The perverse incentive of the taut plan, made mechanical. Next period’s target ratchets up from this period’s output — so the rational manager hides capacity. Rewarding output can suppress it.
source Joseph Berliner, Factory and Manager in the USSR (1957) documented it; Martin Weitzman, “The ‘Ratchet Principle’ and Performance Incentives,” Bell Journal of Economics (1980) formalised it. · ROOM: THE STATE — the planning ratchet as a principal–agent game.
Soviet enterprises ran to a plan, not a market. A manager’s bonus turned on meeting the assigned target — the taut plan the ministry set as high as it dared. Berliner’s interviews with émigré managers found the target for each year was set by looking at last year’s output.
Weitzman (1980) gave it a name and a model: the ratchet principle. Because the planner cannot observe true capacity, it infers it from realised output and revises the quota upward — a one-way ratchet the manager learns to game.
The manager knows a plain fact: revealing capacity today raises tomorrow’s quota. A record year is punished with a harder target next year — forever. So the optimal move is to meet the target and hoard the slack: produce enough to earn the bonus, never enough to disclose the ceiling.
This is not laziness. It is a rational response to being measured against your own past. The engine (centre) computes exactly how much the manager withholds as the ratchet α tightens.
The ratchet outlived the USSR. It became a founding case in mechanism design and principal–agent theory: how do you write an incentive when today’s performance sets tomorrow’s standard?
Weitzman’s work seeded the “new Soviet incentive model” (reward a manager for an honest forecast, decoupled from realised output). The same trap recurs everywhere targets are set from history: sales quotas, budget baselines, the “use-it-or-lose-it” budget.
Next period’s target is ratcheted up from this period’s output:
target₂ = base + α · output₁
α ∈ [0,1] is the ratchet strength. α=0 is commitment (the target is fixed, whatever you make). α>0 means every extra unit today raises tomorrow’s quota by α — a standing tax on being seen to produce.
The manager is rewarded for output, pays a convex effort cost, and pays for the raised future quota the ratchet imposes.
The manager’s period-1 payoff is b·q − ½c·q² − p·α·q. Setting the derivative to zero gives a closed-form optimum:
q₁* = (b − p·α) / c
At α=0 the manager gives full effort to the true optimum qmax=b/c. Turn the ratchet on and the optimum slides down: q₁* falls, capacity qmax−q₁* is hoarded, and — because period 2 is unchanged — total output over both periods is strictly less than under commitment.
Punishing success with a higher quota suppresses the very output it chases. This is the pathology behind “taut plans,” chronic under-fulfilment, and pripiski (padded reports).
Two documented symptoms follow directly:
· Slack-hoarding — capacity kept off the books so next year’s quota stays reachable.
· Pripiski (приписки) — outright falsified output, when withholding is not enough to hit the taut target. Ties the-potemkin-village (why the data lied) and gobernet (the plan that could not price).
“Higher targets get more output — pressure lifts production.” Cut. Only if the target is exogenous. When it is ratcheted from your own output, the equilibrium hides capacity: the harder you push, the more the manager withholds. The ratchet suppresses the output it chases.
“Managers just needed more discipline.” Examined. Withholding is the disciplined, optimal play against a ratchet — not indiscipline. The fault is in the incentive, not the man.
Assert the intuition the planners had: “rewarding output must raise it — so the ratchet MAXIMISES total output.” The engine holds the real equilibrium against that claim.
The witness recomputes the equilibrium: under α>0 the optimal output falls and total output drops below the α=0 case. The claim contradicts the math, so the witness turns red.