Every manager knows both failure modes, usually from the inside. Hoarding: every decision routes through her, the queue grows, the team learns helplessness, and vacation becomes a single point of failure. Dumping: exhausted by the queue, she declares the team empowered, withdraws, and discovers months later that three people made three incompatible calls on the same question and nobody flagged the collision. Hoarding at least fails visibly. Dumping fails in silence and gets called empowerment until the invoice arrives.
The way out is not a personality adjustment along some control-freedom dial. It is architectural, and the evidence for that is unusually good.
Delegation is a bad-weather asset
Aghion, Bloom, Lucking, Sadun and Van Reenen studied what happened to firms through the Great Recession using two large datasets across ten OECD countries and US administrative data. Firms that had delegated more decision authority from headquarters to local plant managers before the crisis outperformed centralized peers in the sectors hit hardest. The mechanism is the value of local information under turbulence: when conditions change fast, the person nearest the facts decides better than the person highest in the chart. The authors are careful to note decentralization is not universally optimal; it is specifically an advantage in bad times. For a manager in a volatile market, that caveat reads as a recommendation.
The complementary finding, from Bloom, Sadun and Van Reenen's study of roughly 4,000 manufacturing firms across countries, is that trust is what makes delegation feasible, and that trust-enabled decentralization lets firms grow larger and more productive because the top stops being the bottleneck. A manager who cannot delegate is a growth cap wearing a lanyard.
What the guardrails are
If delegation works, why does it so often produce the dumping disaster? Bain's research across more than 350 organizations offers a blunt answer: only about 15 percent practice effective decision-making, and the recurring pathology is ambiguity about who actually decides. Their RAPID framework exists to force one clarification above all, a single named decider per decision, with the roles of recommending, advising and executing distributed explicitly. The specific labels matter less than the discipline: a delegated decision with no named owner has not been delegated, it has been abandoned.
The second guardrail concerns information. Kathleen Eisenhardt's classic study of strategic decisions in high-velocity environments found the counterintuitive pattern that fast, successful deciders used more real-time information and developed more simultaneous alternatives than slow ones. Speed came from being close to the data, not from skipping deliberation. For the delegating manager, the implication is that she does not protect quality by pre-approving decisions; she protects it by making sure the deciders are saturated with current information, dashboards, customer contact, the real numbers.
The third guardrail is a coordination rule. As Ricardo Semler has told it, when Semco's factory workers began setting their own schedules in 1985, management predicted the assembly line would collapse; it did not, because the policy came with one instruction: talk to the person next to you. Parents took mornings, single workers took evenings, and the line ran. The design insight is that the alternative to central approval is not hope, it is a stated protocol for handling interdependence, cheap enough to run constantly. Adults coordinate when coordination is named as part of the job.
Delegating the decision, keeping the standard
Assembled, the architecture looks like this. Name the decider, in writing, per class of decision, and include what is explicitly reserved (spending caps, irreversible commitments, matters touching other teams). Flood the zone with information; a team deciding on stale data is a hoarding failure that has changed clothes. State the coordination rule, the modern equivalent of talk to the person next to you: who must be informed before a decision of each type lands. And review decisions like a portfolio, on quality of process, not on whether each call matched what the manager would have done, because grading against her own hypothetical choices reinstates the bottleneck through the back door.
Abdication says: you decide, good luck. Delegation says: you decide, here is the information, here is who else needs to know, here is where my authority still applies, and here is when we will look at how it went. The first is a manager subtracting herself. The second is a manager finally doing the part of the job that cannot be delegated, which is designing how deciding works.
Sources
- Aghion, P., Bloom, N., Lucking, B., Sadun, R. & Van Reenen, J. (2021). Turbulence, Firm Decentralization, and Growth in Bad Times. AEJ: Applied Economics 13(1). Firms that delegated to local managers before the Great Recession outperformed centralized firms in the hardest-hit sectors; mechanism is the value of local information under turbulence. https://www.aeaweb.org/articles?id=10.1257%2Fapp.20180752
- Bloom, N., Sadun, R. & Van Reenen, J. (2009/2012). The Organization of Firms Across Countries. NBER WP 15129 / QJE. ~4,000 firms; higher regional trust predicts more decentralization, which supports larger, more productive firms. https://www.nber.org/digest/dec09/international-differences-decentralization-and-productivity
- Rogers, P. & Blenko, M. (2006). Who Has the D? How Clear Decision Roles Enhance Organizational Performance. Harvard Business Review / Bain. Across 350+ organizations, only ~15% decide effectively; RAPID assigns a single accountable decider. https://www.bain.com/insights/who-has-the-d/
- Eisenhardt, K. M. (1989). Making Fast Strategic Decisions in High-Velocity Environments. Academy of Management Journal 32(3). Fast, successful deciders used more real-time information and more simultaneous alternatives, not less. https://acawiki.org/Making_fast_strategic_decisions_in_high-velocity_environments