Last week I took a crack at projecting Josh Donaldson over the time horizon of a possible contract extension. The logical follow-up from that is what an extension could look like, and whether the incentives line up for both sides to find common ground
Let's start with the big picture. At a strategic level, the front office has positioned the Jays with a high degree of future flexibility whatever happens in 2018. If the season goes poorly, a rebuild/reset of some sort becomes inevitable. But with very modest salary commitments beyond 2018, they're not going to spend the first year or two just digging out. On the other hand, if 2018 goes well and the Jays make the playoffs or are at least highly competitive (high 80s win total), there's a real possibility of threading the needle between the present and the future. At the very least, you're very likely looking at contending in 2019, and then you're close to the Vlad/Bo era.
Where does Donaldson fit into this? In the event of a rebuild, it wouldn't be a complete waste to have extended him, since he should still be a productive player. But realistically, the bulk of Donaldson's value is the near term wins he can provide a contending team. Conversely, if the Jays have a successful 2018, then the combination of strong revenue/high payrolls and lack of long term commitments would leave them well-positioned to pay going market rates to keep him.
In other words, it only really makes sense for the Jays to do an extension if they can secure a discount compared to what the expected market will be this winter. From last week, my best estimate would be that Donaldson will produce about 20 WAR from 2019 onwards. Applying a market rate of $9- to $10-million per win, and you get something like a $180-200 million commitment (or $200-225 million including 2018 at his current salary). .
That leads into the other major factor at play, the value of already having Donaldson under contract for 2018 and the optionality it creates. In decision science, there's a concept called the expected value of perfect information. Basically, it's the value that could be realized by having a crystal ball and knowing what will happen in the future to make decisions now accordingly. It's not perfectly analogous to the situation with Donaldson, but the ability to see how 2018 plays out and how that impacts expected future value (simple Bayesian updating) is very valuable.
We often refer to aging curves for players, but these are just expectations, the average decline from year to year of a large group of players. Some players do follow these career paths, but at an individual level what tends to be more common is that at some point they experience rapid decline. Essentially, their skills collapse at some point, often more or less out of nowhere. History is littered with players who were highly productive in one season, and a couple years later were out of baseball or essentially done. Jose Bautista went from 148 wrc+ to 80 in the space of two years.
Exactly how valuable could the added info from letting 2018 play out be? Let's go back to my comparison set of the eight elite third basemen. In their age 32 seasons, three of them had montser seasons (A-Rod, Schmidt and Brett). Over the remainder of their careers, they average 3,800 PA, a 129 wrc+, and just under 23 WAR. The other five, who all had good-to-very good seasons (no total collapses), averaged 3,075 PA, 118 wrc+, and just under 19 WAR. If 2018 can give us information about the future on the order of a 4 WAR difference, that's upwards of $30- to $40-million difference.
In particular, the career arcs of several of the comparison players are instructive. Eddie Mathews was consistently a force in his prime, with his worst season after his rookie year being 5.7 WAR and 120 wrc+. In 1964, at age 32, he slipped to 117 wrc+ and never recovered to his previous heights as he played just three more full seasons. Likewise, Scott Rolen had a down year at age 32 and while he was a solid contributor for parts of five more seasons, never had another 5 WAR season (six previously).
There's even more cautionary tales in some of the players I screened out because they had already declined prior to age-32. Ron Santo posted seven straight 6+ WAR seasons from 1963-69. His age-30 season in 1970 was very good, but a nonetheless a step down. He had just three more productive seasons and was done before age 35. Sal Bando likewise started declining around age-30, and did not have any productive seasons after age-35. David Wright was coming off a couple of monster seasons on his 31st birthday. That was followed by a significant down year, and then his career was/is effectively over.
The simple reality is that even for elite players with an unblemished track record, when they get to Donaldson's age there's something like a 20-30% chance they only have a handful of productive seasons left. This possibility is especially dangerous for the Jays if they extend him now, since in a rebuild scenario his days as a regular would be winding up as competitive window starting in 2020 or more likely 2021 opens.
Of these two factors, the bigger obstacle to an extension is the actually the uncertainty about the future. The possibility of significant decline in 2018 is at least partially offset by the chance of Donaldson posting another huge season and the positive revision that would have on future output. The risks are to the downside for players in their 30s but there's at least a potential offset.
By my rough back of the envelope figuring, Donaldson's age-33 onwards value to a team rebuilding in the short run is about 25% lower than for a contending team (something like $150-million ompared to $200-million). Right now, it's hard to say which of the two will be the case in 2019, and any offer any to reflect some weighted average of the two, and thus a team specific discount. It's simply the prudent managerial course of action, but not something Donaldson should be inclined to essetially pay for. This wedge will make it even harder to find a deal that makes sense for both sides.