How Incentives Shape Outcomes

In 1902, the French colonial government in Hanoi faced a rat infestation. To reduce the rat population, they offered a bounty for every dead rat, paying for each rat tail turned in. The program appeared to work initially—thousands of rat tails were collected. But then officials noticed something disturbing: rats with no tails running through the streets. Enterprising residents had discovered it was more profitable to farm rats, cutting off their tails for the bounty while keeping the rats alive to breed more bounty-eligible offspring. The government's incentive scheme had created the opposite of the intended outcome.

This story—often called the cobra effect after a similar incident with cobra bounties in colonial India—illustrates a fundamental truth: incentives shape behavior in powerful and sometimes unexpected ways. What you reward, you get more of. What you punish, you get less of. But the relationship between incentives and outcomes is rarely straightforward. Incentives can align interests, motivate effort, coordinate action—or they can backfire spectacularly, creating perverse behaviors, destroying intrinsic motivation, and producing outcomes worse than doing nothing.

"Show me the incentive and I'll show you the outcome." -- Charlie Munger

Understanding how incentives actually work—the mechanisms through which rewards and punishments shape choices, the conditions under which they succeed or fail, and the unintended consequences they commonly produce—is essential for anyone designing systems, leading organizations, making policy, or simply trying to understand human behavior.

This article explains the mechanics of incentives: what types exist, how they influence decisions, why they often fail, and principles for designing incentives that actually produce desired outcomes without catastrophic side effects.

What Are Incentives? Defining the Core Mechanism

At its most basic, an incentive is anything that motivates or influences behavior by changing the costs or benefits of different choices. Incentives operate through consequences:

Positive incentives (rewards): Benefits for performing a desired action
Negative incentives (penalties): Costs for performing an undesired action

But this simple definition conceals considerable complexity. Incentives vary along multiple dimensions:

Dimension	Type	Examples
Nature	Financial, social, intrinsic, physical	Money, status, meaning, comfort
Timing	Immediate, delayed	Instant bonus vs. retirement savings
Certainty	Deterministic, probabilistic	Fixed salary vs. commission
Source	External, internal	Manager's approval vs. personal satisfaction
Scope	Individual, collective	Personal bonus vs. team profit-sharing

Moreover, incentives operate not just through direct costs and benefits but through several psychological and social mechanisms:

Signaling: What Organizations Value

Incentives communicate what the incentive-designer values. If a company pays salespeople based solely on revenue, it signals "we care about revenue, not customer satisfaction or ethical behavior." If a university promotes professors based on publication count, it signals "quantity matters more than impact."

People respond not just to the direct payoff but to the implicit message about priorities. This signaling function can be more powerful than the incentive's material value.

Attention Direction: What People Focus On

Incentives direct attention and effort toward incentivized activities and away from non-incentivized ones. Psychologist Daniel Kahneman describes attention as a scarce resource—incentives allocate it.

If salespeople are incentivized on closing deals but not on after-sale support, support quality will deteriorate even if it's nominally "part of the job." People focus where the rewards are.

Information Creation: What Gets Measured

Incentives depend on metrics—measurable indicators of performance. But measurement itself shapes behavior. The famous Goodhart's Law states: When a measure becomes a target, it ceases to be a good measure.

Once you incentivize a metric, people optimize for that metric specifically, often in ways that don't advance the underlying goal. This is the core mechanism behind many incentive failures.

Types of Incentives: The Motivation Toolkit

Different types of incentives operate through different psychological mechanisms and work better for different tasks and people.

Financial Incentives: Money and Material Rewards

Financial incentives—salaries, bonuses, commissions, stock options, penalties—are the most obvious and commonly used. They work through straightforward economic logic: people prefer more money to less.

When financial incentives work well:

Simple, measurable tasks: Call center call volume, manufacturing output, sales
Clear performance metrics: Unambiguous indicators of success
Low intrinsic motivation baseline: Tasks people wouldn't otherwise choose to do
Alignment with goals: What's incentivized matches what's desired

When financial incentives fail or backfire:

Complex, creative tasks: Narrow focus on rewards can reduce creative problem-solving (the overjustification effect)
Judgment-heavy work: Teaching, healthcare, research—quality is hard to measure
Team settings: Individual incentives can undermine cooperation
Ethical dimensions: Incentivizing easily-gamed metrics encourages gaming

Economist Uri Gneezy and psychologist Aldo Rustichini demonstrated this in a famous daycare study: when daycares introduced fines for late pickup, late pickups increased. The fine converted a social norm (don't inconvenience caregivers) into a price (pay to stay late), reducing guilt and making lateness feel acceptable.

"The moment you introduce a monetary incentive to a social norm, you undermine the social norm." -- Uri Gneezy

Social incentives operate through reputation, status, social approval, and belonging. Humans are intensely social; being respected, admired, included, or avoiding shame are powerful motivators.

Examples:

Public recognition: Employee of the month, leaderboards, awards
Status hierarchies: Titles, offices, access to leadership
Peer approval: Team feedback, social media likes
Shame and exclusion: Public criticism, ostracism

Social incentives can be more powerful than financial ones in contexts where:

Social identity matters (professional communities, volunteer organizations)
Reputation has long-term value (academia, small industries)
Peer comparison is visible (rankings, public dashboards)

But they can also backfire:

Public shaming can create resentment and opposition
Zero-sum status competition can destroy collaboration
Visible metrics can create incentive to game rather than perform

Intrinsic Motivation: Autonomy, Mastery, and Purpose

Intrinsic motivation comes from within: the satisfaction of doing work that's interesting, meaningful, or consistent with one's values. Psychologist Edward Deci identified three core sources:

"Human beings can be proactive and engaged, or alternatively passive and alienated, largely as a function of the social conditions in which they develop and function." -- Edward Deci

Autonomy: Control over one's work—choosing how, when, what
Mastery: Growth, learning, getting better at something challenging
Purpose: Work that matters, aligns with values, contributes to something larger

Intrinsic motivation is especially important for:

Creative and complex work requiring insight and innovation
Knowledge work where quality is multidimensional and hard to measure
Long-term projects where sustained engagement matters more than short bursts

The critical insight from decades of research: Extrinsic incentives can crowd out intrinsic motivation. When people are paid to do something they previously found interesting, they often become less motivated to do it in the absence of payment. The external reward reduces the activity from "something I want to do" to "something I do for money."

Negative Incentives: Penalties and Punishments

Negative incentives—fines, penalties, sanctions, termination—attempt to reduce undesired behavior through threatened costs.

They're effective when:

Behavior is clearly defined and easily observed (speeding, deadline violations)
Deterrence is the goal (preventing violations rather than encouraging excellence)
Compliance is necessary regardless of motivation

But they have significant limitations:

Crowding out: Converting ethical obligations into prices (daycare fine example)
Resentment: People resist control and punishment
Focus on minimum: People do just enough to avoid penalty, not strive for excellence
Gaming: People avoid detection rather than change behavior

Moreover, excessive punishment can create a culture of fear that suppresses innovation, hides mistakes, and encourages cover-ups rather than improvement.

The Mechanisms: How Incentives Actually Change Behavior

Understanding how incentives influence behavior reveals why they sometimes produce unintended outcomes.

Direct Effect: Changing the Cost-Benefit Calculation

The most obvious mechanism is direct economic impact: incentives change the relative payoffs of different actions. If you pay people more to work overtime, some will choose to work more hours. If you fine littering, some people will litter less.

This mechanism is captured in standard economic models of rational choice—people compare costs and benefits and choose actions with the highest net benefit.

Substitution Effect: Shifting Effort Between Activities

Incentives don't just increase effort on incentivized tasks—they also shift effort away from non-incentivized tasks. Economists call this the substitution effect.

Example: If teachers are evaluated on student test scores, they may:

Increase time on test-relevant material (intended)
Reduce time on non-tested subjects like art, music, critical thinking (unintended)
Focus on marginal students near passing thresholds, neglecting high and low performers (unintended)
Teach to the test rather than fostering deep understanding (unintended)

The result: apparent success on the metric (higher test scores) may not reflect the underlying goal (better education).

Sorting Effect: Attracting Certain Types of People

Incentives don't just change behavior—they also select who participates. Different incentive structures attract different types of people.

Behavioral economist Dan Ariely showed this in experiments where payment schemes attracted people with different motivations:

Pay-for-performance: Attracts competitive, extrinsically motivated individuals
Flat pay with mission emphasis: Attracts prosocial, intrinsically motivated individuals

If you design incentives assuming all people respond the same way, you may attract people whose motivations don't align with organizational goals.

Crowding Out Effect: Undermining Intrinsic Motivation

Perhaps the most counterintuitive mechanism: extrinsic incentives can reduce overall motivation by replacing intrinsic reasons with extrinsic ones.

Bruno Frey and others documented this in domains like:

Blood donation: Paying donors reduced donation rates in some contexts by making it transactional
Volunteering: Paying small amounts can reduce volunteer hours by reducing the "warm glow" of altruism
Creative work: Rewards for creative tasks can reduce creativity by narrowing focus

The mechanism: when extrinsic rewards are introduced for an activity people previously found intrinsically rewarding, they reinterpret the activity as "done for the reward." If the reward is removed, motivation doesn't return to baseline—it falls below it.

Norm Change: Redefining What's Appropriate

Incentives can change social norms—shared expectations about appropriate behavior. The daycare fine example illustrates this: the fine converted lateness from a social violation (breaking trust with caregivers) to an economic transaction (paying for extra service).

Once the norm shifted, removing the fine didn't restore the old norm—late pickups remained higher than before the fine was introduced. The norm change persisted.

When Incentives Backfire: Common Failure Modes

Incentives fail predictably in several ways. Recognizing these patterns can help avoid designing disastrous systems.

Cobra Effect: Creating Perverse Incentives

The cobra effect (from the rat tail bounty story) occurs when incentives create behaviors that advance the metric without advancing the goal—or even work against it.

Classic examples:

Soviet factories: Incentivized by tonnage produced, so factories made excessively heavy, impractical products
Cobra bounties: Paying for dead cobras led to cobra farming
Body counts in Vietnam: Incentivizing enemy kills led to inflated numbers and civilian deaths
Wells Fargo accounts: Incentivizing new accounts opened led to millions of fraudulent accounts

The pattern: Narrow metric + high-powered incentive + weak oversight = gaming.

Metric Fixation: Losing Sight of Goals

Jerry Muller's book The Tyranny of Metrics documents how metric fixation—obsessive focus on quantitative performance measures—leads to:

Teaching to the test rather than fostering learning
Publishing trivial papers rather than conducting meaningful research
Hitting quotas rather than serving customers
Short-term earnings management rather than building value

The core problem: proxies replace goals. The metric (test scores, publication counts, sales numbers) is a proxy for the real goal (education, knowledge creation, customer satisfaction). But once incentivized, people optimize the proxy even when it diverges from the goal.

Multitasking Problem: Neglecting the Unmeasured

When some dimensions of performance are measured and incentivized while others are not, effort shifts toward the measured dimensions. Economists Bengt Holmström and Paul Milgrom formalized this as the multitasking problem.

Example: Police departments that incentivize arrest rates may:

Increase arrests (measured, incentivized)
Reduce community relationship-building (unmeasured)
Target minor offenses for easy arrests rather than serious crimes

The solution isn't always "measure more things"—that can create overwhelming complexity. Sometimes the answer is weaker incentives to avoid extreme distortion.

Ratchet Effect: Punishment for Success

The ratchet effect occurs when high performance leads to higher future targets, effectively punishing success. If a salesperson exceeds quota, next year's quota increases. Rational response: don't overperform.

This creates:

Sandbagging: Deliberately performing just below maximum to avoid raised expectations
End-of-period manipulation: Delaying sales to next period if current quota is met
Reduced effort: Avoiding being the "tall poppy" that gets cut down

Soviet factories famously experienced this—meeting quotas too easily led to higher future quotas, so managers hid capacity and underperformed intentionally.

Crowding Out Prosocial Behavior

In some contexts, introducing incentives reduces the desired behavior by displacing altruistic or social motivations. This is especially problematic for:

Volunteer work: Small payments can reduce volunteer hours
Public goods contributions: Fines for non-contribution can reduce contribution rates
Environmental behavior: Payments for recycling can reduce recycling rates

The mechanism: extrinsic incentives signal distrust or change the frame from "good citizen" to "economic agent." Once the frame shifts, removing the incentive doesn't restore prosocial motivation.

Principles for Effective Incentive Design

Decades of research and painful failures suggest several design principles:

1. Align Incentives with Actual Goals, Not Just Proxies

The tighter the link between the incentive metric and the actual goal, the less room for gaming and misalignment.

Bad: Incentivize number of patients seen (encourages rushing)
Better: Incentivize patient outcomes, satisfaction, and volume together

Bad: Incentivize lines of code written (encourages bloated code)
Better: Incentivize working software delivered, bugs fixed, user satisfaction

2. Consider Unintended Consequences Systematically

Before implementing incentives, ask:

What behaviors will this reward that we don't want?
What unmeasured dimensions will be neglected?
How might people game this metric?
What happens if people take the incentive seriously and optimize it ruthlessly?

Conduct pre-mortems: imagine the incentive has failed spectacularly and work backward to identify failure modes.

3. Use Multiple Metrics to Balance Incentives

Single metrics create tunnel vision. Multiple metrics can balance competing priorities:

Quality and quantity: Measure both output and error rates
Short-term and long-term: Incentivize quarterly results and multi-year growth
Individual and team: Balance personal rewards with collective success

But avoid metric overload—too many metrics create confusion. Aim for 3–7 key indicators.

4. Preserve Intrinsic Motivation

For complex, creative work, avoid high-powered extrinsic incentives that crowd out intrinsic motivation. Instead:

Emphasize autonomy: Give people control over how they work
Provide mastery opportunities: Challenging work, learning, growth
Highlight purpose: Connect work to meaningful outcomes

Use incentives to remove barriers (fair pay, reasonable conditions) rather than to control behavior.

5. Design for Sorting: Attract the Right People

Consider what types of people your incentives attract:

Mission-driven incentives: Attract intrinsically motivated people aligned with organizational values
High-variance incentives (big bonuses for success): Attract risk-tolerant, competitive individuals
Stable, predictable incentives: Attract risk-averse, steady performers

Match incentive structure to the type of talent you need.

6. Build in Safeguards Against Gaming

Anticipate gaming and design defenses:

Audits and oversight: Random checks, third-party verification
Long-term evaluation: Delayed rewards that depend on sustained performance
Peer monitoring: Team-based incentives where peers police gaming
Reputational stakes: Make gaming costly to long-term reputation

7. Start Small and Iterate

Incentive schemes have complex, often unpredictable effects. Rather than full-scale rollout:

Pilot test with small groups
Monitor for unintended consequences
Adjust rapidly when problems emerge
Communicate changes to maintain trust

Incentive design is not a one-time engineering problem—it's an ongoing evolutionary process.

Real-World Cases: Incentives in Action

Examining real-world successes and failures illuminates these principles.

Success: Lincoln Electric's Piecework System

Lincoln Electric, a manufacturing company, has used piecework pay (payment per unit produced) for over a century with remarkable success. Why does it work here when piecework often fails?

Simple, measurable output: Welding equipment—units are clear, quality is verifiable
Worker control: Employees control production pace and methods
Long-term employment: Reduces incentive to sacrifice quality for short-term output
Peer quality control: Work flows through teams; poor quality hurts everyone
Profit-sharing: Aligns individual and company success

This system works because context aligns with incentive structure.

Failure: Soviet Nail Factory

Soviet central planners incentivized nail factories by weight of nails produced. Factories responded by producing huge, useless nails—maximizing weight, not usefulness.

Planners switched the incentive to number of nails produced. Factories responded by producing tiny, useless nails—maximizing count, not usefulness.

The lesson: Narrow metrics enable gaming. Complex goals (produce useful nails) can't be reduced to single metrics.

Mixed: Teacher Performance Pay

Experiments with pay-for-performance in education have produced mixed results:

Some studies show modest test score gains in contexts with strong accountability and well-designed metrics
Other studies show no effect, teaching-to-the-test, cheating scandals, and reduced teacher morale
Long-term effects are often disappointing—gains on tested material don't translate to broader learning

The pattern: when teaching can be reduced to simple, measurable tasks (drill-and-practice for basic skills), incentives can work. When teaching requires complex judgment, creativity, and relationship-building, incentives often backfire.

Success: GitHub's Open Source Incentives

GitHub built a platform where social incentives (reputation, visibility, community respect) motivate massive contributions to open-source software. Contributors receive:

Visible contributions: Public repositories showing expertise
Community status: Recognition from respected peers
Portfolio building: Evidence of skills for employment
Intrinsic rewards: Working on meaningful projects

Financial incentives are absent or minimal, yet the platform has produced enormous value. Why?

Social incentives align with intrinsic motivation (autonomy, mastery, purpose)
Metrics (commits, pull requests, stars) are meaningful proxies for quality
Community norms police bad behavior and reward excellence
Sorting effect: Attracts intrinsically motivated contributors

The Psychology of Incentives: What Behavioral Science Reveals

Traditional economic models assume people respond to incentives as rational, self-interested utility-maximizers. Behavioral science reveals a more nuanced picture:

Prospect Theory: Loss Aversion and Framing

Kahneman and Amos Tversky's prospect theory shows:

Loss aversion: People are more motivated to avoid losses than to achieve equivalent gains
Framing effects: Presenting an incentive as avoiding loss (keep your bonus by meeting targets) is more motivating than framing as potential gain (earn a bonus)

Implication: Framing matters—how you present incentives affects their power.

Present Bias and Hyperbolic Discounting

People heavily discount delayed rewards and overweight immediate ones. A $100 bonus today is far more motivating than a $100 bonus in six months, even though the value is the same.

Implication: Immediate feedback and rewards are more effective than delayed ones, especially for sustaining effort.

People care about fairness, reciprocity, and others' welfare—not just personal material gain. Incentives perceived as unfair (unequal, exploitative) can backfire even when economically beneficial.

Economist Ernst Fehr documented this in ultimatum game experiments: people reject positive offers if they perceive them as unfairly small, preferring zero to accepting an insult.

Implication: Perceived fairness matters as much as magnitude. Transparent, equitable incentives are more effective.

The Endowment Effect and Status Quo Bias

People overvalue what they already have (endowment effect) and resist change (status quo bias). Removing benefits feels like a loss, even if never used.

Implication: Adding benefits is easier than removing them. Be cautious introducing temporary incentives that may be hard to discontinue.

Incentives in Different Domains: Context Matters

Incentive effectiveness varies dramatically by context. What works in sales fails in research; what works in manufacturing fails in healthcare.

Manufacturing and Simple Services

Context: Repetitive tasks, clear output, measurable quality
Effective incentives: Piece-rate pay, productivity bonuses, quality metrics
Risks: Overemphasis on speed sacrificing quality; worker burnout

Creative and Knowledge Work

Context: Complex, judgment-heavy, quality is multidimensional
Effective incentives: Autonomy, mastery opportunities, meaningful work, fair base compensation
Risks: High-powered extrinsic incentives crowd out intrinsic motivation; narrow metrics distort priorities

Healthcare and Education

Context: Outcomes depend on many factors (patient health, student background); quality is hard to measure; professional ethics matter
Effective incentives: Moderate, balanced metrics; peer recognition; professional standards
Risks: Gaming metrics (teaching to test, cherry-picking patients); crowding out professionalism

Public Sector and Government

Context: Goals are often diffuse and political; oversight is weak; accountability is distant
Effective incentives: Mission-driven recruitment; transparency; career incentives (promotion)
Risks: Metric gaming is rampant; perverse incentives are common; public outcry when failures visible

Conclusion: The Power and Peril of Incentives

"It is difficult to get a man to understand something when his salary depends upon his not understanding it." -- Upton Sinclair

Incentives are among the most powerful tools for shaping behavior and coordinating action in organizations, markets, and societies. When well-designed, they can align interests, motivate effort, signal priorities, and produce remarkable outcomes. When poorly designed, they can produce disasters: perverse behavior, destroyed motivation, eroded trust, and outcomes opposite of those intended.

The central lesson from decades of research and practice: incentives are not simple levers. They operate through multiple psychological and social mechanisms. They have unintended effects that are often stronger than intended ones. They interact with intrinsic motivation, social norms, professional ethics, and organizational culture in complex ways.

Effective incentive design requires:

Deep understanding of the actual goals, not just convenient proxies
Anticipation of how people will respond, including gaming and unintended consequences
Balance between incentivizing performance and preserving intrinsic motivation
Continuous monitoring and adjustment as context changes

Most fundamentally, incentive design is not a purely technical problem—it's a human problem. Understanding how real people (not idealized rational agents) actually respond to rewards and punishments is essential for creating systems that produce desired outcomes without catastrophic side effects.

The rats of Hanoi remind us: when you change incentives, you change behavior—but you may not change it in the way you intended.

Key Researchers and Their Contributions

The scientific study of how incentives shape behavior draws from economics, psychology, and organizational theory, with a relatively small group of researchers producing the foundational empirical and theoretical work.

Bruno Frey (born 1941) is a Swiss economist at the University of Zurich and later at the University of Basel who developed the crowding out theory of motivation in the 1990s. His 1997 book Not Just for the Money: An Economic Theory of Personal Motivation challenged the standard economic assumption that more money always increases motivation, presenting evidence that extrinsic financial rewards can reduce intrinsic motivation by changing the psychological framing of an activity. Frey's research drew on experiments in fields from blood donation to volunteer work to professional tasks, consistently finding that adding financial rewards to activities people found intrinsically meaningful reduced their motivation to perform the activity when the reward was removed, and sometimes even when it was present. His 2001 survey paper with Reto Jegen, "Motivation Crowding Theory," synthesized this evidence for an economics audience and helped move the crowding out hypothesis from a psychological curiosity into mainstream economic discussion.

Uri Gneezy is an Israeli-American behavioral economist at the Rady School of Management at UC San Diego who has conducted some of the most influential field experiments on incentive effects. His 2000 paper with Aldo Rustichini, "A Fine is a Price," published in the Journal of Legal Studies, demonstrated that introducing a fine for parents who arrived late to pick up their children from a Haifa daycare center caused late pickups to increase rather than decrease. The paper became one of the most cited in behavioral economics and illustrated the norm-transformation mechanism: the fine converted a social obligation into an economic transaction, removing the guilt that had previously motivated on-time pickup. Gneezy has subsequently studied incentives for exercise, charitable giving, educational performance, and workplace productivity, consistently finding that the effects of financial incentives depend heavily on context, framing, and the social norms they activate or disrupt.

Edward Deci (born 1942) is a psychologist at the University of Rochester who, with Richard Ryan, developed Self-Determination Theory (SDT), the most influential psychological framework for understanding the relationship between extrinsic and intrinsic motivation. Deci's 1971 paper "Effects of Externally Mediated Rewards on Intrinsic Motivation" in the Journal of Personality and Social Psychology demonstrated that paying people to perform intrinsically interesting activities (puzzle-solving) reduced their subsequent engagement with those activities when the payment was removed, compared with a control group that was never paid. This finding, initially controversial because it contradicted straightforward economic predictions, has been replicated hundreds of times across different tasks, populations, and reward structures. SDT distinguishes between motivation that is fully autonomous (intrinsic) and motivation that is controlled by external pressures (extrinsic), and predicts that controlled motivation produces less creative, less persistent, and lower-quality performance than autonomous motivation.

Bengt Holmstrom (born 1949) is a Finnish-American economist at MIT who received the Nobel Prize in Economics in 2016 (jointly with Oliver Hart) for his contributions to contract theory and incentive design. His 1979 paper "Moral Hazard and Observability" in the Bell Journal of Economics established the foundational result that incentive contracts should link compensation to all observable signals that carry information about the agent's effort. His 1991 paper with Paul Milgrom, "Multitask Principal-Agent Analyses," formalized the multitasking problem: when agents perform multiple tasks, high-powered incentives on some tasks cause agents to redirect effort away from other tasks, often with damaging consequences. This result is the formal theoretical foundation for the observation that incentivizing test scores causes teachers to neglect untested subjects, incentivizing sales volumes causes salespeople to neglect customer relationships, and incentivizing quantity causes workers to sacrifice quality.

Steven Kerr (1941-2022) was an organizational behavior professor and chief learning officer at General Electric who wrote what became one of the most cited papers in management: "On the Folly of Rewarding A, While Hoping for B," published in the Academy of Management Journal in 1975. Kerr documented dozens of real-world cases in which organizations designed incentive systems that rewarded one behavior while hoping to achieve a different outcome, including university tenure systems that reward publication while hoping for teaching quality, political systems that reward short-term visibility while hoping for long-term policy effectiveness, and military systems that reward body counts while hoping for enemy defeat. The paper's central insight is simple but underappreciated: if you measure and reward one thing while caring about another, you will systematically get more of the measured thing and less of what you actually want.

Historical Case Studies That Changed the Field

Several specific experiments and organizational interventions have provided the empirical foundation for understanding when and why incentives succeed or fail.

The Daycare Fine Study in Haifa (1998-2000). Uri Gneezy and Aldo Rustichini's study of ten daycare centers in Haifa, Israel is one of the most elegant natural experiments in behavioral economics. The researchers observed baseline late pickup rates for four weeks, then introduced a fine for late pickups at six centers while leaving the other four as controls, observed the centers for twelve more weeks, then removed the fine and observed for another four weeks. The results were striking: in the six centers with fines, late pickups immediately increased and remained elevated even after the fine was removed. In the four control centers, late pickup rates remained roughly constant throughout. The experiment provided clean causal evidence that the fine had converted a social norm (don't inconvenience the teachers who stay late for your child) into an economic transaction (pay a modest fee for extra service), and that this normative shift persisted even after the economic incentive was withdrawn.

The Lincoln Electric Long-Run Study (1990s-2000s). Lincoln Electric, a Cleveland-based manufacturer of welding equipment, has used a piece-rate pay system combined with a no-layoff policy and annual profit-sharing bonuses for most of its history. Management researchers including John Shea at University of Maryland and researchers at Harvard Business School have studied Lincoln Electric over multiple decades as one of the few large manufacturing companies where piece-rate pay has been consistently successful. The research identifies several conditions that make Lincoln Electric unusual: the output is easily measured (welded components with clear quality standards), workers have substantial control over their production methods, the no-layoff policy removes the fear that high productivity will lead to workforce reduction, and the profit-sharing bonus creates alignment between worker and company interests. Lincoln Electric's hourly workers earn significantly more than comparable manufacturing workers elsewhere but work at higher productivity, making the system profitable for the company. The case is regularly taught in business schools as an example of how incentive systems can work when the conditions for successful piece-rate pay are met.

The RAND Health Insurance Experiment (1974-1982). The RAND Corporation's Health Insurance Experiment, led by Joseph Newhouse, randomly assigned approximately 3,900 families in six U.S. cities to health insurance plans with different levels of cost-sharing, ranging from free care to 95% coinsurance with a catastrophic coverage cap. The experiment ran for three to five years per family and is still the largest randomized controlled trial of health insurance design ever conducted. The key finding for incentive theory was that cost-sharing substantially reduced healthcare utilization (families with 95% coinsurance used about 30% less healthcare than those with free care) without significantly worsening health outcomes for the average family, though the sickest and poorest families experienced worse health outcomes with high cost-sharing. The experiment demonstrated that financial incentives significantly affect healthcare utilization decisions, and that the effects are roughly consistent with economic predictions for average patients but can be harmful for vulnerable populations. Its findings directly influenced the design of Medicare and Medicaid cost-sharing requirements and continue to inform debates about health insurance design.

The Wells Fargo Incentive Collapse (2011-2016). Wells Fargo's fake accounts scandal is perhaps the most extensively studied large-scale incentive failure in corporate history. The Senate Banking Committee held hearings in 2016 and 2017 at which CEO John Stumpf testified; independent reports by the law firm Shearman & Sterling (commissioned by Wells Fargo's board) and by the Office of the Comptroller of the Currency documented the sequence of events in detail. The fundamental design failure was that sales quotas (8 products per customer) were set at levels that Wells Fargo's own internal research showed were unachievable for most customers through legitimate means. Branch managers who knew the quotas were impossible nonetheless enforced them because their own compensation depended on their teams' results. The resulting pressure created a situation where fraudulent account creation was the rational response for employees who needed to keep their jobs. The scandal cost Wells Fargo over $3 billion in fines and settlements and resulted in criminal charges against multiple former employees, making it a canonical case study for business school courses on incentive design and corporate ethics.

How These Ideas Are Applied Today

The practical application of incentive research has expanded from academic economics into organizational design, public policy, healthcare, and digital platform management.

Pay-for-Performance in Healthcare. The Affordable Care Act (2010) introduced a series of value-based payment programs that link hospital and physician reimbursement to quality outcomes rather than volume of services. The Hospital Value-Based Purchasing Program, implemented by the Centers for Medicare and Medicaid Services beginning in 2012, adjusts hospital payments based on clinical outcomes, patient experience, efficiency, and safety measures. Early evaluations found modest but consistent improvements in the incentivized quality measures, with some evidence of the multitasking problem: hospitals that improved on incentivized measures sometimes showed deterioration on non-incentivized measures. The Commonwealth Fund and the RAND Corporation have conducted extensive evaluations of these programs, finding that financial incentives work better when they are large enough to motivate change, when the measures closely align with actual quality, and when hospitals have the organizational capacity to respond.

Behavioral Incentives in Retirement Savings. The SECURE Act of 2019 and SECURE 2.0 of 2022 in the United States further institutionalized behavioral economics insights about incentive design in retirement savings, building on the Pension Protection Act of 2006. SECURE 2.0 requires employers starting new 401(k) plans to automatically enroll workers at a default contribution rate of 3-10% and automatically escalate contributions annually, directly implementing Richard Thaler and Shlomo Benartzi's "Save More Tomorrow" program at the level of federal law. The legislation also created new incentives (tax credits) for small businesses to establish retirement plans, addressing the sorting problem that many low-wage workers lack access to employer-sponsored retirement savings vehicles. Vanguard's 2023 "How America Saves" report, which analyzes data from over 5 million plan participants, documents that automatic enrollment has increased participation rates to 93% in plans that use it, compared with 67% in opt-in plans.

Algorithmic Incentive Management in the Gig Economy. Companies including Uber, Lyft, DoorDash, and Amazon Flex use algorithmically designed incentive systems to manage their gig worker populations. Uber's "surge pricing" mechanism is an incentive designed to attract more drivers during periods of high demand; academic research by economists at Uber and by independent researchers including Jonathan Hall and Alan Krueger has documented that surge pricing does increase driver supply. However, the algorithmic management of gig workers also illustrates the multitasking problem: drivers optimized for surge earnings show patterns of location-seeking behavior that can reduce service reliability in low-demand areas. Research by sociologists Alex Rosenblat and Luke Stark documented how Uber's algorithmic management creates information asymmetries that systematically advantage the company over drivers. These findings have informed legislative debates about gig worker classification in California (Proposition 22, 2020), the UK Supreme Court's Uber ruling (2021), and EU platform worker legislation.

References

Ariely, D., Gneezy, U., Loewenstein, G., & Mazar, N. (2009). Large stakes and big mistakes. Review of Economic Studies, 76(2), 451–469. https://doi.org/10.1111/j.1467-937X.2009.00534.x

Deci, E. L. (1971). Effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology, 18(1), 105–115. https://doi.org/10.1037/h0030644

Fehr, E., & Gächter, S. (2000). Cooperation and punishment in public goods experiments. American Economic Review, 90(4), 980–994. https://doi.org/10.1257/aer.90.4.980

Frey, B. S., & Jegen, R. (2001). Motivation crowding theory. Journal of Economic Surveys, 15(5), 589–611. https://doi.org/10.1111/1467-6419.00150

Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of Legal Studies, 29(1), 1–17. https://doi.org/10.1086/468061

Holmström, B., & Milgrom, P. (1991). Multitask principal-agent analyses: Incentive contracts, asset ownership, and job design. Journal of Law, Economics, & Organization, 7, 24–52. https://doi.org/10.1093/jleo/7.special_issue.24

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291. https://doi.org/10.2307/1914185

Kerr, S. (1975). On the folly of rewarding A, while hoping for B. Academy of Management Journal, 18(4), 769–783. https://doi.org/10.5465/255378

Lazear, E. P. (2000). Performance pay and productivity. American Economic Review, 90(5), 1346–1361. https://doi.org/10.1257/aer.90.5.1346

Muller, J. Z. (2018). The tyranny of metrics. Princeton University Press. https://doi.org/10.2307/j.ctvc77bmd

Pink, D. H. (2009). Drive: The surprising truth about what motivates us. Riverhead Books.

Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic Literature, 37(1), 7–63. https://doi.org/10.1257/jel.37.1.7

Weitzman, M. L. (1976). The new Soviet incentive model. Bell Journal of Economics, 7(1), 251–257. https://doi.org/10.2307/3003197

Gneezy, U., & Rustichini, A. "A Fine Is a Price." Journal of Legal Studies, 2000.
Deci, E. L., & Ryan, R. M. "Self-Determination and Intrinsic Motivation in Human Behavior." Plenum Press, 1985.
Kahneman, D. "Thinking, Fast and Slow." Farrar, Straus and Giroux, 2011.
Muller, J. Z. "The Tyranny of Metrics." Princeton University Press, 2018.
Pink, D. H. "Drive: The Surprising Truth About What Motivates Us." Riverhead Books, 2009.
Frey, B. S., & Jegen, R. "Motivation Crowding Theory." Journal of Economic Surveys, 2001.
Holmström, B., & Milgrom, P. "Multitask Principal-Agent Analyses." Journal of Law, Economics, & Organization, 1991.
Kerr, S. "On the Folly of Rewarding A, While Hoping for B." Academy of Management Journal, 1975.
Ariely, D. "Predictably Irrational: The Hidden Forces That Shape Our Decisions." HarperCollins, 2008.
Fehr, E., & Gächter, S. "Cooperation and Punishment in Public Goods Experiments." American Economic Review, 2000.

Field Experiments That Reshaped Incentive Design Practice

Several specific experiments have generated findings so clear and surprising that they directly changed how practitioners design incentive systems in healthcare, education, and organizational management.

The Safelite Glass Productivity Study (1994-1996). Economist Edward Lazear at the Hoover Institution at Stanford studied what happened when Safelite AutoGlass, a national auto glass repair company, switched from hourly wages to piece-rate pay. The natural experiment, involving approximately 3,000 workers, was published in the American Economic Review in 2000. Productivity increased by 44% on average after the switch to piece rates. Critically, the increase came from two sources of roughly equal magnitude: workers who had been employed under the hourly wage system increased their productivity, demonstrating a direct incentive effect, while worker composition shifted because the piece-rate system attracted higher-ability workers and caused lower-ability workers to leave, demonstrating the sorting effect. Lazear's data also showed that the quality of installations did not decline despite the increased production pace, because Safelite built in a quality accountability mechanism: workers were required to return and repair at their own time cost any installation that failed quality checks. The study is considered the gold-standard evidence that piece-rate pay can work in service settings when output is measurable and quality accountability is maintained.

The Israeli High School Cash Incentives Experiment (2001-2002). Economist Victor Lavy at Hebrew University conducted a randomized controlled trial paying Israeli high school students cash bonuses for passing matriculation exams, publishing results in the Review of Economics and Statistics in 2009. The experiment randomly assigned students in 49 schools to receive either no bonus, a bonus for passing one or more subjects, or a bonus for improving their grade point average. Students who received the subject-specific bonus showed a 3.7 percentage point increase in the probability of passing the incentivized exams, but also showed decreased performance in non-incentivized subjects, providing clean causal evidence of the multitasking substitution effect in an education context. Students who received the GPA-based bonus, which incentivized overall performance rather than specific subjects, showed a 4.5 percentage point increase in passing rates without the negative cross-subject substitution effect. The study demonstrated that incentive scope design, whether to incentivize specific outputs or overall outcomes, materially affects whether beneficial spillovers or harmful substitution effects dominate.

The Moral Licensing Effect in Sustainability Incentives (2014). Economists Alison Caris, Reto Foellmi, and Josef Zweimuller conducted a study in collaboration with a Swiss electricity utility examining whether customers who received positive feedback about their green energy consumption subsequently reduced their overall energy conservation effort. The study found a moral licensing effect: households informed that their previous conservation behavior was "excellent" subsequently increased their energy consumption by approximately 8%, while those given standard information maintained or reduced consumption. This finding, replicated in broader meta-analyses of pro-environmental behavior by Nina Mazar and Chen-Bo Zhong, has influenced how utility companies design customer feedback and incentive programs. Several major utilities including Pacific Gas and Electric and Enbridge Energy have restructured their behavioral energy efficiency programs to avoid positive feedback framing that triggers moral licensing, replacing "you're doing great" messaging with comparative "you're using X% less than your neighbors" framing that activates social comparison rather than self-completion effects.

Cross-Domain Applications: How Incentive Research Travels Across Fields

Incentive research developed in one domain regularly produces insights that transform practice in entirely unrelated fields, reflecting the underlying universality of the mechanisms involved.

Incentives in Software Development: Stack Overflow and Open Source. Stack Overflow, the programming help platform founded in 2008, designed its reputation system as an explicit experiment in social incentive architecture. Co-founder Joel Spolsky described the design process in public posts, drawing directly on behavioral economics research on reputation systems. Users earn reputation points for upvoted answers and questions, with a carefully calibrated system of privileges that unlock as reputation increases. A 2013 study by researchers at Carnegie Mellon University found that the reputation system produced measurably higher-quality answers than comparable platforms without reputation feedback, and that answer quality increased predictably with answerer reputation scores, validating the sorting effect: the incentive structure attracted highly skilled contributors and motivated them to produce high-quality content. By 2023, Stack Overflow had accumulated over 58 million answered questions, representing an enormous commons of technical knowledge produced primarily through social incentives with no financial compensation for most contributors. Research by David Gefen at Drexel University found that Stack Overflow contributors with higher reputation scores reported significantly higher intrinsic motivation (measured via self-determination theory scales), not lower, contradicting the simple crowding-out prediction and suggesting that well-designed social incentive systems can complement rather than replace intrinsic motivation when they align with users' competence needs.

Incentive Design in Climate Policy: Carbon Tax Versus Cap-and-Trade. The long-running debate between carbon taxes and cap-and-trade systems is fundamentally a debate about incentive design, and economists including Joseph Stiglitz at Columbia University, William Nordhaus at Yale, and Nicholas Stern at the London School of Economics have contributed extensive theoretical and empirical analysis. A 2019 study by researchers at Resources for the Future examining twelve existing carbon pricing systems found that jurisdictions with explicit carbon prices (taxes or cap-and-trade) had achieved 20 to 50% larger reductions in electricity sector emissions than similar jurisdictions without carbon pricing, controlling for fuel prices, economic growth, and renewable energy policies. The British Columbia carbon tax, implemented in 2008 at $10 per tonne and rising to $65 per tonne by 2022, provided one of the most extensively studied natural experiments. A 2015 study by economists at the University of Ottawa published in PLOS ONE found that British Columbia's fuel consumption fell by 16% relative to the rest of Canada after the tax was implemented, with no significant effect on provincial GDP growth, suggesting that a well-designed price feedback mechanism can achieve emission reductions without the economic costs that opponents predicted.

Incentives for Whistleblowing: The SEC's Dodd-Frank Program. The Dodd-Frank Wall Street Reform Act of 2010 created the SEC Whistleblower Program, which pays informants 10 to 30% of SEC enforcement actions exceeding $1 million where the whistleblower's information was original and led to a successful enforcement. The program was explicitly designed to overcome the specific incentive failures that had prevented Enron and Bernie Madoff whistleblowers from coming forward: fear of retaliation, inadequate financial compensation for the career risk involved, and lack of legal protection. Between its 2011 launch and fiscal year 2023, the program received over 67,000 tips and paid approximately $1.9 billion in awards to 416 individuals. A 2019 study by researchers at the University of California San Diego, published in the Journal of Financial Economics, found that SEC enforcement actions involving whistleblower tips resulted in significantly larger financial penalties and higher probability of successful prosecution than comparable cases without tips, demonstrating that the financial incentive did increase the quantity and quality of actionable information reaching regulators. The program has since become a model for whistleblower programs in the Commodity Futures Trading Commission, the Internal Revenue Service, and financial regulators in the European Union, UK, and Australia.

Frequently Asked Questions

How do incentives influence behavior?

People respond to rewards and punishments—incentives signal what's valued, create motivation, and shape choices through consequences.

What are types of incentives?

Financial (money, equity), social (status, reputation), intrinsic (autonomy, mastery, purpose), and negative (penalties, consequences).

Why do incentives often backfire?

Narrow focus, gaming, crowding out intrinsic motivation, unintended consequences, or measuring wrong things.

What's the difference between intrinsic and extrinsic motivation?

Intrinsic is internal drive (interest, values); extrinsic is external rewards. Extrinsic can undermine intrinsic for complex tasks.

When do financial incentives work?

For straightforward tasks with clear metrics—less effective for creative, complex, or judgment-heavy work.

What are perverse incentives?

Incentives creating opposite of intended behavior—like bounty for dead rats leading to rat farming.

How do you design good incentives?

Align with actual goals, consider unintended consequences, balance multiple incentives, and maintain intrinsic motivation.

Can you over-incentivize?

Yes—too many competing incentives confuse priorities, and excessive focus on external rewards can destroy intrinsic motivation.

When Notes Fly

Search

Popular Topics