To underscore my point (post below), why do you say "broken" rather than "healed" or "saved?"

Employee evaluation is like science in the sense that how well it's done affects all employees through it's effect on the success of the company. It's not like science because it isn't under public scrutiny and isn't expected to be reproduced. It's also not like science in that there is no hypothesis being tested against the measurements.

Evaluation by objective criteria is like the rule of law in that it forces the employer to describe goals in a way that the employee can guide their actions toward (whether the expressed goals turn out helpful or not). We think the rule of law is good for [edit] certain general commands in [/edit] a command economy. Coase said that firms have economic reasons to be like command economies. Is it a categorical switch whether it's appropriate to act like a command economy or a market in all ways? Does the transaction-cost model mean that small companies should be more commandy and large companies should be more markety? (If so that would suggest a reason why bosses in larger companies bend the rules more.) But somehow rules for a year of an employee's behavior don't seem to me necessarily right for small companies either.

Overall, I guess companies need to *signal* objectivity and fairness, but in their actual behavior negotiate and communicate hard-to-articulate, fluctuating values and make sharp-edged, hard-to-justify decisions, i.e. act political and unfair.

But that's a guess, my question is whether politics or objectivity would be more "broken" in this context. [edit] Also, how political and politically-driven is the process of setting up the on-the-surface objective criteria and review processes? What needs would you expect the set-up process to serve and what needs does it actually serve? [/edit]

And what has been done or what would you do to find out?

Expand full comment

Are there ways to evaluate whether "objective" (i.e., formal rule-driven) performance evaluations are a more accurate, or (separate question) less political, way to evaluate the value of an employee to an organization? Otherwise this seems a one-sided comparison. Is politics correlated with accuracy, correlated with inaccuracy, or orthogonal? Are the formal review criteria used by large organizations more or less accurate than those used in small organizations? More or less political in large organizations? Even the notion of accuracy in predicting the value of an employee seems suspect to me, since it's like looking at a single price within a market and asking whether it's "correct".

When I ask how political formal criteria are, I don't mean whether they specify objective measures. I mean whether the measures selected are selected for more or less political reasons or by more or less political processes.

Expand full comment

Does the concept of a 'misery frontier' resonate with anyone else. Diminishing marginal utility of a profession combined with high job switching costs. "Do what you love" --> "Do what you hate the least"

Expand full comment

News at 11? Exactly what you would expect, really.

Expand full comment

Basketball affords natural experiments in relative player skill when an injury swaps 1 player for another in an otherwise fixed starting 5. You can get several hundred minute sample sizes controlled for those swapped players. The resulting +/- stats are the cleanest performance eval that exists in team sports.

Several times in recent years the new player has boosted +/- by, say, 6-8 points per 48 minutes played (huge margin), and when the old player recovers, he typically regains his starting job.

In one instance the recovered player (declining superstar Ray Allen) didn't get his job back; he was replaced by a young guy, Avery Bradley.

When the season ended Allen bolted from the Celtics to their prime rival Miami for less money (and still a backup role!). Celtics culture suffered, seemingly strong bonds were severed nastily, and the core of the Celtics broke-up permanently one year later.

All this despite the existence of an eval far simpler and more robust than exists in a corporate situation, and in an organization small enough so that all info is common.

Expand full comment