2026-09-19 Weekly Good Reads: Meetings Are Expensive, But Not Necessarily a Waste

Why I picked it

Over the past week, WeChat forwards clustered around a few interlocking questions: whether management is building castles on sand, whether power comes from position or from dependency, what ordinary people can trade upward, and what abilities can truly travel with you. On the surface they're about relationships, position, and prospects, but underneath they all circle one thing: who holds information in an organization, how it flows, and how it turns into decisions.

The new study Meetings by David Deming, Katrine Løken, Alexander Willén, and Yaling Xu happens to enter through a door everyone knows and everyone resents: meetings. In the interpretation published on September 19, 2026, the authors use a survey of 9,099 Norwegian employees matched with employer–employee administrative data to ask why firms are willing to spend such expensive employee time meeting together.

Its most valuable contribution isn't whitewashing meetings — it's pulling apart two questions that often get conflated: meetings obviously have costs, but having costs doesn't mean having no output; meetings being correlated with higher wages, higher income, and faster wage growth doesn't mean "more meetings gets you a raise." What the research really offers is a large set of descriptive facts that lets us re-examine information flow in organizations, rather than memorizing another slogan about "fewer meetings" or "more communication."

The paper's core argument

The survey was conducted by Norstat between May and July 2025, with a sample of Norwegian employees working at least 20 hours per week. The paper ultimately analyzes 9,099 survey participants and links their responses to Norwegian employer–employee administrative data, allowing it to simultaneously observe individual work patterns, firm characteristics, and subsequent wage changes. The survey came from a continuously maintained online panel and offered compensation for participation; the authors report a response rate of about 90%.

Start with scale. Employees in the sample averaged 4.7 hours of meetings per week, roughly 12% of a standard workweek. If you price that time at attendees' wages, the direct labor cost firms put into meetings is about 14% of total payroll. That doesn't yet count pre-meeting prep, post-meeting follow-up, commuting, and attention switching, so the authors treat it as a conservative cost estimate.

These numbers aren't precise stopwatch readings either. The questionnaire bucketed meeting duration into ranges like "under 1 hour, 1 to 3 hours, 4 to 7 hours," and researchers converted using interval midpoints; the top bucket of "over 15 hours" was coded as 20 hours. So 4.7 hours is best read as an estimate at the sample scale, not something to evaluate a specific team against.

Who meets most is also revealing. Managers and professionals spend significantly more time in meetings, while demographic differences like age and gender are relatively small. The most common activities in meetings aren't socializing or administration but planning and strategy, collaboration and problem-solving, project updates, and information sharing. These uses are fairly stable across occupations, industries, and work arrangements.

Differences across firms aren't random noise. Firm effects explain about one-fifth of the variation in weekly meeting hours; about one-quarter on work-from-home days, and about 17% on office days. This means "organizing work through meetings" is itself a firm design choice, not just an individual employee habit.

Next comes the most easily misread result. Firms with more meetings tend to have higher firm wage premia, higher revenue, and higher revenue per worker. Employees at high-wage firms have more expensive time, so the opportunity cost of holding the same meeting is higher — yet they still put more meeting resources in. If meetings were mainly unproductive bureaucratic waste, this pattern would at least be hard to explain.

Similar associations appear at the employee level. People who spend more time in meetings see faster subsequent wage growth; the relationship holds after controlling for individual characteristics, occupation, industry, firm, and initial wages. Compared with focused independent work, administrative tasks, training, travel, and email, meeting time predicts wage growth more strongly. Employees in meeting-intensive environments also report more on-the-job learning, and interacting with more senior colleagues is also associated with faster wage growth.

The authors therefore propose that meetings may be a kind of "organizational capital investment": they consume the expensive time of multiple people being present at once, but they also let knowledge scattered across different roles be combined, let problems be jointly defined, and let decisions and collaboration actually happen. The more specialized modern work becomes, the less any one person holds the whole picture, and the greater the need for cross-role coordination.

That said, the paper never proves that meeting more itself causes firms to earn more or employees to get raises. The authors explicitly say the research is descriptive. People who are more capable, have more complex responsibilities, and are more ambitious may simply be assigned to roles with more meetings; high-performing firms may also simultaneously have better products, talent, and management. Firm fixed effects can't eliminate selection from different task assignments within the same firm.

Points worth questioning

First, the sample is from Norway. Norway's labor relations system, union coverage, organizational hierarchy, and work culture all differ from Chinese private firms, public institutions, small teams, or individual projects. The research can show that "meetings may carry organizational capital," but it can't directly imply that any country or any team should add meetings.

Second, the survey sample doesn't fully represent the entire working population. Compared with the administrative population, respondents were on average older, with 60.9% holding a university degree versus 46.6% overall; 69.2% were union members versus 50% overall; 50.1% were public-sector employees versus 36% overall; and immigrants were a smaller share. Respondents also came more from large firms and from occupations that already depend on coordination. This positive selection may amplify the importance of meetings.

Third, meeting duration, purpose, learning perception, and productivity ratings mainly come from self-reports. People misremember time, and they may describe activities aligned with their role as more valuable. Administrative data improves the credibility of outcomes but can't eliminate survey measurement error.

Fourth, the 14% wage cost only counts attendance time, missing prep, follow-up, and attention residue; at the same time, it doesn't directly price meeting output. An expensive meeting may avoid even more expensive rework, or it may change nothing. Cost alone can't tell you net benefit.

Fifth, the research finds average associations and can't tell us which specific meeting was worth holding. High-wage firms having more meetings doesn't mean stuffing another hour into the calendar will replicate a high-wage firm's capabilities; wage growth correlating with meeting time could also be because more important roles naturally have both more meetings and more promotion opportunities.

Sixth, the paper treats information exchange, problem-solving, and knowledge transfer as the explanations most consistent with existing patterns, but it doesn't randomly assign "who attends which meeting, how meetings are run, whether decisions get made." The truly actionable questions remain unanswered: which topics need synchronous discussion, what information suits asynchronous sharing, what attendee mix produces knowledge transfer, and when coordination costs have already exceeded the benefits.

Takeaways tied to recent interests

This research helps correct an overly simple judgment: management having lots of meetings isn't enough to prove they're "building castles on sand"; meetings correlating with firm performance isn't enough to prove meetings create value. Judging a meeting isn't about whether it happened, but whether it solved a problem that only multiple people synchronously could solve.

Recent forwards repeatedly mention "information value." Meetings are precisely an institutionalized entry point for information flowing up, down, and sideways. If a person just sits in a meeting hearing news, what they get is temporary information; if they can read each party's constraints, fill in missing facts, articulate trade-offs clearly, and push decisions to land, what accumulates is closer to a transferable ability. Positions change, specific news expires, but the ability to identify key problems, organize evidence, and coordinate different roles is portable.

This also explains why power isn't just appointment. A formal position gives you the right to convene meetings, but it doesn't guarantee others will offer real information, let alone that decisions will be executed. Real influence in an organization partly comes from whether others believe you can lower coordination costs: turning vague disputes into judgeable questions, assembling scattered information into actionable plans, and being accountable for post-meeting outcomes.

For managers, rather than vaguely shouting "fewer meetings," keep asking three things in sequence: why can't this be done asynchronously; what mutually scattered information needs to be merged on-site; after the meeting ends, whose what action will change. If all three questions can't be answered clearly, the meeting is probably just status performance. If they can be answered clearly, still leave behind the decision, the basis, the owner, and the review time — otherwise information only flows through in the moment and doesn't settle into organizational capital.

For ordinary participants, don't treat "being invited" itself as a sign of prospects. Meetings may offer access to senior colleagues and understanding of the big picture, but the value depends on whether you complete the learning loop: can you form your own judgment before the meeting, identify information others hold that you lack during it, and after it re-explain the trade-offs without the meeting's exact words and handle similar problems independently. If you only increase exposure without increasing judgment, meetings are just a more advanced form of busyness.

For individual projects and small teams, be even more wary of copying large organizations. The research discusses coordination under specialized division of labor; when one person simultaneously takes on most roles, the knowledge-combination value of meetings is very low, and asynchronous records and direct execution are usually more cost-effective. Only when there genuinely exist different knowledge, different responsibilities, or decisions that must be jointly committed does synchronous discussion have a chance of being worth the cost.

How to read it

Start with the authors' interpretation on VoxEU, focusing on three sets of facts: average meeting hours and direct wage costs, systematic differences in meetings across firms, and the associations between meetings and firm performance, employee wage growth, and on-the-job learning. When you read "meetings may be good for us," immediately write in the margin "descriptive, not causal."

Then read Section 2 of the paper and Appendix Table A.2, comparing the 9,099 respondents with the administrative population on demographics, education, union membership, public sector, and firm size. That determines how far the conclusions can be extrapolated, and prevents directly applying the average pattern of large Norwegian organizations to your own work.

Next, look at the interval conversion of meeting hours, firm fixed effects, and the wage growth regressions. Note what the authors controlled for, and which selection mechanisms remain uncontrolled. You don't need to recompute all the formulas, but you should be able to answer: why "correlation" still can't be called "return."

Finally, take one meeting you attended in the past week and do a small audit: write down what information each party held before the meeting, what new judgments formed during it, and which action changed as a result afterward. If there was only "syncing progress" without new facts, trade-offs, or commitments, convert it to asynchronous; if it genuinely merged scattered knowledge, keep the meeting but tighten the attendees, inputs, and outputs. Read this way, the research isn't defending meetings — it helps you identify which meetings are actually producing organizational capital.

Author interpretation: Deming et al.: Why we spend so much time in meetings

Paper page: NBER Working Paper 35706: Meetings