With another baseball season underway, I've been reflecting on the revolution within the game. Since Michael Lewis' Moneyball hit the shelves, baseball analytics departments have become the "thing" in front offices all around the game. Sabermetrics [1] has taken hold. New statistics have replaced "old" ideas like batting average and ERA. Newfangled things like OPS and OPS+, FIP and ERA+ are bantered about on message boards and SportsCenter alike. [2] And even someone with a liberal arts background (like me) can see what's going on. But this blog isn't about baseball. Not really. It's just the starting point. And the real place to start is the statistic that fascinates me most:
Wins Above Replacement, or WAR. It's the closest thing baseball has to a Grand Unified Theory of Everything. It breaks every event that unfolds during a baseball game, offensively or defensively, into a unit. And that unit is part of the thing that matters more than anything else in sports. Wins. Which means that for the first time, we can actually tell how much a player is worth. Or, rather, how many wins a player has been worth -- especially when compared to a baseline -- and then predict how many wins that player will be worth in the future. [3]
Now: What if we could do that with job creation?
What I'm talking about is Jobs Above Replacement. Call it JAR for short. A way to measure how many jobs, above a baseline, any particular investment or initiative might be expected to produce. A Grand Unified Theory of What-Actually-Works-in-Economic-Development. Instead of guessing at a dollars-per-job number in advance, or waiting to retrospectively report (and then argue over what's included in a measurement), a formula that can help guide us, public or private, to the things that create the most jobs. And not just any jobs. Good jobs. Because like WAR, the formula behind JAR would be weighted too.
A smooth defensive left fielder with a cannon for an arm isn't worth the same as a sweet-swinging second baseman with a howitzer for a bat (they're pretty close to being the inverse of each other). So why should we be counting a $10 per hour job without benefits the same way as a salaried position with benefits, training and advancement opportunities? We shouldn't. They're not the same now, and their effect on the future (individually and collectively) isn't the same either. We need a better way to measure what matters.
To get started creating this new statistic, we have to figure out what we're actually trying to measure. Ideally, we'd be able to determine how many jobs are expected to be created absent whatever investment or intervention we're analyzing. Said differently, we're talking about the jobs we'd expect to be created if (name your geography/company/institution) just kept doing what it's already doing. This would represent the "replacement level" part of the JAR equation.
There are any number of ways to develop a formula that would result in a JAR figure. It will take time, and experimentation, to get it right. But here's where I think we might start:
1. Jobs Created: With weighted adjustments for the type of job. Full-time, part time. One could imagine an "expected" number, with an adjustment for "actual" when evaluating outcomes.
2. Good Jobs Adjustment: Yes, this will be contentious, but benefits matter. So do career advancement opportunities. And living wages. [4]
3. Jobs Saved: A job that isn't lost, because of a particular investment, might be worth the same as a job created. In baseball, a defensive run saved is worth the same as an offensive run created. [5] In energy efficiency, a kilowatt of electricity saved is worth the same as a kilowatt of electricity generated. So we need to count jobs saved. And we need to guard against "false job anxiety." Those are my words for dire pronouncements of impending job cuts, designed to force some type of investment or concession. In other words, we need a way to count jobs saved that are truly saved, and that's going to be tricky.
4. Velocity: How fast are the jobs going to be created? A job this year is worth more than a job next year. A dollar in the bank this year is worth more than a dollar in the bank next year.
5. Multiplier Effect: Is a particular job-creating investment likely to lead to follow-on investment? Indirect jobs should count. A cleanup hitter may not be able to control who's on base when they're at bat, but the extra runs they knock in sure do count.
6. Cost Per Job: Even if it's hard to measure in advance, the projected cost per job is important. If the expected cost is $100,000 per job created, that's not as valuable as $100,000 investment that's expected to create10 jobs. Just like Troy Tulowitzki's projected WAR of 4.5, at a salary of $20 million, aren't worth the same as Dustin Pedroia's 4.5 at $12.5 million. [6] Value matters here.
7. The Baseline: As outlined above, all of this has to be divided by what we'd expect in job creation absent the investment being considered.
Making JAR work smoothly is going to take time. It will probably require some statistically savvy baseball fans with economics backgrounds to dig in. [7] But there's precedent here too. When Bill James wrote his first Baseball Prospectus, the modern world of sabermetrics didn't yet exist. The Prospectus begat the Baseball Handbook, which evolved into the world of advanced statistical analysis we see across the sports world today. [8]
The lesson I draw from all of this is: once you decide what it matters to measure, you can get to the matter of measuring it. Jobs matter, so let's figure out the measuring.
[1]: Sabermetrics is, essentially, advanced statistical analysis applied to baseball.
[2]: On-base Plus Slugging (OPS), On-base Plus Slugging Plus (OPS+), Adjusted Earned Run Average (ERA+), Fielding Independent Pitching (FIP)
[3]: A (very simplified) way of thinking about it: Every event on a baseball field contributes to some percentage of a win. A home run is worth X. A stolen base is worth Y. An error subtracts Z. Put it all together, once the formula is determined, and you get WAR.
[4]: There are a number of groups that work on measuring (and creating) good jobs. Fund Good Jobs is one of them. [Disclosure: Fund Good Jobs is an Impact client of Womble Carlyle]
[6]: There are multiple ways to calculate this. I used Steamer.
[7]: Yes. I just described the front offices of the Red Sox and Dodgers, among others. Or maybe Nate Silver can give it a shot?
[8]: Bill James is considered the father of the modern statistical movement in baseball. A legion of dedicated baseball fans has collectively researched and contributed to new evolutions in statistics, including the membership of the Society for American Baseball Research (SABR).