978-0393919684 Chapter 13 Lecture Note

Unlock access to all the studying documents.

View Full Document

CHAPTER 13

Mechanism Design

Teaching Suggestions

A good way to get students to understand the key issues and concepts of this chapter may

be to remind or tell them about Karl Marx’s claims about the communist society he envisioned (in

his 1875 Critique of the Gotha Program [Rockville, Md.: Wildside Press, 2008]): “From each

according to his ability, to each according to his need.” As we know, the reality of communist

states was entirely different. One important obstacle to realizing Marx’s vision was the problem

of information. The government did not know the abilities or needs of individual citizens. In his

own interest, each citizen would pretend to have too little ability and too much need. Nor could

the government ensure that everyone would work to the extent of his ability, especially when it

came to the quality of effort. The government would have to devise mechanisms to incentivize

citizens to reveal ability, exert effort, and tame their claims of need for goods and services. The

most repressive governments did have incentives in the form of sticks at their disposal, but they

could not use them well. Their detection technologies were inaccurate, often relying on

denunciations from fellow workers. Therefore, the incentives became a matter of damned if you

do and damned if you don’t. Get students thinking along these lines, and they will start to see the

role of the price system, outcome-based payments, and other schemes from this chapter and from

their own observations.

The idea of outcome-based payments as incentive schemes is usually self-evident and

students should be able to think of many examples. Discussion of these examples can lead to

more sophisticated points about such schemes including (1) short-run versus long-run effects of

awarding employees and managers the company stock and stock-options, and (2) good and bad

effects of the competition induced by relative performance schemes.

We have conducted successful discussions about incentive payment schemes by raising a

question that actually arose during the production of this book. Both of us had too many

commitments of teaching, research, and administration to do a thorough and careful job of

reading the galley proofs. Therefore, we hired a student to do the work for us. In choosing an

appropriate payment scheme for the student, we realized that several interesting questions of

moral hazard arose. If she were to be paid a flat sum, the student could just sit on the book for a

couple of weeks and then come back and say that it was error free. We could not check this

conclusion without doing the work ourselves, and the full truth might not be known for years

until other readers found errors and informed us. (By the way, please do so if/when you find more

errors.) If paid a piece rate per error discovered, the student might be afraid that the book would

actually be error free; to meet her participation (individual rationality) constraint the piece rate

would have been too high for us to afford. The actual payment scheme had to be a balance

between the two. This example leads nicely into the numerical example of a salary and bonus that

we present in the text. You can then reveal our actual solution—$1 per page plus $1 per error

found—and let them discuss whether the compensation offer was too high, too low, or reasonable.

By the way, the student found more than 300 typos and other errors; if you think that is too many,

try writing a book this long.

This example does not involve some other kinds of moral hazard. For example, the

student cannot “create” errors and get paid for fixing them; that would be a problem if the

typesetter himself were to be paid extra for finding and correcting typographical errors. Adverse

selection is not a serious problem. The quality of the student can be judged on the basis of her

performance in one’s own previous courses or by asking colleagues. And students have built-in

longer-run incentives to perform tasks well for their professors: to ensure the best letters of

recommendation for graduate school or jobs!

The multitask incentive example (teaching and research) in the text argued that the

incentives for each had to be less powerful to avoid excessive substitution of effort toward the

other. But what if teaching and research are complements? Then rewarding research with a more

powerful incentive will indirectly increase the reward to teaching effort because that will increase

the professor’s research productivity and vice versa. This will make it optimal to have more

powerful incentives, not less, for each. You can have a discussion to get the students to recognize

this point, and perhaps also get their views on whether teaching and research are complements or

substitutes. The relationship may depend on the time horizon. On a day-to-day basis more time

devoted to one task leaves less time for the other, but in the long run the research frontier enters

teaching and ideas that arise while teaching become research questions.

The substitutes-complements distinction also has implications for the design of

organizations. If teaching and research are substitutes, it is better to have two separate kinds of

institutions—teaching colleges and research institutes—so the faculty in each can specialize in

one task without any diversion of effort. But if the two are complements, they are better housed

together so the faculty can benefit from the synergy of the two types of effort. You can bring out

this point in class discussion, and extend the idea to examine the best organization in other

contexts: Should firms focus on core competence or be conglomerates? Should computer

operating systems and Web browsers be bundled together? And so on.

You can look ahead to Chapter 16 and get students to think of auctions as mechanisms to

elicit willingness to pay, and ask them to suggest some issues that have to be considered when

devising a good auction form.

Another good but huge topic for discussion is the process of constitution building.

Citizens want to be governed by politicians who (1) have innate skills at detecting and solving the

most pressing problems facing the society, and (2) exert diligence and effort to devise and

implement the institutions and agencies best suited for this purpose. Constitutions should be

mechanisms for the purpose of selecting governments and giving them the appropriate powers

and incentives, subject to the necessary checks and balances. If you were given the task of writing

a new constitution for the United States in the twenty-first century, how would you design it?

A design problem from sports may appeal to some students. If you want to find the best

one out of say a dozen contestants (individuals or teams), while appealing to spectators, and

fitting everything within a season of appropriate length, what are the relative merits of a league

where everyone plays everyone else, perhaps more than once, and a knock-out tournament? In the

light of your analysis, what can you say about the systems that are used in various sports:

baseball, basketball, football, hockey, tennis, golf, soccer, and so on?