four potential free agent comps, all of whom are near in age to where Shields will be when he hits the market. All signed five-year deals in early-to-mid-March of their free agent year. (If Shields signs before his birthday on December 20th, he'd be just shy of 33 years of age, making him slightly older than the other four were.)

How does Shields stack up to that quartet? Let's look at both three-year averages and walk-year performances to get an idea, and look at the salary numbers through the lens of inflation (present dollar value estimated via US Inflation Calculator). I have used ERA and fWAR not to suggest that those are the proper means by which to value pitchers, but to represent two sides of the overall picture: the former captures pure results, while the latter incorporates FIP as its baseline and thereby captures some of the underlying talent and rate-based results that teams surely examine closely when signing deals of this magnitude.

Needless to say, Lee paced the grouping — at least in the eyes of advanced metrics and in dollars achieved. He still fell well short of Greinke's guarantee, though of course he inked his deal a few years back and reportedly could have signed for seven years and $148MM. 

As for Shields, his current three-year and last-year ERA both stand at 3.15. He has been worth 12.9 fWAR over the last three seasons, and earned 4.5 wins in 2013. Assuming he hits the market with approximately the same profile — i.e., he logs another 200+ innings of pitching in the low-to-mid-3.00 ERA range and has underlying metrics to support a 4-5 fWAR season — where might he land?

The answer, probably, is not in the realm of Greinke's contract (or the commensurate offer that Lee could have signed). Even accepting that baseball inflation has outpaced that of the general economy — or, at least, understanding that MLB salary levels are subject to wide variation due to the league's comparatively tiny player market, small sample of transactions, and range of non-market-based influences — the Wilson-Lackey-Burnett line of arms equate at most to a five-year, $100MM present-day value.

Importantly, Dierkes noted, those three were clearly the best available arms in their free agent years. While it is still conceivable that Shields will ultimately be the most attractive starter available when the market opens, odds are he will slot in behind Max Scherzer and possibly Jon Lester. (Of course, if Scherzer repeats his 2013 season, he might be on another plane altogether.) And if Justin Masterson hits the market, he might be more appealing to some clubs because of his significantly younger age. Regardless, unless each of those arms is extended off the market or takes a large value hit, Shields will not be the sole target for clubs looking to add an impact starter.

Moreover, as I recently noted in discussing extensions, teams have evinced an increasing willingness to value age. That attention to the aging curve probably also implies a corresponding unwillingness to roll the dice on continued production on its downside. Still-excellent pitchers Hiroki Kuroda (age 39) and A.J. Burnett (age 37) just signed one-year deals that will pay them $16MM. Can Shields really beat that AAV on a five-year pact that would buy out his age 33-37 campaigns, especially if there are alternative arms on the market? As Adams noted, Shields' slightly more advanced age means that a fifth year would push his guarantee into somewhat uncharted territory as compared with his most direct comps. It is worth wondering whether clubs would be willing to guarantee another year at such an elevated rate.

In the aggregate, it looks as though Shields could reasonably expect to top out at five years, with a $20MM AAV a seeming longshot at that term. On a four-year deal, Shields might land in the $75MM to $80MM range, perhaps with a vesting/club option of some kind attached on the end. But barring a massive jolt to the market that leaves Shields as the unquestioned prize with multiple suitors, it is somewhat difficult to imagine him reaching the $100MM threshold — if he is able to get a guaranteed fifth year at all.

One possible twist would be the inclusion of an opt-out clause. As Dierkes notes on Twitter, Lee actually had one offer that included an opt-out before he signed on with the Phillies. In his excellent piece on the use of opt-out clauses, Dierkes wrote that such provisions may not be that onerous to the team since, if exercised, the club is able to re-assess whether to take on further obligations (and, I would note, can do so with the advantage of insider knowledge on that pitcher). If clubs are unwilling to guarantee as much cash as Shields hopes for, he could potentially press instead for a deal that includes an opt-out, allowing him to re-enter the market if he carries his production into his mid-30's. And at that point, perhaps the Lowe contract would take on increasing relevance.