Manager Promotion: Why Eager Candidates Make Worse Managers

Manager promotion just got an empirical headache. A Quarterly Journal of Economics paper just landed with an uncomfortable finding: the people most eager to step into management roles tend to be the worst at the job once they get there. HR Dive picked up the result this morning and paired it with a Dayforce data point that hits the same nerve from the other end. Only 42% of front-line workers in 2025 felt their leaders actually understood the problems they faced day to day. In 2024 that number was 62%. Eighty-nine percent said unresolved shift-level issues hurt their well-being. Seventy-one percent had thought about quitting.

Two different studies, but they connect at a single point. We pick managers using a signal, eagerness to step up, that has weak correlation with the skills that make a manager useful. Then we end up with leaders who don't read the room they're leading, and twenty points of trust evaporate over twenty-four months.

The QJE researchers ran their experiment cleanly. They had a population of workers, looked at who self-nominated for managerial roles, and compared the team output of those self-nominated managers against managers chosen using economic decision-making and fluid-intelligence signals. The structured-selection group produced materially better teams. The eager self-promoters produced worse ones than random selection would have. In other words, self-nomination is not just a weak signal, it's actively negative.

This shouldn't be a huge surprise if you've spent any time inside a mid-size company. The loudest candidate for a manager role is usually the one who's been mentally rehearsing the case for promotion every Sunday. They're confident, they're prepared, they have a list of accomplishments at the ready. None of that maps to whether they can absorb a half-baked complaint from a junior, frame it back to them clearly, and turn it into a useful next step. The skills the QJE paper found correlate with real managerial performance, decision-making under constraint, fluid intelligence, social inference, show up in entirely different artifacts.

Why most promotion processes filter for the wrong signal

The default process at a 100-to-500-person company looks like this. HR opens a promotion window. Employees who want a manager track raise their hand. They write a memo or pitch their case in a calibration meeting. Their direct manager either backs them or doesn't. A senior leader signs off. Done.

Every step in that chain rewards eagerness and rewards visibility. The quiet senior IC who would be a great line manager, but doesn't lobby for it because she's busy doing the work, never enters the funnel. The mid-level engineer three months into rehearsing his pitch, whose pitch is convincing, sails through.

This pipeline also has the property that nobody upstream sees how the candidate actually behaves in the things that predict management performance. Does she listen well in a peer 1-on-1? Does he handle a junior asking the same question three days in a row without sighing? Does she pull a hesitant teammate into a discussion instead of dominating it? These are observable behaviors, and they show up consistently in peer interactions. They just don't show up in a self-pitched memo.

The structured signal: competency matrix plus 360

The fix isn't a new framework or a six-month consultation. It's combining two artifacts most HR stacks already have but rarely use together for promotion decisions.

The first is a competency matrix tied to grades. Not the generic kind that says "leadership: meets / exceeds." The specific kind, where the manager-track competencies are written as observable behaviors at each grade level. "Gives clear, kind feedback in 1-on-1s without softening to the point of confusion." "Delegates a task with enough context that the report can finish it without re-checking." "Absorbs a disagreement from a peer without escalating to a senior." Each one has a defined evidence pattern. Each one shows up in artifacts the team has already produced.

The second is 360 feedback, used differently than the annual ritual most companies run. For promotion decisions, you run a targeted 360 where the respondents are explicitly the people who would observe each competency. Peer feedback on collaboration. Direct-report feedback from people the candidate has informally mentored, on listening and delegation. Cross-functional feedback on decision-making under ambiguity. The instrument is the same one you'd use for development cycles. The population, the scale anchors, and the analysis lens are tuned to promotion.

When you combine them, the promotion case stops being "this is what the candidate says she can do" and starts being "this is what fifteen people who have worked with her have observed across these specific competencies, at this density of evidence."

That's a defensible signal. It's also one that doesn't preferentially reward the candidates who raise their hand loudest.

What this looks like operationally

Three practical shifts.

Shift one. Replace self-nomination with manager-nomination, then sanity-check with peer-nomination. Self-nomination stays in as a tie-breaker, not the gate. Managers nominate candidates they believe ready, peers add or remove names blindly, and the candidate pool that emerges is broader and quieter than the one self-selection produces.

Shift two. Run the 360 against the manager-track competency matrix, not the generic instrument. Same software, different question set. The questions are anchored to observable behaviors at the target grade level, and the scale is concrete enough that "yes I've seen this" and "no I haven't" mean the same thing across respondents.

Shift three. Separate manager and IC tracks at the grade level. This is the structural piece. If your competency matrix bundles "Senior" as one thing and lets people slide laterally into manager roles at the same grade, you've already lost the signal. Treat the manager track as a different competency profile starting two grades earlier, so the development conversation happens in time to be useful.

None of these are new ideas. The QJE finding doesn't introduce them. What it adds is empirical confidence that the default, promote the person who pitched it well, produces worse teams than the structured alternative. That's a useful piece of evidence to bring into a conversation with a founder who instinctively prefers the self-starter narrative.

Where DTPulse fits

We didn't build the Grades and Competency Matrix module around this specific research, but the QJE result lines up cleanly with how we set it up. Each grade level is tied to behaviors. The 360 instrument lets you pick which competencies to assess this cycle, so you can run a promotion-track 360 with the same software you use for development cycles, against the manager-track grade. Manager and IC tracks are separated by default in the seed library, adjustable but separated.

What that gets you is a promotion meeting where the case is "here's the candidate, here are the seventeen pieces of evidence across these four competencies from these eleven respondents, here's the gap, here's the development plan if she's not ready yet." A structured promotion meeting takes about the same time as the old self-pitch one. It just feels less performative and produces measurably different decisions.

What to do this quarter

If you have the next promotion window scheduled in the second half of 2026, the easy first step is one change to your existing process. Don't switch the whole pipeline. Take your current 360 instrument, swap in a manager-track competency set against the candidate pool, and add a manager-nomination step before the self-pitch one. Run it once, compare who comes through that filter against who came through the old one, and see how much the two pools overlap.

The QJE paper's quiet implication is that the overlap will be smaller than you expect, and the second pool will produce better managers eighteen months in. That's testable inside your own org. You don't have to take the researchers' word for it.