The Joy of Economics:  Making Sense out of Life
 Robert J. Stonebraker, Winthrop University

Important note:  Because I am about to retire, my university plans to remove this site from its server.  As a result, this page might not be available after May 15, 2017.  If you wish to continue accessing the material, feel free to copy it (with appropriate citation) or move it to another site or venue.

Educational Lemons




            He who can, does.  He who cannot, teaches.

                                                                                    .....George Bernard Shaw



            The alleged decline of American education always is a hot topic.  U.S. students are being outperformed by students abroad in a wide variety of fields.  Theories abound, but one oft-heard explanation blames low U.S. teaching salaries.  Low wages create shortages of high-quality teachers, particularly in math and sciences, and the educational process suffers.


            There is some truth to this.  Some teachers undoubtedly are underpaid, especially those in fields for which high-paying non-academic jobs are available.  Most public schools offer one-size-fits-all salary packages; instructors with the same rank and the same experience are paid equally.  A $50,000 salary might seem generous to an art teacher with few alternative occupations.  But if physicists can command $120,000 salaries in the corporate sector, it is folly to expect large numbers to be interested in teaching for $50,000.  If physics teachers are paid no more than art teachers, shortages of physics teachers ultimately will surface.


            Is it fair to pay physics teachers more?  Don't art teachers work just as hard and, therefore, deserve the same pay?  Not necessarily.  Work effort is largely irrelevant.  It is quite possible that the school custodian, who moonlights at the local Holiday Inn to make ends meet, works harder than either the physics or the art teacher. 


            What about equal pay for equal work?  Don't all teachers do the same work and, therefore, deserve the same pay?  Not necessarily.  What matters is the result, not the effort.  Should a student who studies three hours for an exam and earns a 39 percent be "paid" with the same grade as another who studies three hours and scores a 93 percent?  If I train and practice my basketball as hard as LeBron James, should I be paid an equivalent salary?


            In addition, instructors in different fields do different work.  A physics teacher needs different training, different skills, and faces different challenges in getting the material across to students.  More important, the physics teacher faces different opportunity costs.  The typical cost in terms of foregone opportunities of becoming a physics teacher far exceeds that of becoming an art teacher.  Higher wages are essential in such cases.


            What about other fields?  Are teachers underpaid in general?  Would higher salary scales attract better-quality teachers?  Perhaps.  But recent studies suggest that changing the system of teacher compensation might be more effective than changing the level of compensation.


Lemons: a digression


            Many real-world markets are characterized by asymmetric information; that is, sellers have better information than buyers about product quality.  Used cars are an obvious example.


            Imagine that you are in the market for a used car -- a 2002 Honda Accord.  You know that the going price for a "typical" '02 Accord is $4,000, but lack the mechanical sophistication to judge the relative quality of a particular car you find for sale.  The seller or owner will naturally insist that the car is a "creampuff" even if she knows otherwise.  As a cautious buyer unable to judge whether the car is of above or below-average quality, you are not likely to offer more than $4,000 for the car.


            According to economist George Akerlof, sophisticated sellers will know this ahead of time.1  If I own an unusually good Accord worth $6,000, I will understand that potential buyers, because of poor information, will not pay more than $4,000.  As a result, I will either keep the car myself or else sell it privately to a relative or friend who knows (and will pay for) the car's true value.


            On the other hand, suppose I own an '02 Accord that is a real "lemon."  It breaks down every third Tuesday on the freeway and is worth only $1,500.  If you know my car's history, you would never pay more than $1,500.  But you do not know.  There is asymmetric information; you cannot differentiate between the clunkers and the creampuffs.  With a going price of $4,000, I have every incentive to sell, and you, lacking the needed information to scare you off, are willing to buy. In a nutshell, Akerlof argues that when prices are based on average quality, high-quality items will be withheld, and only the low-quality items will be sold in the open market.2  Lemons squeeze out the creampuffs and, in the long run, quality and prices will fall.


            Unless.  Unless sellers can find a way to signal buyers that their products are of higher-than-average quality and, therefore, warrant higher-than-average prices.  What signals might work?  Warranties are a common ploy.  Sellers are less likely to market lemons if they assume all repair expenses.  Building brand-name reputations are another solution, especially in markets with frequent repeat purchases.  Lemons could destroy a reputation and wreck future sales.


Lemons in academia


            What does this have to do with teaching?  Have you caught on yet?   What if teachers are lemons?


            As with used cars, teacher quality is very hard to measure.  How effective was your third-grade teacher?  How can we know? Should we measure comparative student performance on exams?  Not necessarily.  The exams might be biased, some students might be brighter than others, and current knowledge might not reflect long-term retention and use of that knowledge.  Moreover, as a result of what I term the Mileah Effect, the performance of students in one subject might critically depend upon the quality of their teachers in other, seemingly unrelated fields.  For example, students whose math teacher explains concepts especially well will be able to grasp algebra more quickly.  With less need to spend long hours studying math, they can spend more time trying to decipher the poorly crafted lessons of their less-than-competent history teacher.  If so, their improved history performance will owe more to the abilities of their math teacher than their history teacher. And what about intangibles?  How important is the classroom enthusiasm a teacher might generate?  How might it impact future learning?  How can it be measured?


            Since administrators will inevitably make mistakes in judging performance, teachers have resisted efforts to base their salaries on questionable measures of merit or productivity.  Rather than risk a mistaken assessment of their talents, they opt for a safer system that pays them uniformly. Potentially strong teachers do try to establish brand-name reputations by earning degrees from prestigious institutions, but this has little impact.  In a world of union-scale salaries, a teacher with an Ivy League pedigree can be paid no more than one with a degree from the proverbial Podunk State. 


            And we are right back at Akerlof's model.  If teachers are paid uniformly on the basis of average productivity, those with above-average abilities will be underpaid.  Where possible, the most-proficient instructors will relocate to occupations with merit-based pay scales that will appropriately reward their superior talents.  On the other hand, "lemon" teachers, knowing that they are overpaid, will be content to continue teaching.  As in the used-car market, the best products/people will leave.  Both teacher quality and their long-run price (or wage) will fall.  Economists John Barron and Mark Loewenstein do find supporting evidence.3  Average wage rates are higher in occupations with merit-based pay than in occupations, like teaching, with little or no merit pay.  They conclude that the current one-size-fits-all pay structure:


            ...not only causes teachers to reduce their work effort, but it also results in a sorting of high ability workers out of the teaching occupation.  One consequence is lower earnings in the teaching occupation.  A second is less capable teachers.4


            If this analysis is correct, raising the level of pay across-the-board will not solve the problem.  As long as teachers resist merit pay, the lemon model will remain applicable.  Moving to a system, albeit an imperfect one, of higher pay for better performance, might be more effective.





1.         Akerlof, George A., "The Market for Lemons: Qualitative Uncertainty and the Market Mechanism," Quarterly Journal of Economics, volume 84, 1970, pp. 488-500.

2.         Think of another example.  Many people are reluctant to look for mates at singles bars or through match-making services.  They reason that the best potential mates, the ones who are worth more than the average, already have taken themselves off the open market and are marketing themselves privately. Those who remain in the open markets of bars and dating services are more likely to be lemons.        

3.         Barron, John and Loewenstein, Mark, "On Imperfect Evaluations and Earning Differentials,"  Economic Inquiry, volume 24, October 1986, pp. 595-615.

4.         ibid., page 613.



Testing Yourself


To test your understanding of the major concepts in this reading, try answering the following:


1.         Critique the arguments that all teachers should be paid the same salary because they work equally hard and do the same job.

2.         Explain the concept of asymmetric information and give an example.

3.         Explain why lemons might dominate the used-car market.

4.         Use Akerlof’s lemons model to teaching and explain the pros and cons of merit pay for teachers.


Permission to reproduce or copy all or parts of this material for non-profit use is granted on the condition that the author and source are credited.  Suggestions and comments are welcomed.

Return to Contents

Last modified 04/15/17