The Ratchet Effect: Why Your Best Employees Learn to Hide Their True Capacity

Priya runs a customer-support team that closed 1,400 tickets last quarter against a target of 1,000. Her reward? A new target of 1,450 — plus a quiet comment from finance that her team 'clearly has spare capacity'. Next quarter, Priya's team closes 1,460. Exactly enough. Nobody asks how. Priya has just learned the most expensive lesson in performance management: never show the organisation your ceiling.

What the ratchet effect actually is

The ratchet effect describes what happens when an organisation uses an employee's or unit's past performance to set its future standard. Like a mechanical ratchet, the standard moves in one direction only: up. A great year becomes the new normal; a normal year becomes a shortfall. Anyone who understands this — and high performers usually understand it quickly — faces an incentive to under-deliver, delay, or hide their real capacity.

Economist Joseph Berliner documented the behaviour among Soviet factory managers in Factory and Manager in the USSR (1957). Plans were set from last year's output, so managers who overfulfilled were rewarded briefly and then punished permanently with higher quotas. Managers learned to fulfil the plan by a small, safe margin and keep reserves hidden. Martin Weitzman formalised the logic in 'The Ratchet Principle and Performance Incentives' (1980), and Laffont and Tirole (1988) showed why a planner who cannot commit to future targets gets less honest information from the people it manages.

The idea is not only about central planning. In 1952, sociologist Donald Roy spent months working in a Chicago machine shop and described 'quota restriction' in detail: experienced machinists deliberately capped their piecework output because they believed management would cut the piece rate if earnings rose too high. What managers read as a ceiling on ability was a ceiling on disclosure.

“A manager who overfulfils the plan this year is rewarded with a higher plan next year.”

Why smart people sandbag

From the employee's side, the arithmetic is simple. If revealing 40% more capacity earns a one-time bonus but permanently raises expectations by 40%, the long-run value of revealing is often negative. The more farsighted the employee, the stronger the incentive to hide. That is the uncomfortable part: the ratchet effect disproportionately silences your most strategic people.

Where it shows up in modern HR

Goal-setting research adds an important nuance. Locke and Latham's decades of work show that specific, difficult goals generally raise performance compared with vague 'do your best' goals. But the benefit depends on commitment and trust. A goal that feels like a trap — hit it and it rises forever — erodes commitment, and people start optimising for survival rather than for the goal.

How to break the ratchet

A worked example

A regional sales team exceeds its quota by 30%. Under a ratchet system, next year's quota rises by 30% plus company growth. Under a redesigned system, HR and sales operations first ask why: a new product launched in the region, which explains roughly half the jump. The quota is rebuilt from the territory's addressable market, the rep's pipeline quality, and the launch effect. The top rep's new method for qualifying leads is packaged into training; she receives a two-year recognition bonus tied to its adoption. The next year, forecast accuracy improves and deal-slipping between quarters drops — because telling the truth no longer costs anyone.

Limits and honest caveats

• Not every steady performer is sandbagging. Some roles genuinely have stable output.
• Removing all links to past performance is not the answer — history is useful information. The problem is using it as an automatic formula.
• Much of the formal evidence comes from economics, planned economies, and piece-rate settings; applications to knowledge work are well-reasoned but less directly measured.

What the experiments show

The ratchet effect is not only a story told about Soviet factories. It has been reproduced in controlled experiments. Cooper, Kagel, Lo and Gu (1999) ran a planning game with Chinese students and with real Chinese factory managers. When the 'planner' could raise targets after seeing strong output, participants playing workers deliberately held output down. The experienced managers learned to do this faster than students did — practice in a ratcheting system teaches people to game it.

Charness, Kuhn and Villeval (2011) found a related result: when workers knew that a good result would reset the standard, they restricted effort early and revealed capacity only at the end, when the ratchet could no longer bite. In plain terms, people time their honesty. They reveal what they can do when it is too late to be used against them.

The ratchet's close relatives

A diagnostic you can run this quarter

You do not need new software to test for a ratchet. You need three data pulls and one honest conversation.

1. Attainment distribution: plot each team's or rep's result as a percentage of target for the last eight periods. A cluster just above 100% with thin tails is a warning sign. Genuine performance usually looks messier.
2. Period-end timing: compare the last two weeks of each period with the first two weeks of the next. If output drops sharply once the target is safe and rises again at the start of the next period, work is being moved, not done.
3. Forecast bias: compare forecasts with actuals. Forecasts that are beaten by a small, consistent margin are negotiation, not prediction.
4. The safe question: in a skip-level conversation, ask, 'If next year's target could not change because of what you tell me, what could your team realistically deliver?' Compare the answer with the current target.

What to say to a manager who uses the ratchet

Most leaders who ratchet targets are not cynical. They are under pressure to grow, and last year's result is the easiest number to defend in a planning meeting. So the argument has to be about business outcomes, not fairness. A useful script: 'Our current method gives us a target we can defend, but it also trains our best people to hide information from us. We are paying for accuracy we are not getting. Let's build the target from market potential and commit to it for two cycles, and see whether forecasts improve.'

Knowledge work: where the ratchet hides

In sales, the ratchet is visible because quotas are numbers. In knowledge work, it hides inside estimates. An engineering team that delivers a project in six weeks against a ten-week estimate may find every future estimate cut by 40%. The team responds by padding estimates, keeping improvements to itself, or quietly spending the saved time on work nobody asked for. Agile practices such as velocity tracking can accidentally become ratchets when velocity is used as a target rather than a planning aid — which is exactly the misuse Goodhart's law warns against.

• Use velocity and throughput to plan, never to rank teams or set individual goals.
• When a team finds a faster method, give the saved time a named purpose — reducing technical debt, learning, or rest — before it is absorbed by new work.
• Separate the conversation about 'what did we learn' from the conversation about 'what is next quarter's target'.

Checklist: is your target system ratchet-proof?

• Targets are built from opportunity, capacity, or market data — not only from last period's actuals.
• Targets are committed for more than one period where the business allows.
• People who share a better method gain lasting benefit from it.
• Forecast accuracy is recognised, not only over-achievement.
• Some freed capacity is protected for improvement or learning.
• Leaders can explain, in one sentence, how next year's target will be set before this year ends.

Worked case: the support team that stopped improving

A customer support team of twelve handles about 1,800 tickets a month against a target of 1,700. In March, a senior agent builds a set of saved replies and a better triage rule. Output jumps to 2,300 tickets. The head of operations is delighted and, at the next planning cycle, sets the target at 2,300 'because the team has proven it can'. Nobody gets a raise. The agent who built the tool gets a thank-you on Slack.

Six months later, output has settled at exactly 2,310. The saved replies have not been improved since. Two newer agents who found further shortcuts kept them private. In exit interviews a year later, one of them says: 'Everyone knew that if you got faster, you just got more work.' The team did not become lazy. It learned, correctly, what the system rewards.

What a better response would have looked like

1. Name the gain and its source in public, and pay a one-off bonus or give recognition linked to the specific improvement, not to the new output level.
2. Split the saved capacity on purpose: for example, one third to extra volume, one third to quality work such as reducing repeat contacts, one third to time for further improvement.
3. Hold the new target for at least two cycles before reviewing it, and say so in writing.
4. Ask the team to propose the next target, with a reason, and agree it together.

What each role can do on Monday

Common mistakes when fixing a ratchet

• Promising 'we won't raise targets' and then raising them anyway. One broken promise teaches people more than ten good speeches.
• Swinging to the opposite extreme and freezing targets forever. That removes the ratchet but also removes the link between plans and reality.
• Fixing the sales plan and ignoring estimates, budgets and headcount requests, where the same behaviour lives.
• Treating sandbagging as a character problem and disciplining individuals for a pattern the system created.

A 90-day plan