It’s obvious, of course, and nothing more than basic behavioral psychology: if you reward a behavior, you will get more of it. Whether in business, sports, entertainment, industry, or education, the assumption is always pretty much the same. It underlies the typical enterprise compensation strategy: foster productivity by rewarding people with bonuses and incentives based on performance, and motivate the pursuit of excellence by offering huge rewards for star performance. Money motivates. Harnessed greed can be useful, but…

Except for one thing: the power of money as a motivator is a myth as old as human endeavor, a myth that achieved a false scientific imprimatur when the results of studies using lab animals doing simple tasks for immediate rewards were generalized inappropriately to complex human behavior. Money may control behavior, but it is, in fact, poor motivator with significant risk of unintended consequences..

What do we actually know about rewards, motivation, performance, and human behavior? A great deal - but not enough. 

Economic and behavioral studies show two very important but also significantly limited ways in which money (or material compensation more generally) functions as an effective motivator:

  • Money (material compensation) works well, and exactly as we were taught and expect, only for simple, mechanical, algorithmic tasks.
  • In the settings where money works, you have to pay (compensate) above a minimal threshold so people think about the job, not about their lack of pay. But you only have to pay enough to take money off the table. 

The genesis of our current understanding of this field dates to the early 1970s and research by Ed Deci and Richard Ryan  (the founders of Social Determination Theory) at the University of Rochester. Their work opened a new window onto the fascinating and complex relationships between rewards and behavior, which they and countless others have now shown to be quite different from what most of us have been taught and what we intuitively but incorrectly believe.

One of their typical early studies, done initially on graduate students but replicated since in a wide variety of settings, involved subjects asked to perform a timed task solving a series of problems of increasing difficulty. One group was paid a flat nominal fee for participation. The other group was promised compensation proportionate to the number of problems correctly solved within the time limit. When the problem solving time had elapsed, the experimenter entered the room, collected the responses leaving the problem sheet, and said he/she would return in 10-15 minutes with their compensation. The behavior of the subject during this waiting period was observed. The majority of those paid a flat nominal fee continued to work on solving the unsolved problems, or reviewed their work for accuracy. The majority of the group to be compensated based on correct solutions paid no further attention to the problems, instead busying themselves with the magazines and snacks in the room. Structured interviews showed that the nominal fee group tended to frame the task as ‘how many problems can I solve correctly?’ while the group compensated based on outcome tended to frame the task as ‘how many problems can I get paid for?’ For the first group, the motivation was solving problems, and the behavior tended to maximize the number solved. For the second group, the motivation was getting paid for the task, and their behavior reflected a limitation on problem solving behavior imposed by the compensation. Further work has shown that the larger the compensation scale, the more prominent the effect. Further, external material compensation based on productivity targets increases the frequency and flagrancy of behaviors such as cheating, lying, and gaming the system. Again, the greater the potential compensation, the more likely one is to see these undesirable behaviors. (Repeatedly seen in headlines about medical care.)

At the time, this was contrary to what was believed and being taught about rewards and motivation, and created quite a stir. Subsequent work by them and many others evolved into theories such as SDT and became important in other branches of behavioral psychology, cognitive neuroscience and behavioral economics.

What have we learned about money and motivation over the last three decades?

Studies from different disciplines (psychology, economics, sociology), in multiple cultures, at all age and educational levels, and with different types of material compensation yield consistent results. For simple, mechanical, non-cognitive tasks, material compensation is the powerful and predictable motivator we expect, and greater compensation results in better performance. As soon as the task becomes more complex, or involves even minimal cognitive, creative or abstract thinking, the expected relationship between reward and performance falls apart. For highly creative, cognitively demanding, or abstract tasks, the standard motivation by extrinsic reward paradigm simply does not apply. In fact, for these higher level tasks - the things we often consider as defining us as human rather than machine or animal -  there is often an inverse relationship between offered rewards and response, with higher compensation impairing behavior and yielding poorer outcomes.

A typical early and often repeated experiments is the following. A group of graduate students are given a collection of tasks to do, ranging from simple physical or mechanical activities (hammer nails, fill glasses with water), through more complex like memorization or simple arithmetic problems, to simple word games, and ultimately to complex problem solving. Various combinations of incentives from one of three levels of reward are offered, based on performance. For simple mechanical tasks like tossing balls into a basket, performance is highly correlated with the size of the reward available. For letter sequence memorization or simple arithmetic tests, there is little difference between the low and moderate rewards, and only a slight improvement with the large reward. For the more complex tasks, there is no difference between groups offered the low or moderate rewards, but the groups offered a high reward demonstrate poorer performance. The positive relationship between reward and outcome is seen where the outcome is measured solely by number or volume, rather than by quality, and where little or no imagination, information processing, or decision making is required. These studies with variations in complexity, setting, population, rewards have been replicated again and again and again.

What’s going on here? It’s complicated and far from fully understood, but it appears based on a variety of behavioral studies coupled with results of functional scanning and TMS (transcranial magnet stimulation) that the parts of the brain associated with complex cognitive processing, creativity, and long term planning can be inhibited by the parts of the brain activated by awareness of extrinsic compensation, and that the inhibition potential is related to the size of the reward. The awareness of rewards tends to activate those cognitive processes associated with short term and mechanistic thinking. In short, the higher the reward, the more likely the person is to think about the reward and the less cognitive and creative resources are available for them to apply to the task. 

Cognitive and creative resources are finite and humans do not multitask. When we think we are multitasking, we are actually serially monotasking, with multiple transitions between the individual monotasks, a process that is associated with delays, loss of focus, dropped information and impaired performance. Ask yourself, do you want your airline pilot to be mentally calculating his pay and benefits for the month while landing your plane in the snow and with cross winds, or your neurosurgeon to be thinking about her planned vacation to Curacao while removing the tumor from your brain? I thought not. It stands to reason - and is supported by research - that presentation of a powerful stimulus (and money is, indeed, a powerful stimulus) reframes and redirects our internal cognitive processes. The more we are aware of the monetary implications of a task, the less we are able to focus on the performance of the task itself. 

If money is not an appropriate incentive, what is? What have we learned about the factors that do reliably motivate humans for complex, abstract and cognitively demanding tasks? A useful framework, adapted eclectically from multiple authors and stealing shamelessly from a schema popularized by the writings of Daniel Pink  (see also his TED talk  and his book for more information), is that there are three key components to effective intrinsic motivation: autonomy, mastery, and purpose (AMP).

  • Autonomy is the drive to direct our own lives. It involves being able to work independently, with control over pace, target, environment, and process. The importance of autonomy to human behavior is not new. Arthur Seligman (sometimes called the father of Positive Psychology with more information here) demonstrated that feeling in control  of one’s life is necessary for durable happiness, and conversely, that being a ‘victim of circumstances’ is highly correlated with unhappiness.
  • Mastery is the desire to experience and demonstrate skill, and skill growth, at important tasks. Watching a child at play illustrates this: there is nothing as productive or self-driven as children learning through independent play and exploration.Csikszentmihalyi’s  research is about the positive response he labels as ‘being in the flow’ that occurs when one feels well prepared, confident on one’s skill level, and subject to arousal by the challenge of a task that can be accomplished with this skill.
  • Purpose is the need to feel part of something bigger, better, or more durable than ourselves. This need for purpose is often manifest as some form of altruism (such as doing good that benefits others, makes a mark on society, lasts after we are gone) or connectedness (such as being part of a social group that transcends either spatial or temporal constraints). This is widely recognized in both academic and non-acaemic literature, and appears in work by such scientists and thinkers as: Brafman and Beckstom (The Starish and the Spider)  in their discussion of Apaches versus Aztec/Mayans, Joseph Campbell  whose works including Monomyth and Hero’s Journey talk about ‘following your bliss’, and David Pearl’s emphasis on finding a Deep Intent. A common business aphorism reflects the core truth of this: “It is not the opportunity to increase shareholder value that gets passionate, creative, dedicated entrepreneurial people out of bed in the morning.”

(Note that there is an irony in the contrast between autonomy-mastery on one hand, which tend to be very self-oriented, and purpose on the other hand, which is almost always other-oriented. There is some very pertinent discussion in writings by psychiatrist and author Anthony Storr about the range of ways humans resolve what he conceptualizes as the universal tension between two primal - and incompatible - needs: independence and connectedness. Read Solitude for more detail and a nice explanation of the normalcy of those who prefer solitude.)

Management approaches developed during the early 20th Century were breathtakingly successful at the time because they were so well adapted for the rote and repetitive tasks expected of unskilled labor in an industrial setting, where the use of piecework and scalar compensation is effective. The United States changed in a little over two generations from a rural and agrarian society with a population of individual independent workers on farms and in family businesses in small towns to a world-dominating economic and manufacturing power with big cities populated by large masses of unskilled but highly productive laborers. 

The only constant is change, and things have changed again. In our present 21st Century information-based work setting where skilled workers do complex tasks in a constantly changing environment, intrinsic motivation is more effective: hire skilled workers, foster individual and local decision making, recognize and support mastery, growth, experiment, and change, and encourage the development of community. Management by rules and protocols will –at  best - result in grudging compliance, but will usually generate performance that is limited to - rather than enhanced by - external goals. For engagement and creativity, self-direction and intrinsic motivation works best.  A quote by Meg Wheatley (author of Leadership and the New Science applying Complex Adaptive Theory to management) is pertinent: “Even if we insist on obedience, we will never gain it for long, and we only gain it at the cost of what we wanted the most: loyalty, intelligence and responsiveness.” 

High end consultants certainly get it.  But the rest of the world is painfully slow to catch on. It is sad, but probably not surprising, that the old - and counterproductive - paradigm of using salaries, incentives and bonuses to drive performance remains so prevalent, despite ample evidence that it doesn’t work in the current culture. (There is plenty of published evidence that pay-for performance does not deliver the expected results in education or medicine, as reported in such places as the NYTimes andBMJ. In general, the health care industry has taken little notice and continues to talk about the great things that P4P will do to reform and streamline medical care. As illustrated in the famous Invisible Gorilla study, people rarely see what they are not looking for.)

I say ‘probably not surprising’ because in my experience, autonomy is deeply threatening to those who believe people will try to do the least they can get away with, mastery is disorienting to those who need to micromanage even what they do not understand in order to support their self-image, and the purpose motive is inexplicable to those who defend power by hiding information and decision making in organizational silos. 

Reflect a moment on what happens where you work? Is it like Atlassian or Google where autonomy and creativity are highly valued, and the workplace is built around fostering them? Or is it more like a 1950s factory, where you are assigned tasks and quotas that are determined by management and communicated by your supervisor, who also serves as your only interface with the rest of the organization, and where feedback and collaboration consist of a Suggestions Box that may or may not be used? How does it feel? How well does it work? And which would YOU prefer?

For me, the answer is clear. I’d like to AMP up my work place.


Peter Elias, MD

Some additional things to read include the following.  If you have other resources, please add them in Comments

A discussion of the Social Determination Theory  

Engaging Leadership by Marlier and Parker is a good book about how this applies to leadership

Other resources that address motivation, leadership, collaboration, and community include:

For information about the changes in management from the 19th to the 21st Century:

Helpful information about behavioral economics.

Resources pertaining to science, human consciousness, human neuroscience and creativity:

Links to more on this topic: