THE MODERN FIRM

Laura Pagani

FIRST TERM 2021/2022

MSc Marketing e Mercati Globali

Track Global Management

1
About myself

Laura Pagani laura.pagani@unimib.it

Office hours

Syllabus and course material:

https://elearning.unimib.it/course/view.php?id=38895

2
About yourself…

3
 ABOUT THIS COURSE
Prerequisites
Microeconomics (math and statistics help, too)

Textbook
Edward P. Lazear and Michael Gibbs, Personnel Economics in Practice, 3th Ed., Wiley

For some topics, it will be complemented by Garibaldi, P., Personnel Economics in imperfect labour markets,
2006, Oxford University Press

WARNING: the textbook is required, it is NOT enough to use my slides in preparation for the exam

4
 ABOUT THIS COURSE

Exam:

Written individual exam

Partial exams Midterm: 22/11/2021 - 26/11/2021 – Final: January 2022

General exam Official dates

5
 ABOUT THIS COURSE

How to do well in this course

«Attend» regularly asynchronous classes

Participate actively during classes: ask questions

Study along the way: do not procrastinate

Read the relevant chapters before class

Use «office hours»: I’m here to help

6
 THE MODERN FIRM - OVERVIEW

Performance of organizations (and economies) result of actions of many individuals, that

combine to create innovation, economic growth, job opportunities …

Importance of understanding how modern firms organize themselves and manage their
employees

Considering:
Importance of human resources to many organizations’ success and to the world economies (70% of world-
wide wealth is in the form of human capital)

Labor accounts for the majority of business costs (in large corporations 3/4 of total costs)

focus on personnel economics, that is, the economic theory applied to traditional topics in
the study of human resources management within the firm
7
 THE MODERN FIRM - OVERVIEW
Interesting questions our course aim at answering:
Which applicant should be hired?
Should the firm promote an internal candidate or hire an outside applicant?
How can the manager encourage the worker to do her best work?
When do firms choose to use one form of compensation over another?
When are teams important?
The list extends to any decision an employer has to make with respect to its dealings with
employees
While many managers solve these problems using gut instinct (typically some combination of
experience and personal preference) in this class you will learn how to recognize simple models
of human behavior and apply them to these and other human resource problems

Vast variety of employment relationships such as:

tenured government employees with fixed salaries
bonus-driven sales people
completely independent contractors
…
8
 THE MODERN FIRM - CONTENT
PART I: Sorting and Investing in employees
Setting hiring standards (L&G Ch.1)
Recruitment (L&G Ch.2)
Investment in skills (L&G Ch.3)
Managing turnover (L&G Ch.4)
PART II: Job design
Team work (L&G Ch. 8)
PART III: Paying for performance
Performance evaluation (L&G Ch.9)
Optimal compensation schemes (Garibaldi, Ch.6)
Pay for performance with wage constraints (Garibaldi, Ch.7)

9
 PART 1: SORTING AND INVESTING IN EMPLOYEES
Employees bring to the workplace both
Innate abilities (e.g. ability to work with numbers or to interact with colleagues)
Acquired abilities (through education, experience, training…)

The topics of this part are:

1. How to sort employees by

their innate or accumulated
2. How to invest
skills
further in their
skills

3. How to manage their

exit from the organization

Several important economic concepts are introduced, such as

asymmetric information, optimal investment, labor markets
constraints on firm policies and different methods of contracting
 CHAPTER 1 – HIRING STANDARDS

After completing this chapter, you will be able to answer questions such as:

1. How many workers should my firm hire?

2. When should I hire low wage workers over higher wage workers?
3. Should the skills of my workers change if my firm increases its investment in capital equipment?
4. What are the strategies for hiring risky workers?

11
AN EXAMPLE: HIRING RISKY WORKERS
Position to fill in a London investment bank
Salary = 100K; will work for the firm T years

Two candidates
Gupta predictably (standard background: degree in economics, MBA with focus in finance, few
year experience as a financial analysist…) produces 200K per year
Svensen (unusual background: appears talented but has few years of experience) may be a “star”
(50%), producing 500K; or a “lemon” (50%), losing 100K
Ignore discounting for simplicity
Assume the firm is risk neutral

Expected output from Svensen:

0.5 × 500𝐾 − 0.5 × 100𝐾 = 200𝐾

Equal expected productivity. Which is a better hire?
12
 Hire Gupta Hire Svenson

Net = 100K x T Net = 400K x T Net = -200K

Expected net = (2T-1) x

100K

Hiring Svenson better if

(2T-1) x 100K > 100K x T

Hiring risk isn’t always bad, even if expected values are equal!
A risky hire is a “real option” if firms can limit downside risk by firing “lemons”
13
This simple model suggests other factors that are important in deciding whether to hire a risky
worker:
Downside risk: the more the potential there is for an employee to destroy value, the less
likely it is to be optimal to take a chance on a risky worker
Upside potential: the higher the profit if Svenson turn out to be a star, the greater the option
value from a risky hire (creative jobs with high upside potential)
Termination cost: the more costly it is to fire a worker, the more costly is a risky candidate
(legal or social restrictions can make the option of firing a worker after one year costly)
Risk aversion: the higher the risk aversion of the firm, the lower the option of hiring a risky
worker
Length of evaluation: the time it takes to evaluate whether Svensen is a star or a disaster
affects the value of hiring a risky candidate
Length of employment: the value of a risky hire will usually be larger the younger the new
hire and the lower the turnover of the company (so that employees tend to stay with the
firm longer)
14
If the firm can pay stars less than their high productivity… but the market value of stars
should (eventually) rise risky hires profitable if:

Switching jobs costly to workers

Stars are not easily identified by competitors (asymmetric information)

Productivity is firm-specific

15
SETTING HIRING STANDARDS
Before it actually begins recruiting employees, the firms has to establish its hiring standards
Labor is not an homogeneous factor of production: highly skilled worker different factor of
production with respect to unskilled worker

Balancing Benefits Against Costs

Is it optimal when hiring to obtain the best quality workers?… The most productive workers are also
likely to be the most expensive. Should the firm hire least expensive workers?
A trade off emerges and the firm has to find the right balance between cost and productivity (optimal
skill ratio)
In general, the best skill composition is not necessarily the cheapest nor the most productive
composition
Technological considerations are also very important: In some firms, only one type of labor will be
chosen, while in other type of firms it will always be optimal to have a positive quantity of both type
of skills

16
KEY CONCEPTS FOR FORMAL ANALYSIS
Q=firm output
H=high school graduate labor; wage WH
C=college graduate labor; wage WC
𝑊𝐶 = 𝑊𝐻 (1 + 𝛾) 𝛾 =skill premium
The firm‘s objective is to maximize profit; it uses two types of labor to produce a given output Q with
a given amount of capital
𝑄 = 𝑓 𝐻, 𝐶, 𝐾
Assume
no constraints on the firm's ability to hire as much labor as it desires
the price at which the firm sells its output and the price per hour it pays employees are constants

Marginal productivity of labor is the change in output resulting from hiring an additional worker
holding constant the quantity of other inputs:

∆𝑄
𝑀𝑃𝐻 = = 𝑓𝐻
∆𝐻 𝐶=𝐶, 𝐾=𝐾

∆𝑄
𝑀𝑃𝐶 = = 𝑓𝐶
∆𝐶 𝐻=𝐻, 𝐾=𝐾
17
ISOCOST

𝑊𝐻 𝐻 + 𝑊𝐶 𝐶 = 𝑇𝐶

𝑇𝐶2
𝑇𝐶1
𝑊𝐻
slope −
𝑊𝐶

18
ISOQUANT
Combination of H and C that yields the same amount of output
𝑓 𝐻, 𝐶, 𝐾 = 𝑄

Three key properties: C

1. downward sloping
2. do not intersect
3. higher isoquants: higher levels of output

DC

DH H

Moving from x to y:
∆𝐶 × 𝑀𝑃𝐶 +∆𝐻 × 𝑀𝑃𝐻 =0

∆𝐶 𝑀𝑃𝐻
slope 𝑚𝑎𝑟𝑔𝑖𝑛𝑎𝑙 𝑟𝑎𝑡𝑒 𝑜𝑓 𝑡𝑒𝑐ℎ𝑛𝑖𝑐𝑎𝑙 𝑠𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑜𝑛 MRTS = =−
∆𝐻 𝑀𝑃𝐶
19
THE METHOD OF PRODUCTION

Three scenarios representing different approaches to production, to see what effect the method of
production has on our analysis

1. Productivity Depends on Co-Workers

A manager describes her production as follows:

«This firm manufactures small appliances. We find it better to have a combination of worker types. High
school graduates are cheaper and more cost-effective in the short run, but we find that we can’t keep their
skill levels up without some college graduates around. The high school graduates forget what they knew.
The college graduates keep the high school graduates sharp. So we like to have both kinds of workers.
The problem is that I’m not sure about the appropriate balance»

Here workers interact with one another: college graduates affect the output of high school graduates,
and vice versa

20
PRODUCTIVITY DEPENDS ON CO-WORKERS (AN EXAMPLE)

The larger the number of high school graduates in the workforce, the higher the value of adding college
graduates to the workforce
Similarly, the more valuable that high school graduates are, the more college graduates are employed
In this case the firm wants a balance of college and high school graduates

When workers interact on the job, a worker’s contribution to output includes the effect on co-worker
output

21
GRAPHICAL SOLUTION

Solution is given by the tangency:

𝑤𝐻 𝑓𝐻
= (relative marginal cost = relative marginal product)
𝑤𝐶 𝑓𝐶

H
FORMALLY
The firm will choose the labor composition that minimizes its labor costs 𝑊𝐻 𝐻 + 𝑊𝐶 𝐶 for
given output Q:

𝑀𝑖𝑛𝐻,𝐶 𝑊𝐻 𝐻 + 𝑊𝐶 𝐶

𝑠. 𝑡. 𝑓 𝐻, 𝐶, 𝐾 = 𝑄

𝐿 = 𝑊𝐻 𝐻 + 𝑊𝐶 𝐶 + 𝜆 𝑓 𝐾, 𝐻, 𝐶 − 𝑄
FOC:
𝑊𝐻 = 𝜆𝑓𝐻
𝑊𝐻 𝑓𝐻 𝑊𝐻 𝑊𝐶
𝑊𝐶 = 𝜆𝑓𝐶 = 𝑜𝑟 =
𝑊𝐶 𝑓𝐶 𝑓𝐻 𝑓𝐶
𝑓 𝐾, 𝐻, 𝐶 = 𝑄

This example illustrates two simple but important economic principles:

1. Always think in terms of tradeoffs between costs and benefits: in this example, the desire for high
quality workers must be balanced against their higher cost
2. Always compare your approach to your best alternative
EXAMPLE: COBB–DOUGLAS PRODUCTION FUNCTION
𝑄 = 𝐻𝛽 𝐶 𝛼
𝑀𝑃𝐻 = 𝛽𝐻𝛽−1 𝐶 𝛼
𝑀𝑃𝐶 = 𝛼𝐻𝛽 𝐶 𝛼−1
the marginal product of each type of labour depends on the number of people of the other type that are
currently hired
The optimality condition is

𝑀𝑃𝐻 𝛽𝐻𝛽−1 𝐶 𝛼 𝛽𝐶 𝑤𝐻
= = =
𝑀𝑃𝐶 𝛼𝐻𝛽 𝐶 𝛼−1 𝛼𝐻 𝑤𝐶
And the optimal skill ratio is
∗
𝐶 𝛼 𝑤𝐻 𝛼
= =
𝐻 𝛽 𝑤𝐶 𝛽(1 + 𝛾)

The optimal skill ratio increases with 𝛼 and decreases with 𝛾 and 𝛽
2. PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS
A second manager describes production in his unit as follows:
«My team is a sales force. Each salesperson works independently. The organization consists of my salespeople and
me. What kind of worker should I hire?»
Each worker’s sales depend on his own ability and effort, irrespective of the efforts of other salespeople

AN EXAMPLE

College graduates are about 28%

more productive than HS graduates
College graduates cost more than
HS graduates

25
Both high school and college graduate profitable to employ, but apparently more profitable to employ a
single college graduate compared to a single high school graduate:

Monthly profit from hiring HS graduate = 122,917 – 2,254 = 120,663,

Monthly profit from hiring college graduate = 156,944 – 3,593 = 153,351.

However…
Suppose your firm wants to hire enough workers to produce 1 unit of output ($1 million in monthly
sales)
This would require 6.4 college graduates (1,000,000/156,944), at a cost of $22,995, or 8.1 high
school graduates (1,000,000/122,917), at a cost of $18,257
High school graduates more profitable because high school graduates have lower cost per unit of
output

26
PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS (FORMAL ANALYSIS)
Assume productivity of each type of labor is constant: with 𝑎 > 𝑏 𝑎𝑛𝑑 𝑎 = 𝑏(1 + 𝛿)
𝑀𝑃𝐶 = 𝑎
𝛿 𝑖𝑠 𝑡ℎ𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝑝𝑟𝑒𝑚𝑖𝑢𝑚
𝑀𝑃𝐻 = 𝑏
Two cases are possible:

𝑊𝐻 𝑊𝐶
1. > → optimal hiring only college graduate labor
𝑏 𝑎

𝑏(1 + 𝛿) 𝑊𝐻 (1 + 𝛾)
>
𝑏 𝑊𝐻

𝛿>𝛾
𝑊𝐻 𝑊𝐶
2. < → optimal hiring only HS graduate labor
𝑏 𝑎

𝑏(1 + 𝛿) 𝑊𝐻 (1 + 𝛾)
<
𝑏 𝑊𝐻
𝛿<𝛾

With independent workers only one type of labor should be used and the solution is obtained by
comparing the value of the wage skill premium with the value of the productivity premium 27
FOREIGN COMPETITION
Analysis useful for thinking about globalization of labor markets and the role of foreign competition
Often argued that countries with low labor costs drive companies in countries with high labor costs out of
business. Is that accurate?

The real issue is not whether labor is cheaper, but whether it is more cost-effective: cheap labor is not
necessarily low-cost labor
Similarly, high-productivity labor is not necessarily the most profitable labor
Should seek low cost per unit of output, whether that arises from low wages, high productivity, or both 28
3. PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS, BUT DEPENDS ON CAPITAL

A third manager describes the production process as follows:

«We are a large clothing company that has our men’s dress shirts produced by a factory in Malaysia.
Each worker uses a sewing machine, which costs us $7.50 per day to rent. We can use skilled labor,
which produces an average of 4 shirts per day, or professional labor, which produces an average of
6 shirts per day. Skilled labor costs $7.50 per hour, and professional labor costs $12 per hour. The
sewing machine company says that it will rent us a new machine that doubles output per worker, but
the better machine costs $16.50 per day to rent. Should I rent the new machine? What kind of labor
should I hire?»

Producing a shirt involves both machines and labor

29
PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS, BUT DEPENDS ON CAPITAL (AN EXAMPLE)

Using the new machines, the firm should hire professional rather than skilled workers because when the new
machines are used, the cost per shirt is lower with professionals than with skilled labor: when expensive
capital is employed, it may be cost effective to use it intensely

30
PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS, BUT DEPENDS ON CAPITAL (FORMAL ANALYSIS)
Assume the cost of capital is Ck
Only college graduate labor should be hired if: Cost of 1 unit of labor with a capital machine

𝑊𝐶 + 𝐶𝐾 𝑊𝐻 + 𝐶𝐾
<
𝑎 𝑏 Productivity of 1 unit of labor L

𝑊𝐻 (1 + 𝛾) + 𝐶𝐾 𝑊𝐻 + 𝐶𝐾
<
𝑏(1 + 𝛿) 𝑏

𝑊𝐻 1 + 𝛾 + 𝐶𝐾 < 𝑊𝐻 1 + 𝛿 + 𝐶𝐾 (1 + 𝛿)

𝑊𝐻
𝛿>𝛾
𝑊𝐻 + 𝐶𝐾

May be optimal to hire college graduate labor even if 𝛾 ≥ 𝛿 (this depends on how costly machines are
and on how costly H is)

31
 PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS, BUT DEPENDS ON CAPITAL (AN EXAMPLE)

Productivity premium
d:
6 =4(1+d) d=0.5
12=8(1+d) d=0.5

Wage premium g:
96=60(1+g) g=0.6

1) Old machine:
δ = 0.5
𝑊𝐻 60
𝛾 = 0.6 = 0.53 ⇒ ℎ𝑖𝑟𝑒 𝑜𝑛𝑙𝑦 𝐻
𝑊𝐻 +𝐶𝐾 60+7.5

2) New machine:
δ = 0.5
𝑊𝐻 60
𝛾 = 0.6 = 0.47 ⇒ ℎ𝑖𝑟𝑒 𝑜𝑛𝑙𝑦 𝐶
𝑊𝐻 +𝐶𝐾 60+16.5

32
PRODUCTIVITY IS INDEPENDENT OF CO-WORKERS, BUT DEPENDS ON CAPITAL

Professionals use the machines more efficiently, which leads us to conclude that a firm should
improve the quality of workers that it employs as it increases the amount or quality of its
capital stock. More specifically, the optimal level of skill rises as the use of capital relative to
labor increases

This helps explain why the president of a firm should be very highly skilled: His or her labor is
combined with the entire capital stock of the firm, in a sense: it makes no sense to waste the
capital by placing it under the stewardship of a low-skilled individual
We will see that the labor market has valued highly skilled workers relatively more over time;
one explanation for this is that firms have made increasing use of valuable, and very
productive, capital in the form of new information technology

33
HOW MANY WORKERS TO HIRE?

The firm should continue to hire workers so long as the incremental profit from hiring an additional
worker is positive

This approach implies that there is a limit to the number of workers the firm should hire, because of
the principle of diminishing marginal productivity: As workers are added to an organization, the value
of an additional one falls. Why?

The main reason is that workers are combined with other resources: computers, machines, your time
as their manager, etc.

The more workers that you hire, holding other resources fixed, the more thinly are those resources
spread across each worker

34
HOW MANY WORKERS TO HIRE? (AN EXAMPLE)

As more workers are hired, the marginal productivity (extra sales) from each additional worker declines
The firm should hire workers up to the point where they are no longer profitable; that is, when marginal
productivity is less than or equal to marginal labor cost

The general result is that profit is maximized by using any resource, including employees, up to the point
where the marginal benefits just equal the marginal costs
35
OTHER FACTORS

AVAILABILITY OF WORKERS
In many communities, more high school graduates than college graduates are available

This does not mean that a firm should hire high school graduates as they are cheaper because most
employers (even very large ones) employ a small part of the local labor force, so the total availability of
workers is irrelevant

One exception is when a firm employs a very large fraction of the local labor force and hiring more of a
certain type of worker drives up the wage

Another exception is when the type of labor being hired is very specific, the market for it may be thin
(there are few buyers for this type of worker). If so, there may be significant search costs to finding a
worker with the right skills and the wage must build in these amortized search costs

36
OTHER FACTORS

THE FIRM’S FINANCIAL CONDITION

Suppose the firm is in financial distress. How should this affect its hiring decisions?

None of the analysis above makes mention of the firm’s financial condition and choosing the
wrong kind of labor will only make the financial condition worse

A firm in financial distress may have trouble paying employees because of cash flow
problems; however, this is a financial problem, not a labor problem

The best solution to such a problem would be to arrange financing to cover the short-term
cash flow problems, so that the firm can hire workers when it is profitable to do so

In fact, creditors should encourage this, if it increases profits, since it makes it more likely that
the debt will eventually be paid off.

37
 CHAPTER 2 – RECRUITMENT

How firms bring employees into the organization and patterns of careers they have once they are there

After completing this chapter, you will be able to answer questions such as:

1. How can I reduce interviewing and hiring costs for my firm?

2. Should credentials be required when hiring?
3. How would I implement a probationary hiring system?

1
 INTRODUCTION
CAREER PATTERNS IN ACME INC

 Management ranks have 8 hierarchical levels, from entry (Level

1) to CEO (Level 8)
 Most management employees in the first 4 levels
 At Levels 2-8, most employees were not hired from the outside
but promoted from within  managers in upper levels have
substantial experience in the firm on average
 Movement between levels (promotion) is more rapid at lower
levels

Two patterns:
1) Many leave Acme very quickly after being hired  evidence of sorting in the first few years on the job

2) If employees survive the sort, they often enjoy careers at the company that last for many years 
value to having employees remain with the firm

2
 SCREENING JOB APPLICANTS

 Once your firm has decided which types of workers to hire, it must recruit for those types
 Two general issues:
 how weed out undesirable applicants
 how attract the right types of applicants

 The problem arises because of asymmetric information: One party knows what type they are (in this case, a
high or low quality job candidate), and the other does not
 If the firm offers a wage equal to the average productivity, only the wrong kinds of workers are attracted to the
firm  adverse selection

 A number of approaches can be used to mitigate the problem of adverse selection in recruiting

3
 EFFICIENCY WAGES

 One strategy for attracting good quality job applicants: offer high level of pay or benefits:
 Assumption: reservation wage function of ability (better outside options) 
 Low wage: only low-productivity applicants
 High wage: larger pool of applicants and higher average quality

Example:
L (20%) M (40%) H (40%)
Reservation wage 16k 21k 24k
Productivity 32k 44k 56k

1) w=16+ε  only L profit per worker Π=32-16=16

! $
2) w=21+ε  L and M profit per worker Π=" (32-21)+" (44-21)=19
! $ $
3) w=24+ε  all types profit per worker Π=& (32-24)+& (44-24)+ & (56-24)=22.4

 Higher wage related to higher profits (wage is not only a cost)

4
 EFFICIENCY WAGES
 Generalizing with two types of workers:
High productivity: Low productivity:
 Productivity MPH  Productivity MPL < MPH
 Reservation wage wH  Reservation wage wL< wH
 Share qH  Share 1-qH

 Profits if w=wL  ΠL= MPL-wL

 Profits if w=wH  ΠH= qH(MPH-wH)+(1-qH)(MPL-wH)
 More profitable wH if
ΠH > ΠL
qH(MPH-wH)+(1-qH)(MPL-wH)>MPL-wL
qH(MPH-MPL)>wH-wL
 wH more profitable:
 the higher the share of highly productive workers
 the higher the productivity gap
 the lower the reservation wage gap

 Drawback: hired also low productivity workers (lower profits compared to no asymmetric information, see example)
5
 CREDENTIALS
 An approach is to look for credentials (e.g. job and promotion history, type of training, quality of school
attended) that distinguish some applicants from others
 What makes a credential useful for hiring?
1. Informativeness of the Credential
 positive correlation between ability to obtain the credential and ability to perform well on the job

2. Cost of Obtaining the Credential

 it should be easier for well-qualified workers to obtain compared to poorly qualified workers
 on the other hand, a credential extremely expensive for all workers few applicants will have it (also
well-qualified)
 For a credential to be effective, it must be that most qualified applicants possess the credential
while most unqualified ones do not
3. Return on investment in the credential
 If the difference in wages between those who have credentials and those who do not is not very
great, small differences in credentials will signal large differences in ability

6
 SCREENING

 Suppose that you selected a subset of candidates with appropriate credentials

 What should you do next? It make sense to expend some resources to screen them further

 There are a variety of methods that firms use to screen applicants: tests to see how they perform on
specific tasks, psychological profiling, personal interviews…

 All of these examples involve some costs

 To what extent your firm should invest resources in screening applicants carefully?

7
EXAMPLE: Screening Bankers

TYPE
A B C D E
% of job applicants 10% 20% 40% 20% 10%
Investment bank -250 0 125 200 450
Productivity
Commercial bank 95 100 110 120 125

 Each bank expects to pay 100,000

 Investment bank would want to avoid A and B types
 Commercial bank would want to avoid A types

 No screening:
 Both banks have 110,000 average productivity from each new hire and average profit of 10,000

8
EXAMPLE: Screening Bankers
 Applicants can be put through a series of tests that cost 2,000 per person and give definitive information
on which type the applicant is. How valuable is such information?

 Investment bank
 Would reject A and B, and accept 70% of all applicants
 Average productivity of C, D and E hired would be about 193,000
 Screening cost per worker actually hired would be 2,000 x 10/7 (2,857 per hire)
 Average profit from each new hire would rise to about 90,100  profit greatly from screening applicants

 Commercial bank
 Would reject A
 Average productivity would rise to about 112,000
 Screening cost of 2,000 x 10/9 (2,222 per hire)
 Profit per new hire would fall to about 9,800  not benefit from screening
9
EXAMPLE: Screening Bankers

 Why the difference?

1. Investment bank screen out three times as many workers as the commercial bank. The point of
screening is to avoid hiring the applicants who would not be profitable
2. Downside from hiring poor candidates worse at the investment bank; more at risk from hiring the
wrong type of worker

 In general:
 Screening is more profitable when the test is more effective: cheaper to administer, more accurate
(correctly distinguish between desirable and undesirable job applicants), more discriminating (it weeds
out a higher fraction of candidates, recommending a smaller fraction for hiring)
 Screening is more profitable when the stakes are higher: the greater the downside risk from hiring the
wrong person, the more value there is to screening. Similarly, the longer that a new candidate can be
expected to stay with the employer, the more valuable will be the screen

10
SCREENING JOB APPLICANTS (FORMAL ANALYSIS)

 Two types of job applicants:

 E (productivity QE, probability p)
 D (productivity QD, probability 1-p)
 QE > QD

 The firm pays wage W to those it hires such that QE > W > QD
 profit from E types, but loss from D types

 Expected profit from random new hire:

! "# − % + (1 − !) "* − %

 The firm has a screen available that costs s, with accuracy q

 Expected profit with screening
!+ "# − % + 1 − ! 1 − + "* − % − ,

11
SCREENING JOB APPLICANTS (FORMAL ANALYSIS)
 Change in profits from screening compared to not screening:

∆ #$%&'( = −# 1 − , -. − / − 1 − # , -0 − / − 1

 First term negative (loss from mistakenly rejecting candidates of type E)

 Second term positive (gain from appropriately rejecting D types)
 Third term is also negative

2∆ #$%&'(
>0
2,
2∆ #$%&'(
<0
21
2∆ #$%&'(
<0
2#

 The test is more effective when it is

 More accurate
 Cheaper
 More discriminating

12
SIGNALING

 Screening methods described above may be useful, but are imperfect: only proxies for how the person actually
performs the job

 In some cases, firms may use signaling in addition to screening

 Workers generally have a good idea about their skills (work ethic, ambition, …)

 If workers share this information honestly with employers, a firm could recruit employees of a certain type

 Not an effective approach…

 Suppose that through screening the bank to weed out Types A through C easily, but harder to distinguish between
Types D and E (the bank would like to hire E)
 Instead of screening, it can construct a job offer involving probation, up-or-out promotion, and a raise on
promotion that is attractive to the E types, but not to the D types
13
SIGNALING (EXAMPLE)

 Two year contract, wage W1 and W2

 The bank can figure out what type an employee is after observing them on the job for one year

 The accuracy of this judgment is not perfect: 10% of the time, the wrong decision is made
 10% of D types are promoted when they should not be
 10% of E types are not promoted when they should

 In other jobs:
 D types can earn 175,000 (350,000 for two periods)
 E types can earn 200,000 (400,000 for two periods)

4. Selezione avversa e reclutamento 14

 Value of applying for D = !" + 0.9 ( 175+0.1 ( !,

Value of applying for E = !" + 0.1 ( 200+0.9 ( !,

 Probation can generate good self selection if the firm pay a sufficiently low amount during probation, and a
sufficiently high amount after probation
 The firm is demanding that each applicant post a bond (accepting less than they could earn elsewhere) during
probation and in return if they perform well and are promoted the firm will give them a reward (by paying them
more than they could earn elsewhere)

15
 Type E receive a smaller reward on promotion, and pay a larger cost during probation, than do type D:
 Up-front bond (W – W1) larger for E, since their outside alternative is larger
 Deferred reward (W2 – W) smaller for E

 How can this type of job offer deter them from applying D, while motivating E types to apply?

 The evaluation must result in sufficiently high probability that E types will be promoted, and sufficiently low
probability that D types must be promoted: the low success rate for D types reduces the expected value of the
job for them compared to E types.

16
 This discussion illustrates the general economic idea of signaling:
 High quality type signals his or her type to the market by incurring a cost
 If low quality types are not willing to incur this same cost, then the signaling is effective: the fact that
someone incurs the cost proves that they are the high quality type

 In our employment example, Type E’s can signal their type (and confidence in their ability to perform well
and earn promotion) by their willingness to accept low pay in the first period

 This only works if the D types are not also willing to accept the same contract

 Another application of signaling in the business world is when venture capitalists demand that entrepreneurs
invest all of their family’s personal funds in a new business venture

 This helps the venture capitalist separate out the most confident and serious candidates from the least

10/19/2021 17
SIGNALING AND PROBATION (FORMAL ANALYSIS)

 We now incorporate signaling into the probation model described above and show how the wage must be
structured each period to ensure signaling

Assumptions
 There are types E and D, as defined before
 The firm offers W1 and W2 in two periods
 In period one, workers are observed on the job
 those deemed to be good fits are promoted and paid W2
 the rest are fired and earn their outside pay
 q is the accuracy of the promotion decision

 Outside alternatives respectively WE and WD with WE > WD

10/19/2021 18
SIGNALING AND PROBATION (FORMAL ANALYSIS)

 In order to deter D types, but attract E types, we must meet both of these conditions

!" + 1 − & !' + &!( < 2!( (1)

!" + &!' + (1 − &)!. > 2!. (2)

.
 (1) says that D types expect to do worse at this firm
 (2) says that E types expect to do better

 This scheme can induce self selection if:

!" < !( + 1 − & !( − !' < !(
And
!. − !"
!' > !. + > !.
&

19
SIGNALING AND PROBATION (FORMAL ANALYSIS)
 In fact, the optimal wages (that minimize compensation cost) are:
1 − '(
!" = !$ − (! −!$ )
2' − 1 +

2−'
!( = !+ + (! −!$ )
2' − 1 +

 These imply that W1 < WD < WE < W2 (deferred pay)

 It is easy to see that:
 The more accurate the test (larger q), the larger is W1, and thus the smaller is the bond posted by E types
 W2, and thus the reward on promotion, is smaller the more accurate is the test (E types more willing to accept since there
is less chance of a mistake)
 The smaller is WD, the smaller is W1
 The larger is WE, the larger is W2
 The larger the differences in productivity between the two ((!+ −!$ ), the larger the reward on promotion

20
 PERFOMANCE-RELATED-PAY AS A SCREENING MECHANISM

 Performance-related-pay is an alternative approach that may induce candidates to self-select

 Intuition: high/low productivity candidates will/will not accept because they will be paid a high/low wage
 2 types:
High productivity: Low productivity:
 Productivity yH  Productivity yL < yH
 Reservation wage wH  Reservation wage wL< wH

 If production is measurable, the firm pays w=by

 Through the right choice of b the firm can attract/weed out the H/L types
 H applies if byH>wH --> b>wH/yH
 L does not apply if byL<wL --> b<wL/yL
 Necessary condition: wL/yL>wH/yH
Example

L (20%) H (40%)
b x 56 > 24  b>0.43
Reservation wage 16k 24k b x 32 < 16  b<0.5
Productivity 32k 56k 0.43<b<0.5

 Problem: measurability of production

21
EDUCATION AS A SIGNAL

 Assume students learn nothing useful in school but more talented students find it easier to learn the
material quickly

 Then they might be able to signal their talents to the labor market by investing in more education than less
talented students

 The labor market recognizes this, paying more to those who have obtained more schooling (indeed those
with more schooling do earn more)

 There is evidence that education does have some role in screening workers, for example, those who almost
complete four years of college earn less than those who go a little bit further and complete their degree
(formal credential), hard to explain by training alone

10/19/2021 22
SIGNALING (FORMAL EXAMPLE)

 Junior accountants can invest in some education or on-the-job training

 If they complete the training, they become Certified Public Accountants (CPAs)

 There are two types of accountants:

 “quicks” (fraction α), more productive and find it easier (less cost) to become CPA
 “slows” (fraction 1–α)

!# = employee′s productivity
3# = 4678 69 6586<7<78 8ℎ: 3;< 4C:=:78<6>
< = F, 7

 Labor market pays accountants their expected productivity  if cannot tell the two types of accountants apart
pay to all
!@ = A!B + (1 − A)!C
NB:
!@ < !B
D
!@ > !C

 Adverse selection
23
 If the quicks do succeed in distinguishing themselves, they will be paid their productivity

 Those who do not signal will then be assumed to be slows, and paid Qs

 In order for signaling to work it is necessary that:

!B − 3B > !C (1)
!B − 3C < !C (2)

 These two together imply that:

3B < !B − !C < 3C

 The gain from signaling must be higher than the cost for high ability types, but not so high that low ability types
are also motivated to signal

24
 It is also necessary that for quicks their profit must be higher than what they would get if none of them
signaled at all (if none signal, everyone is paid average productivity !@):

G
!B − 3B > ! (3)

 This condition is stronger than (1) since G

! > !C : very large α implies that !@ is very close to !B , making
it more likely that condition (3) cannot be met

 Intuitively: signaling to distinguish is more likely to be profitable for quicks the rarer that they are

 If these conditions are not met, neither has an incentive to obtain the credential, and quicks do not
distinguish themselves from slows (pooling equilibrium): no signaling

 If the conditions are met, quicks signal and slows do not. This is called a separating equilibrium

25
WHICH TYPE OF FIRM IS MORE LIKELY TO USE SIGNALING?
 Signaling is helpful
 when employers do not have enough information about job applicants
 when differences in talent among potential employees matter a lot to productivity

 Signaling more important in jobs where skills are most important (at high levels of the hierarchy, in
research and development, and in knowledge work, professional service firms, such as consulting,
accounting, law firms, and investment banks):
 In such professions, small differences in talent can lead to large differences in effectiveness on the job,
so sorting for talent is very important

→ Signaling also more likely to be used where there is not much information already available about job
applicants (e.g. workers who are new to the labor market) although firms can use these techniques even for
hiring experienced talent at very high levels

26
 CHAPTER 3 – INVESTMENT IN SKILLS

Until this point: assumed that workers have fixed talents

However workers learn over time in many ways
The two most important for our purposes are:
Formal education and other pre-labour market training
On the job training
Economics and business world view education and training as investments which can be modelled like
any other kind of investment (Human Capital theory)

1
 EDUCATION

In capital theory, investments are made if the present value of the cash flow or other benefits generated
by the investment exceeds the present value of the costs of the investment

Suppose an individual is choosing whether to drop out or finish college this year (period 0)
Drop out: earnings Ht per period
Continue school: earnings Kt per period

t=1… T

→ Increased earnings from finishing school: (Kt – Ht) each year

Interest rate r

Present value of the return on the education investment

𝐾 𝐻
1 𝑟

2
 EDUCATION

Two costs of investments in education:

Direct (tuition, textbook, supplies, etc) C0
Opportunity cost of the time spent on education F0

Investment should be made as long as the present value of the return exceeds the present value of the cost

𝐾 𝐻
𝐶 𝐹 0
1 𝑟

For early years of schooling, the returns to schooling exceed the costs:
Much to be learned when an individual knows very little and a little bit of school can affect productivity
dramatically (but usually diminishing returns)
Costs of going to school are very low during the early years of schooling and foregone earnings during the
early years of schooling are very low
It pays for almost everyone to invest in some formal education, but that there is also an optimal stopping point
for each individual
The stopping point is the year when the net present value of investment in education switches from positive to
negative 3
 EDUCATION
Previous expression has several implications

1) Costs
Increases costs reduce enrolment: students close to the margin (net present value of education close to zero) will
now find that the costs exceed the benefits.

Those with high-paying jobs will reluctant to go back to school, all else equal (education is a better investment
when labour market opportunities are weaker)

2) Career Length
The longer the work life, the larger is the optimal investment in schooling people more likely to invest in
schooling when young

3) Specialization of Human Capital

Education tends to have diminishing returns portfolio of skills rather than focus?

This happens at lower levels of education (almost all education systems require student to develop a general
education with a little knowledge of many different subjects) but at relatively advanced stages specialization
becomes important because of comparative advantage and gains from trade: If individuals focus on one area and
become relatively expert, they can trade their output with those who specialize in other areas of expertise
4
 EDUCATION
4) Effectiveness of learning
When (K – H) increases, net present value of education rises and schooling should rise

Plausible that those with more innate ability tend to learn more efficiently in school, leading to higher K benefits of
schooling are higher for those who are already highly talented (increases inequality in skills and in earnings)

Improvements in school quality should have positive effect on K Technological innovation in education, changes
in the effectiveness of teaching methods or quality of teachers expected to increase investments in education

An important factor is the level of technology associated with the average job: education is complementary with a
technologically advanced society K and overall levels of education are higher in advanced societies than they
were in 1900

This logic can also help explain current patterns of education across different societies, and trends in returns to
education over the last several decades

5
Internal rate of return (implied interest rate on the investment) from education is generally quite high
(about 11% per year in the U.S., Asia, or Europe) and increased in recent decades:
Strong trend toward relatively higher earnings for those with a college degree
Increase in 90th/10th percentile of earnings

Most important explanation: increasing use of

advanced technology in the workplace

As discussed, capital tends to be complementary

with skilled workers, then greater use of
technology, increases the value of having highly-
skilled workers

This increases the demand for skilled workers,

which increases their labour market value

6
 INVESTMENT IN ON THE JOB TRAINING

Investment like education in many ways:

Increases worker skills and productivity
Benefits both the employer and the worker
There are direct costs (books or other resources, compensation for trainers, etc.) and indirect costs (on-the-job
training may take the worker’s time and attention away from regular duties, lowering productivity)

We start by considering when a possible investment in on-the-job training would be economically

profitable (gains in productivity exceed the costs of the investment)

7
 INVESTMENT IN ON THE JOB TRAINING: AN EXAMPLE

Small Silicon Valley startup provides enterprise software that does tax optimization
Typical employee must know
something about tax laws
programming in Java (unusual combination of skills)

Many firms value both skills independently, but few employers value an employee with the same
combination of expertise in tax law and Java as this firm
an employee who leaves the startup will find difficult to find a firm that can make use of all of the skills
that he acquired in the first firm

To what extent should you invest in skills and knowledge that help you in this job and a career in this
firm?
To what extent should you instead invest in training that would improve your job prospects in the outside
labor market?
8
 INVESTMENT IN ON THE JOB TRAINING: AN EXAMPLE

Current productivity at the startup $10,000

3 options for on-the-job training
1. Completely on Java
2. Completely on tax
3. Splits the training time between the two in
proportion to how the employer values those skills
in this job
Total costs of training $5,000 (same for all three choices)

Option 1 (only Java)

Since job involves spending about 40% of time on Java, productivity would rise by $3,200 (tax-related
productivity would not change)
This training affect productivity at a new firm depending on the kind of job
If elsewhere emphasize Java at 80%, productivity in an alternative job would rise by $6,400 on average
9
INVESTMENT IN ON THE JOB TRAINING: AN EXAMPLE

Option 2 (only tax)

Current job involves spending about 60% of
time on tax, productivity would rise by $4,800
If another job emphasizes tax work only
about 20% of the time, training would raise
productivity at other jobs by about $1,600 per
month on average

Option 3 (mix)
If mixed training, productivity rises in both tasks
Total productivity rises by $5,200, more with current firm than in the outside market.
That is because the training is designed to match current job’s skill mix, not the average skill mix required by
the labor market.
The best on-the-job training for this employee depends on where he is most likely to be employed after the
training
The optimal investment maximizes expected productivity
If expects to stay at the current firm, should train in both tax law and Java programming
If expects to quit this firm, should focus only on Java

In general,
If expect to stay at the current firm, best strategy is focuses training on the skills that current employer
values the most
If expect to leave, best strategy is to invest in skills that the labor market values the most

11
ON-THE-JOB TRAINING (FORMAL ANALYSIS)

Assume
firm and the worker share costs and benefits
investment decision made by the employee

Worker invests in skills

J (Java)
T (tax)

Cost of training C 𝐽 𝑇

Firm gives weight

λ to skill J
1-λ to skill T

Potential earnings at the current firm:

𝑊 𝜆𝐽 1 𝜆 𝑇

Wages determined similarly at other firms, but weights λ vary from firm to firm

12
 Two periods:
I: worker invests in on the job training
II: may or may not switch employers

Probability that the worker stays at the current firm next period = p

Worker chooses J and T to maximize net earnings:

1
max 𝑝 𝜆𝐽 1 𝜆 𝑇 1 𝑝 𝜆̅𝐽 1 𝜆̅ 𝑇 𝐽 𝑇
, 2

𝜆̅ = expected weight on Java skills at potential other employers

FOC:

p𝜆 1 𝑝 𝜆̅ 𝐽 0

p 1 𝜆 1 𝑝 1 𝜆̅ 𝑇 0

13
 𝐽∗ p𝜆 1 𝑝 𝜆̅

𝑇∗ p 1 𝜆 1 𝑝 1 𝜆̅

Investment is weighted average of the relevant skill-values inside the firm and outside (weights depend on
the probability of separation)
If p = 1 (continuation certain), the only skill value that matters is λ (current employer’s relative valuation of skill J)
If p = 0 (separation certain), current firm’s valuation does not matter and only 𝜆̅ matters

If a worker after investment switches to another firm where

wage = W’
weight given to skill J=λ’

Change in earnings:
𝑊 𝑊 𝜆 𝜆 𝐽∗ 𝑇∗

Sign uncertain
Typical case is low probability of separation and the worker invests with emphasis on the skills that the
current employer values
However, if high probability of separation or not too unusual current firm’s relative valuation of skills,
worker’s investment will tend more toward 𝜆̅
If so, leaving the firm might lead to an increase in earnings
14
GENERAL V. FIRM-SPECIFIC HUMAN CAPITAL
There are two training possibilities
General Human Capital: training is equally valuable inside and outside the firm, skills or knowledge
raise productivity equally at both the current employer, and with many other employers (e.g. MBA,
knowledge of a foreign language, most skills acquired outside of the workplace, such as at a
university)
Firm-Specific Human Capital (less common): it has no value outside the firm (e.g. knowledge of how to
operate unusual machine designed for a specific firm, intangible knowledge such as understanding of
informal networks and power relationships inside firm, working relationships with clients, and
knowledge of their particular organizations)

However: Most training falls in between and many skills have a value both inside and outside the firm,
though the values may be different

15
 WHO SHOULD PAY?
EDUCATION
Some employers pay for schooling of their workers is this a good investment for the employer?
In general no (general human capital) but there are reasons why a firm might pay some of the tuition costs

Implicit Cost to Employee & Benefit to Employer

The employee is in fact paying for the tuition by accepting lower salary at the job in order to get access to the tuition
benefits (often employers impose a contractual obligation on employees to pay back the tuition if they quit the firm within
a few years of graduation)
Matching
The firm offers tuition benefit to employees that are a strong match: The firm expects that these employees will stay at
the firm for many years, and expect to recoup some of the benefits of the schooling if the worker has incentive to stay
with the firm (firm and the employee split the profits from the investment in education)
Recruiting
Offering a certain benefit may generate useful self-selection in recruiting
UPS, for example, offers tuition reimbursement to its employees and so attracts a harder working, more ambitious
workforce and also younger
Arbitrage
The firm might have a cost advantage in paying for schooling if there are tax benefits to paying for education or training
16
 WHO SHOULD PAY?
ON-THE-JOB TRAINING
If skills are pure general human capital (new training valued equally by both other employers
and the current employer), when the worker gets the training, his or her market value rises
The firm will thus have to pay the worker a higher salary once the training is completed, or risk
having the worker quit
For this reason, the general rule is that if skills are completely general human capital, the
worker should pay for 100% of the investment, and receive 100% of the benefits.

17
ON-THE-JOB TRAINING – FORMAL ANALYSIS
Perfect competition (perfect information, no mobility cost or rigidities, no institutional constraints) wage
related to productivity
One worker hired for two periods (r discount rate)
Production function y=y( ) where is training with y’>0 and y’’<0
Cost of training c=c( ) with c’>o
t
t=1 t=2
• Training at cost c( ) • Productivity y1=y( ) > y0
• Productivity y0=y( =0) • Wage w1=w( ) > w0
• Wage w0

Firm Objective:

𝑦 𝜏 𝑤 𝜏
max 𝜋 𝑦 𝑤 𝑐 𝜏
1 𝑟
P=…Rent Sharing and Compensation
18
ON-THE-JOB TRAINING – FORMAL ANALYSIS
FOC:

𝑐 𝜏 0P=…

𝑦 𝜏 𝑤 𝜏
𝑐 𝜏 → 𝑀𝐵 𝑀𝐶
1 𝑟

In case of general training, competition among firm push the wage up to the new productivity level:
𝑤 𝜏 y 𝜏

MB of training is null for the firm and the optimal training to provide is 0

19
ON-THE-JOB TRAINING – FORMAL ANALYSIS
Since benefits of training belong completely to workers in terms of higher wages, they should pay for
training
Worker’s choice in case she pays for training:
Worker’s objective:

𝑤 𝜏
max 𝑈 𝑤 𝑐 𝜏
1 𝑟
FOC:

𝑐 𝜏 0P=…
Since in competitive labour market wage equates productivity

𝑐 𝜏 → 𝑀𝐵 𝑀𝐶

20
 𝑦 𝜏 𝑦′ 𝜏
1 𝑟 1 𝑟

𝑐 𝜏 𝑙𝑖𝑛𝑒𝑎𝑟

𝑐′ 𝜏

𝜏* 𝜏 𝜏* 𝜏

Since worker’s benefit is equal to total benefit of training, the training level is efficient (consider 𝜏 < 𝜏 *
and 𝜏 > 𝜏 *

21
ON-THE-JOB TRAINING – FORMAL ANALYSIS
Although generally firms formally pay for training, they can transfer cost of training to workers by paying
lower wages while training is provided (e.g. stage). Workers are willing to accept because their future wage
will increase
𝑤 𝑦 𝑐 𝜏∗

y 𝜏∗

𝑦
𝑦 𝑐 𝜏∗

training time

22
ON-THE-JOB TRAINING AND WORKERS CHARACTERISTICS – FORMAL ANALYSIS
Assume
𝑦 𝜏 𝜋𝜏 where π represents innate ability and 𝛼 1
Linear cost function 𝑐 𝜏 𝑐𝜏
Utility function

𝜋𝜏
𝑈 𝑤 𝑐𝜏
1 𝑟
FOC:

0 ⇒ 𝜏∗ P=…

Investment in training increasing in 𝛼 and 𝜋 and decreasing in 𝑐 and r

Cost of training (𝑐) tend to be higher and returns to training (𝜋) lower for less talented workers they will
have less incentive to invest in training

23
ON-THE-JOB TRAINING AND WORKERS CHARACTERISTICS – EXAMPLE
.
𝑦 𝜏 100𝜏
𝑐 𝜏 5𝜏
𝑦 200
𝑤 200
r=0.05
Utility function

100𝜏 .
𝑈 200 5𝜏
1 0.05
FOC:

𝜕𝑈 50𝜏 .
0⇒ 5 0
𝜕𝜏 1.05𝜏 .

50
𝜏∗ 90.7
5 1.05

𝑐 𝜏 5 90.7 453.5 𝑝𝑎𝑦𝑚𝑒𝑛𝑡 𝑡𝑜 𝑓𝑖𝑟𝑚 253.5

In the second period, the worker will obtain the return to the investment in training:
.
𝑤 𝜏 𝑦 𝜏 100 90.7 952.38

24
ON-THE-JOB TRAINING AND WORKERS CHARACTERISTICS – FORMAL ANALYSIS
An important factor is the length of the period over which the new skills will be used
Assume that the worker will work for T periods after training (ignore discounting)
Utility function
𝑈 𝑤 𝑐𝜏 𝑇𝜋𝜏
FOC:

0 ⇒ 𝜏∗ P=…

Investment in training increasing in 𝑇

more training when workers are young

Low wage when young (workers pay training with lower

wage), wage increases thank to increased productivity,
investment in training decreases at older age (less
incentive), wage increases but at decreasing rates
25
 WHO SHOULD PAY FOR SPECIFIC TRAINING?
If human capital partly or fully firm-specific, worker’s outside value after training is lower than inside, even
though the labor market does value the training. What happens in this case?

Consider our software programmer and assume that best investment choice is training in both Java and tax laws
Two periods
Direct and indirect costs of training (C + F) = 5,000

productivity after training at current firm

return on training
productivity after training the labor market as a whole.
if stays at the firm return on training if
leaves the firm
Productivity w/o training
Cost of training

net productivity Because the return on the education

investment it is larger if he stays, this
investment is somewhat specific to this firm
26
WHO SHOULD PAY FOR TRAINING? ON-THE-JOB TRAINING
Suppose worker pays for the investment, expecting to earn the returns in the second period (i.e. firm
agrees to pay the worker a salary equal to his productivity in both periods)
Firm earns no profit or loss in either period
Worker
First period: loss (pay is 5,000 less than he could earn elsewhere)
Second period: profit (pay is 5,200 more than he could earn elsewhere)
After the investment is made, if the firm pays 15,200, it pays more than it has to keep the employee
(employee can only earn about $14,400 elsewhere) tempted to pay a little more than $14,400
The worker will accept since he can only get $14,400 elsewhere
The firm may be tempted to renege on its promise, and renegotiate after the investment is made
because paying less than 15,200 captures some of the profits from the investment that the worker made
The worker may foresee this risk and then will be unwilling to make the (profitable) investment

27
WHO SHOULD PAY FOR TRAINING? ON-THE-JOB TRAINING
Suppose the firm pays for the investment paying the worker what he would get if there was no
investment (10,000 in each period) and provide training
The firm would
pay the cost, since productivity would be less than pay during training
capture the benefits, since productivity would be higher than pay after training
Same risk of renegotiation: once investment is made, the employee may be tempted to renegotiate
pay higher than 10,000 because his market value has risen to $14,400 and he might even ask for pay
close to his value (15,200)
Firm tempted to accept since they would lose $5,200 in profits if he left
But if they do, the employee will get some of the profits from the firm’s investment!

Hold-Up Problem: one party makes an investment and expects to earn benefits later, but a second
party is tempted to renegotiate after the investment is made
If this risk is foreseeable, the investment may not be made, for fear of renegotiation

28
Hold-Up Problem

How solve this problem? Two possibilities

Rely on the trustworthiness of one or both parties
Split the cost and the return on the investment

benefits split: W2 between what could earn elsewhere,

and actual productivity (if 50-50, W2 = 14.8).
Splitting benefits reduces (though not eliminate) the
temptation to renegotiate

Costs split: W1 between actual net productivity and

what could get elsewhere (if 50-50, W1 = 7.5)
Splitting the costs reduces the risk of the investment
in the first period, since there is less to lose.

Moreover, since both have something to lose if the relationship was broken, both have incentive to avoid
renegotiation
Therefore, investments specific to the firm are likely to be split by the worker and the firm:
during training pay lower than what the worker could earn elsewhere, but greater than net productivity
after the training: pay greater than what the worker could earn elsewhere, but less than productivity 29
IMPLICATIONS OF ON-THE-JOB TRAINING
Turnover
If training is completely general,
the firm made no investment and is earning no return on the training nothing to lose if the worker quits.
the worker takes the full investment with him nothing to lose by switching to another employer
If training is specific firm and worker care about turnover.
With shared firm-specific investments, both lose if the worker leaves and the loss is higher the larger the
difference between productivity at this firm and elsewhere
A relationship arises between the worker and the firm and both have incentive to invest in and maintain the
relationship
The term used for this is that such firms emphasize internal labor markets

Investment
Lower turnover induce workers to invest in firm-specific skills and viceversa
Thus, firms that desire an unusual mix of skills generally adopt policies to reduce turnover
Investment patterns should change with tenure
The longer employee has been with the firm, the more likely is she already invested in firm-specific skills and this
increases his incentive to stay with this employer, reinforcing his tendency to invest even more in firm-specific
skills
Thus, as job tenure rises, workers tend to become more invested in their current employer 30
Compensation
Pay will tend to rise with labor market experience, because most jobs provide some on-the-job training
Moreover, those who have higher tenure in a firm will tend to have higher pay than those who do not,
because they have more firm-specific skills
The more firm-specific are employee skills, the higher the expected loss in compensation from leaving the
firm, the lower the likelihood that the employee will leave the firm

Firm Size
Workers in larger firms will tend to invest more in firm- specific skills because they
Tend to have lower turnover (empirically)
Are more likely to be able to find alternative work for their employees who wish to change jobs if the skill mix is
similar across jobs in the same firm, this thickens the market for an employee’s skills when they work for a larger
firm Rent Sharing and Compensation

31
 CHAPTER 4 – MANAGING TURNOVER

In this chapter we use economic tools to analyse some issues in the management of employee careers
We will study
1. Reasons for TO
2. How reduce TO and cost of TO
3. Bidding for employees
4. Layoffs (who target)
5. Buyouts

Employee TO happens in two different circumstances:

Need to lay off workers due to downsizing
General need to manage regular workforce flows in and out of the firm: every firm should have some employee
turnover as part of its business; the question is how much, and of what kind.

1
 EFFICIENT, INEFFICIENT, CONSENSUAL AND NON-CONSENSUAL SEPARATIONS

Separations can be either firm-initiated (e.g. demand drop, technological change) or worker-initiated (e.g. start
own business, job offer from other firms)
In the absence of firing costs or restrictions, a job is maintained as far as it brings benefits to both parties, i.e.
both the firm and the worker have a positive surplus.

Worker’s surplus:
ഥ −𝑈
𝑆𝑊 = 𝑊 ഥ
Where

ഥ = 𝑤1 + 𝑤2 + 𝑤3 2 + ⋯
𝑊
𝑤𝑇
(𝑤𝑖 = 𝑝𝑒𝑟𝑖𝑜𝑑 𝑖 𝑤𝑎𝑔𝑒)
1+𝑟 (1+𝑟) 1+𝑟 𝑇−1

ഥ = 𝑢1 + 𝑢2 + 𝑢3 2 + ⋯
𝑈
𝑢𝑇
(𝑢𝑖 = 𝑝𝑒𝑟𝑖𝑜𝑑 𝑖 𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑜𝑝𝑡𝑖𝑜𝑛)
1+𝑟 (1+𝑟) 1+𝑟 𝑇−1

The worker is interested in continuing the job as long as 𝑆𝑊 > 0 (non necessary that current wage > current
outside option) 2
 EFFICIENT, INEFFICIENT, CONSENSUAL AND NON-CONSENSUAL SEPARATIONS

Firm’s surplus:
𝑆𝐹 = 𝑌ത − 𝑊
ഥ
Where
𝑦2 𝑦3 𝑦𝑇
𝑌ത = 𝑦1 + + +⋯ (𝑦𝑖 = 𝑝𝑒𝑟𝑖𝑜𝑑 𝑖 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛)
1+𝑟 (1+𝑟)2 1+𝑟 𝑇−1

The firm is interested in continuing the job as long as 𝑆𝐹 > 0

Total surplus:
ഥ − 𝑈)+(
𝑆 = 𝑆𝑊 + 𝑆𝐹 = (𝑊 ഥ 𝑌ത − 𝑊)
ഥ = 𝑌ത − 𝑈
ഥ

3
 EFFICIENT, INEFFICIENT, CONSENSUAL AND NON-CONSENSUAL SEPARATIONS

If either 𝑆𝑊 or 𝑆𝐹 < 0 a separation is initiated.

If both 𝑆𝑊 and 𝑆𝐹 < 0 consensual separation
If 𝑆𝑊 > 0 and 𝑆𝐹 < 0 firm initiated non-consensual separation
If 𝑆𝑊 < 0 and 𝑆𝐹 > 0 worker initiated non-consensual separation

The separation is efficient if total surplus 𝑆 = 𝑆𝑊 + 𝑆𝐹 = 𝑌ത − 𝑈

ഥ < 0 (always for consensual separations)
Non consensual separations can be either efficient or inefficient

Examples
𝑌ത = 10, 𝑊
ഥ = 13, 𝑈
ഥ=8 𝑆𝐹 < 0; 𝑆𝑊 > 0; 𝑆 > 0 → 𝐹𝑖𝑟𝑚 𝑖𝑛𝑖𝑡𝑖𝑎𝑡𝑒𝑑 𝑖𝑛𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑠𝑒𝑝𝑎𝑟𝑎𝑡𝑖𝑜𝑛
𝑌ത = 15, 𝑊
ഥ = 12, 𝑈
ഥ = 13 𝑆𝐹 > 0; 𝑆𝑊 < 0; 𝑆 > 0 → 𝑊𝑜𝑟𝑘𝑒𝑟𝑠 𝑖𝑛𝑖𝑡𝑖𝑎𝑡𝑒𝑑 𝑖𝑛𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑠𝑒𝑝𝑎𝑟𝑎𝑡𝑖𝑜𝑛
𝑌ത = 12, 𝑊
ഥ = 13, 𝑈
ഥ = 14 𝑆𝐹 < 0; 𝑆𝑊 < 0; 𝑆 < 0 → 𝐶𝑜𝑛𝑠𝑒𝑛𝑠𝑢𝑎𝑙 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑠𝑒𝑝𝑎𝑟𝑎𝑡𝑖𝑜𝑛

4
 EFFICIENT, INEFFICIENT, CONSENSUAL AND NON-CONSENSUAL SEPARATIONS

𝑆𝑊 < 0
Consensual
separation (efficient)
𝑆𝐹 < 0 𝑆 = 𝑆𝐹 + 𝑆𝑊 < 0
Firm initiated Efficient
separation
𝑆𝑊 > 0
Non-consensual
separation
𝑆 = 𝑆𝐹 + 𝑆𝑊 > 0
Inefficient

5
 WAGE CUTS

If 𝑆𝑊 > 0 and 𝑆𝐹 < 0 firm initiated non-consensual separation

To avoid separation, the firm and the worker could agree on a wage cut ∆𝑊 that makes the firm’s surplus
positive:
𝑆𝐹 = 𝑌ത − 𝑊
ഥ − ∆𝑊 ≥ 0
The worker’s surplus has to remain positive after the wage cut:
ഥ − ∆𝑊 − 𝑈
𝑆𝑊 = 𝑊 ഥ≥0

If the separation is inefficient (𝑌ത − 𝑈

ഥ > 0) an agreement on a wage 𝑌ത > 𝑊 > 𝑈
ഥ such that both surplus are
positive is possible
If the separation is efficient (𝑌ത − 𝑈
ഥ < 0) such agreement is not possible
However: institutional constraints (minimum wage, collective bargaining), effect on workers’ motivation in case
ത
of asymmetric information (workers don’t observe 𝑌)

6
 WAGE INCREASES

If 𝑆𝑊 < 0 and 𝑆𝐹 > 0 worker initiated non-consensual separation

To avoid separation, the firm and the worker could agree on a wage increase ∆𝑊 that makes the worker’s
surplus positive:
ഥ + ∆𝑊 − 𝑈
𝑆𝑊 = 𝑊 ഥ≥0
The firm’s surplus has to remain positive after the wage increase:
𝑆𝐹 = 𝑌ത − 𝑊
ഥ + ∆𝑊 ≥ 0
If the separation is inefficient (𝑌ത − 𝑈
ഥ > 0) an agreement on a wage 𝑌ത > 𝑊 > 𝑈
ഥ such that both surplus are
positive is possible
If the separation is efficient (𝑌ത − 𝑈
ഥ < 0) such agreement is not possible
ഥ
However: fear of strategic behaviour by the worker in case of asymmetric information (firms don’t observe 𝑈),
fear of effect on other workers’ behaviour

7
 FIRM INITIATED SEPARATIONS AND WAGE CUTS
An Example
Consider a firm that is hiring ten workers and needs to restructure and downsize its labour force
The firm needs to decide which workers should be kept and which workers should be offered a wage-cut
The various workers act separately and the firm is able to distinguish the job of each and every worker

The firm is losing money on three workers (i.d. numbers 3,4 and 9) and these three workers are the obvious
candidate for a separation.

8
 FIRM INITIATED SEPARATIONS AND WAGE CUTS
An Example

On worker 3, the firm loses 8 while the worker enjoys a surplus of 5. A wage-cut up to 5 is profitable for this
worker but such wage cut is not enough to compensate the firm the total surplus is indeed negative and a
separation is efficient.
A similar situation exists for worker number 9, since the firm is losing 20 while the firm has a surplus of 15
On worker 4, the situation is more complicated, since the firm is losing 5 but the worker enjoys a surplus of 11.
Here a job preserving wage cut is possible, since the total surplus is positive a wage cut equal to 6 would
make the job viable

9
 FIRM INITIATED SEPARATIONS AND WAGE CUTS
An Example

On worker 6, the firm is indifferent while the worker has a negative surplus This separation is not firm
initiated but worker initiated, and we typically called it a quit. There is no surplus and the job is going to be
destroyed
On worker 6, conversely, a profitable wage raise is possible

10
 WHY TURNOVER UNDER REGULAR BUSINESS CONDITIONS?

Importance of Sorting
Turnover can be used to continuously sift candidates to find the best ones

Thus, it is important especially in jobs where talent is particularly important and where there is much to be learned
about workers especially early in the career and for knowledge workers (e.g. in leading professional service firms)

Technical Change
New employees are more likely to have new insights, bring different perspectives, and understand the latest ideas,
technology or other developments turnover should be higher in industries where technology advances more rapidly
Organizational change
Current employees are experts at the firm’s current way of doing business, but if the firm needs to change methods,
they are almost certainly no longer the best fit (particularly true for senior management)

Firms that continuously bring in outsiders are more likely to recognize when times change, and adapt effectively

11
Hierarchical Structure

Higher turnover may be necessary when the organizational structure dictates that the hierarchy narrows rapidly at some
level

In this case some turnover is inevitable because there will be few promotion opportunities

Some managers may become frustrated and quit to pursue other opportunities and best workers are most likely to be lost
if promotions are not available

Specific Human Capital

The more specific to the firm is the training, the higher are the turnover costs, therefore, firms with more idiosyncratic
businesses, methods, or cultures are more likely to want to have low turnover

Similarly, in jobs where valuable intellectual property is developed or where workers develop strong client relationships, it
is important to try to reduce turnover

12
 RETENTION STRATEGIES (HOW REDUCE TO)
Tools to reduce turnover

1. Increase compensation:
simple and expensive but solution if key employees (i.e. those very damaging to lose) receive outside job offers
Firm can offer stock, options, make them partners (many professional service firms are organized into some form of
partnership)
Bottom line: firm must pay key employees their market value or will probably lose them

Formal analysis
Assume that a firm employs a single worker
Y is worker’s output
W is wage
𝐴
Probability that the worker quits: 𝑞 𝑊 = 𝑞′ 𝑊
<0
𝑊
𝐶𝑇 is cost of turnover for the firm

Wage increase implies

an increase in wage cost
a decrease in turnover cost

13
 RETENTION STRATEGIES (HOW REDUCE TO)

Firm’s objective:
𝐴
max 𝜋 = 𝑌 − 𝑊 − 𝑞𝐶𝑇 = 𝑌 − 𝑊 − 𝐶
𝑊 𝑊 𝑇

FOC:
𝜕𝜋 𝐴
= −1 + 𝐶 =0
𝜕𝑊 𝑊2 𝑇

𝑊∗ = 𝐴𝐶𝑇
Optimal wage is increasing in turnover cost

The corresponding optimal q is

𝐴
𝑞∗ = >0
𝐶𝑇

14
 RETENTION STRATEGIES (HOW REDUCE TO)

Tools to reduce turnover

2. Tailoring some benefits or characteristics of the work to that worker’s tastes (e.g. flexible working hours if such
flexibility is difficult to find elsewhere)

2. Offering new opportunities like

new training (increases the long-term value of the job)

job enrichment (new tasks or responsibilities make the job more interesting)

early promotion (signal to the employee the value that the firm place on her long term employment at the firm)

15
 REDUCING COSTS OF LOSING KEY EMPLOYEES

Some turnover inevitable, but firms can employ a few strategies to make it less costly
Turnover most costly when the worker has knowledge that other employees do not share, hence:
Have workers collaborate on key tasks so that key knowledge is not monopolized by one employee.
Cross-train (each worker train colleagues in what they do)
Switch tasks periodically,

Job design can also affect turnover costs: the more standardized jobs are, the less costly losing one employee (others
can step into the void)
Knowledge management strategy: attention to procedures by which the knowledge that is created can be documented
for reuse (e.g. set up databases to document new methods, ideas, products)

16
 EMBRACING TURNOVER
Turnover not always bad (some organizations embrace turnover). Two examples:

1. In professional service firms (consulting, law, and accounting) where it is quite common for an employee who
leaves to go to work for their client and this reinforces the working relationship between the firm and its client,
benefitting both

2. Hewlett Packard for many years encouraged employees to quit but also to apply to return to HP in the future.
Why?
Such employees are the best and can increase the quality of its workforce because some may return later
Employees who leave may bring future business
Employees who leave and then return valuable because have a mix of inside and outside experience
Develop reputation as an employer who cares about interests of its employees, which is likely to increase motivation and
reduce conflict

17
 BIDDING FOR EMPLOYEES
Firms are engaged in an active auction market, bidding against each other for employees (especially the most
talented)
The firm decides whether
Raid a competitor for a potential hire costs of time and other resources involved in the bidding war
Not to raid no direct costs

Because of matching and firm-specific human capital, a worker is more likely to be worth more to the current
employer than to the outsider. Under such circumstances, the outsider should not raid
However, there are situations in which a worker has skills well-suited to another firm’s current situation. Under such
circumstances the outsider should ride
However, the problem is that the outsider has less information about the quality of the worker and sometimes
overestimates the worker’s potential productivity the outsider raids but should not

18
19
20
 OFFER MATCHING

Sometimes, current employers refuse to match offers of outsiders policy of “no offer matching”

When is it reasonable to match outside offers and when not?

Determine what affects worker search behaviour: example

Worker currently earning $20 per hour (2000 hours)
One job in his labour market (with 50 firms) will pay $20.50 per hour
He doesn’t know which firm of the 50 is the high payer need to fill out applications (cost X)
If the worker’s current firm agrees to match any outside offer, he will search if the expected present value of
the increased earnings exceeds the costs:

𝑇
1 2000 × 0.5
෍ >𝑋
50 (1 + 𝑟)𝑡
𝑡=0

Low values of X and long work lives tend to lead to higher return to searching for a better job offer

Under offer matching the search may result in the firm having to pay higher wages for any given worker

21
 OFFER MATCHING

A policy of “no offer matching” affect the worker’s behaviour depending on whether the worker will actually leave
to take another job

If he is willing to leave for the new job, then there is no effect on search behaviour of announcing no offer
matching because the returns to the worker from search are the same, with the only difference that the worker will
take the job offer at the new firm.

Which situation does the firm prefer?

With no offer matching the worker leaves.
With offer matching, the worker stays.

Although the firm might not always want to match offers, one would think that the firm would always prefer
having the option to match offers.

If the wage that the firm must pay exceeds the worker’s worth, then the firm can always let the worker go

22
 OFFER MATCHING

A policy of “no offer matching” might discourage certain kinds of search if the worker has a strong (non-pecuniary)
preference for his current firm (co-workers, location, general work environment) and then is unwilling to move to
another firm, even at a higher wage

If the firm matches alternative offers it will raise wages to $20.50, providing an incentive to search for an offer
that he has no intention of accepting

If the firm has a policy of “no offer matching,” the worker would not search, because he is unwilling to accept the
job at the higher wage, then the policy of no offer matching discourages search and saves the firm money

The problem is that it is sometimes difficult for a firm to distinguish between genuine threats and offers that the
worker would decline

For instance, when workers are in the job “for the money,” the non-pecuniary components of the job are less
important and little is gained by a policy of no offer matching because workers will move if the firm does not
match, and the firm cannot discourage needless search by refusing to match (e.g. investment banks rarely adopt
“no offer matching” policies, because money is the driving factor in these industries)
23
 LAYOFFS & BUYOUTS
If firms must downsize by laying off large groups of employees, who to target for layoffs?
Careful about laying off the most highly paid employees (often the most productive), or the less productive
(often the less expensive)
Better approach is to target those employees from which the firm is losing money relative to other employees
(could be high or low paid workers, high or low productivity workers)

Specific Human Capital

An important factor in deciding who to target is the degree of firm-specific human capital
When firm-specific human capital is important, the firm maximizes its profits by laying off from both ends of
the age distribution first: workers who recently started with the firm, and those who are nearing retirement

24
Profiles of a hypothetical worker’s earnings and productivity over the career, with an investment in firm specific human
capital
Kt = productivity at the firm
Wt = wage
At = value of the worker’s best alternative outside of the firm; it depends on:
earnings that the worker could receive in another job (most important factor for younger workers)
value that the worker places on leisure (higher for older worker)
Optimal point of retirement is T (at some point all workers would be better off by retiring due to a rising profile At)

25
Difference between PV(Kt) and PV(Wt) is return earned by the
firm (bottom panel), and it is before increasing and then
decreasing
In general, the firm earns the greatest profits (in present value) on
workers who have completed training and have both high
productivity and relatively long remaining careers (i.e. workers of
medium age)
If productivity falls (e.g. because demand and prices for the firm’s
products decline) drop in Kt to βKt (with β<1)
This corresponds to a downward shift in the present value Pt
No longer pays to make investments in young workers
Similarly, profitable to lay off older workers
Only employees that would be profitably employed would be
workers with middle ages

26
 COSTS OF LAYOFFS
Laying off younger workers less controversial
Typically less legally protected
Invested little in firm-specific skills
Laying off older workers more controversial
Protected by anti-discrimination regulations in most countries
Invested in firm-specific skills, and are now enjoying the returns
Laying off such workers may be perceived as a breach of trust by the employer and (problem if firm cares about
its reputation as a fair employer

In general, a serious cost of layoffs is the litigation that may ensue: many economies provide protection
for workers against wrongful termination, employees may sue if fired

27
 BUYOUTS
Because of the costs of layoffs, many firms opt to offer employees buyouts
contract between a worker and a firm where in exchange for some compensation an employee agrees to end
employment with the firm
One concern is adverse selection:
at any wage category some employees are more productive than others (i.e some are relatively overpaid, and some are
relatively underpaid)
More productive likely to also have better alternative employment elsewhere more likely to accept the buyout
This suggests that buyout packages should be carefully designed, for example, the best performing employees might
not be offered buyouts, or might be offered buyouts that are less lucrative
Similar considerations apply to how buyout packages might vary with the employee’s age:
Workers close to retirement have little to lose from leaving the firm (earned most of the return on their investment in
skills) and require only small buyout packages
Those further from retirement usually require larger buyouts

28
 Analysis of Which Workers to Lay Off

W=wage (assumed constant)

A=value of the worker’s best alternative
K=productivity
Interest rate is about 25

PV(K)=PV(W) for new and retiring workers (competition forces this result)
Last two columns: value of the worker’s productivity falls by 30 percent (β = 0.7) because of a decline in demand
for the firm’s products
Buyout accepted if:
PV(W) – PV(A) < B

In the table, PV(W)>PV(A) for all workers all workers would have to be offered a buyout to be willing to leave

29
Profit to the firm if the worker leaves:

PV(W) – PV(K)

If positive
firm would like the worker to leave and the expression equals the
maximum profitable buyout B that the firm can offer
If negative
firm would prefer to keep the worker

In the table, all workers are profitable

30
 After production falls, this expression changes to

PV(W) – PV(βK)

workers >57 or <30 are no longer profitable

Rule for optimal buyouts

If gains from losing the worker exceed the losses to the worker from leaving, there
is room for a deal to be made: firm can offer a buyout that increases its profits,
and also makes the worker better off when they leave

Putting these together, we can make a deal as long as:

PV(W) – PV(K) > PV(W) – PV(A)

or
PV(A) > PV(K)
A buyout is possible if the present value of the worker’s alternative exceeds the present value of the worker’s productivity at the firm
Low output and good alternatives make buyouts feasible

NB: individuals that the firm would like to lay off are not necessarily the ones for whom a buyout offer is possible:
firm would like to lay off those >57 and <30 but:
Only workers >62 and older have alternatives sufficiently attractive to make an offer feasible
Firm loses money on those 57-61, but not enough that a buyout would be a better option, given the amount of buyout that
they would require
Similar logic applies to those aged 30 and younger
31
 IMPLEMENTATION OF BUYOUTS

Window Plans
Often, announcements of buyout offers are a surprise, and workers are given only a limited time to accept the offer
Window plans because there is a small period or “window” during which the buyout is available
Reasons:
If a buyout is anticipated, a worker has incentives to reduce productivity and a short fuse prevents the worker from
strategically reducing productivity for any significant time period to gain a higher buyout price
It also reduces the chance that the worker can find a suitable outside offer, since there is less time to search

32
Threat of Layoff
To increase acceptance rates for buyout offers the firm can threaten to lay off some fraction of those who do not accept
Suppose that:
firm announce it would lay off randomly 50 percent of all workers who did not accept
if a worker loses his job, he expects that the next job will give him about $10,000 less in present value (minimum
buyout that he will accept)
There is a 50 percent chance that he will lose $10,000 with no buyout willing to accept a buyout of only $5,000 (or
even less, if risk averse)

In general, greater odds of being laid off lowers the buyout that a worker requires
Costs of such a strategy:
lay off costs
Benefits:
increases probability that a worker will accept a given buyout amount
reduces the buyout required to motivate an employee to quit

33
Threat of Layoff (formal analysis)
Profit to the worker from staying (rejecting a buyout):

𝑃𝑉 𝑊 − 𝑃𝑉 𝐴 > 0

The expression is positive, or she would quit on her own

If worker accepts a buyout

receives B + 𝑃𝑉 𝐴
If worker rejects a buyout
with probability 1 − p continues in her employment, earning 𝑃𝑉 𝑊
with probability p is laid off, and earns 𝑃𝑉 𝐴

Accept the buyout if:

B+𝑃𝑉 𝐴 ≥ 1 − 𝑝 𝑃𝑉 𝑊 + 𝑝𝑃𝑉(𝐴)

The minimum buyout that she will accept is:

𝐵∗ = (1 − 𝑝) 𝑃𝑉 𝑊 − 𝑃𝑉 𝐴

This expression is largest when p = 0

Moreover:
𝑑𝐵 ∗
<0
𝑑𝑝
Greater threat of being laid off makes the worker more likely to accept a given buyout offer, and reduces the amount of buyout
that she is willing to accept 34
Speed and Extent of Downsizing
Implementing layoffs quickly and by surprise reduces the amount of organizational trauma
Downsizing highly emotional, workers unproductive because focus on who will be laid off, when, and how
For similar reasons, a firm should lay off more workers than seems apparent at first glance to
minimize the odds that it will have to do so again soon
restructure areas of the organization that need radical restructuring because the costs of firing tend to be lower when
implemented in the context of a larger set of layoffs

Retirement Bridges
Minimum buyout price necessary for a worker close to retirement is relatively small
Because buyout formula offering less to older workers might run into legal difficulties, firms may offer retirement bridge
giving workers seniority credit for the purpose of pension calculations
For example, if retirement age is 65 and a worker leaves at 55 with 18 years of service, he is treated as if he had 28 years of
service for the purpose of calculating retirement benefits
Since number of years awarded by the bridge declines with age, older workers are in effect given a smaller buyout award

35
Job Placement Services Su
Firms sometimes set up job placement services for workers they lay off or offer buyouts to
This may result in cost savings if improving the odds that workers can find outside work lowers the buyout price that
they will require
Whether this logic holds, however, depends on the firm’s efficiency in securing new employment for its workers
The service should be offered only if the firm can provide it more cheaply than the worker can buy a similar service
himself because otherwise, it would be cheaper for the firm to simply pay for outside services that a worker purchases
(e.g., through some kind of voucher system)

36
 TEAMS
Team: aggregation of workers performing a collective task
Compensation usually includes a common component
Team production has both benefits and costs
Teams should be used when the benefits are the highest and the costs the lowest

Two reasons why cautious about using teams

1. Group Decision Making

Decision making tends to be faster and more straightforward when there is a single, clear leader
To the extent that teams involve group decision making too much discussion and politics over key decisions
Thus, it is usually important to establish a clear team leader, supervisor, or mechanism to efficiently resolve
disputes with the group

1
2. Free Rider Effects
In teams workers can often hide behind the productivity of others
Consider
Worker placed in a team with 4 other workers
The team is assigned a project
For each day that the project comes in ahead of schedule, the team will be given a bonus of €100, which will be
split evenly among the five team members
If the worker stay late to work on the project, he will speed up completion of the project by one day
This is worth €100 to the group, but only €20 to him because the bonus is shared
If he could receive the €100 himself, he would be motivated enough to stay at work

Effort below efficient level because the worker who bear the pain of toiling does not reap the full benefit
The company cannot pay individuals rather than teams for their effort because in a team individual effort is
difficult to extract from the group's effort and results
This is because teams are set up in cases where the work is highly interdependent between coworkers and
individual performance is difficult to evaluate accurately

2
Team production function:
𝑁

𝑄 = 𝑒1 + 𝑒2 +. . . 𝑒𝑁 = 𝑒𝑖
𝑖=1

Cost of effort:
𝑒𝑖2
𝐶 𝑒𝑖 =
2
Total cost
𝑁

𝐶(𝑒𝑖 )
𝑖=1

Efficient level of effort:

𝑁 𝑁

𝑀𝑎𝑥𝑒𝑖 𝑒𝑖 − 𝐶(𝑒𝑖 )
𝑖=1 𝑖=1

Solution:
𝑒𝑖 = 1
MB of effort equal its MC
3
Worker’s Utility function:
𝑒𝑖2
𝑈=𝑤−
2
With:
𝑁
𝑖=1 𝑒𝑖
𝑤=
𝑁
Workers choose effort in order to maximize utility:

𝑁
𝑖=1 𝑒𝑖 𝑒𝑖2
𝑀𝑎𝑥𝑒𝑖 −
𝑁 2

Solution:
1
𝑒𝑖 = < 1
𝑁
The optimal level of effort is lower than the efficient level (the more as N increases)

4
 When to Use Teams
Primary benefit from using teams derives from multitasking
In many cases there are so many tasks that are complementary that the resulting job would be
overwhelming for one worker
In that case, the firm must either
separate the tasks (coordination problems)
increase the capacity of labor employed to do the work. This can be done in two ways:
1. Use workers with greater depth and breadth of skills
2. Use teams who work closely together

Activities can be ranked according to the costs and benefits of having teams
Production activities that rank high on the benefits scale and low on the cost scale should be
performed by teams

5
 EXAMPLE FROM A FISHING FIRM: COSTS AND BENEFITS OF TEAMWORK

ACTIVITY BENEFIT COST RANK COMMENT

RANK
Fishing on a 2 5 Tasks too many for one person;
small boat Monitoring easier with small
group; Remove unproductive TM
Fishing on a 1 4 More teamwork needed; More
large boat difficult to monitor efforts; Free
Rider more prominent
Selling fish 5 1 Sales people can work alone;
wholesale Monitoring as a group difficult
Accounting 4 2 Not much benefit from teaming;
for sales Individual monitoring ideal
Selecting the 3 3 Group judgment matters; Joint
site decision can require effort

The best candidate for team production is the actual fishing itself, and the worst one is selling the fish
6
 Other Benefits of Team Production:

SPECIALIZATION

Specialization is an important reason why individuals work together

In a team, each individual can specialize her human capital investments and be assigned a subset of
all of the tasks necessary to the overall business process

Specialization does play an important role in teams because in most teams workers do not perform all
tasks

Rather, they specialize and often rotate tasks between each other over time

7
KNOWLEDGE TRANSFER

Knowledge transfer is the second benefit from team production and probably occur more when
specialization is not too great

For valuable knowledge transfer to occur, individuals must have distinct information sets that are
relevant to each other

If there is too much information overlap, teamwork does not produce much knowledge transfer

If the information that one has is irrelevant to another, then knowledge transfer has no value

8
 KNOWLEDGE TRANSFER: AN EXAMPLE

Kate’s tasks

K T: Tor's information set

K: Kate's information set
Intersection: information shared by both
T Tom’s tasks

Whether teamwork is valuable depends also on informational requirements of the

two jobs
The information required to perform the tasks is represented by the areas in the
ellipses in solid curves
Both Kate and Tor have knowledge to perform about half of their tasks
By working as a team, Tor can transfer all the knowledge necessary for Kate to
perform her tasks, and she can do the same for Tor
Thus, Kate and Tor can obtain from each other knowledge that is necessary to
perform their tasks but is outside of their own knowledge set
With knowledge transfer, they can perform more tasks
9
 Now suppose that the information required to perform the
tasks is represented by the ellipses in dotted curves
Kate’s tasks

K Kate only has information to perform about half of her tasks

but the half that lies outside her task ellipse also lies outside
the information set of Tor
T
The same is true in reverse as well

The non overlapping parts of the information sets are

irrelevant, because they are not valuable to the other person

Tom’s tasks

Thus, teamwork can produce valuable knowledge transfer when two conditions hold:
1. Members of the team have idiosyncratic pieces of information that can flow among members when they are put in a
team together
2. The idiosyncratic information possessed by one member is valuable to some other team members
10
Mechanic’s
tasks

Accountant’s
tasks

An auto mechanic and a cost accountant do not make a good team:

their information sets are non overlapping but their idiosyncratic information is
irrelevant to each other
The mechanic is ignorant of many things needed to do his job (much of the mechanic's
task ellipse lies outside his information set), but those tasks also lie outside the
accountant information set
What the mechanic needs to know but does not know, the accountant does not know
either
The converse also holds

11
 Tasks

Consider two identically trained cost accountants, A and B

There is unlikely to be much knowledge transfer between them, even

B though each has information that is relevant to the other, because their
experiences and knowledge bases arc almost identical

A Neither can do all of the tasks, but because their information is so

similar, it is likely that those tasks that one cannot do cannot be done
by the other

12
 IMPLEMENTATION OF TEAMS

Optimal Team Size

Small teams do not allow enough information transfer because there are fewer opportunities for cross training

However, large teams create communication problems

Peer monitoring and norms to Reduce Free riding

Additional important factor in determining the optimal group size is the free rider effect
In small group, free riding is less a problem because peer monitoring can be effective in small groups
Incentives to punish a shirker are greater in a small group because if one member reduced effort has
significant effect on the earnings of other members

In a large partnership, peer monitoring is not as effective for two reasons

One member’s shirking does not affect another member’s interest as much Benefits of monitoring one's
peers are also reduced, because the benefits are shared with a larger number of partners; incentives to punish
shirking are reduced
More difficult to observe shirkers in large teams

Therefore, free riding effects are more prevalent in larger teams

13
Peer pressure can help reduce free riding in a way similar to peer monitoring
Peer pressure within the group often takes the form of a norm
Norms: informal policies or practices, or a set of beliefs, held by the majority of a group
The level of effort associated with norms depends on the type of sanctions imposed for deviating from the norm
Minor criticism will have little effect
Threat of ostracism is likely to have a strong effect

If the pressure lightens and enforcement lags, norms tend to break down

Firms also engage in activities that build group or company camaraderie

Many practices at firms may have more to do with creating empathy, loyalty, and latent guilt than they do with
performing the nominal tasks themselves

14
 TEAM COMPOSITION

Supervisors should assign individuals to teams when have more information than the individual workers

Teams should be free to choose their own members when workers have better information about one another than the
supervisor does (e.g. worked together for a long time, new employees are friends of current employees, individual
workers engage in tasks that are highly specialized and beyond the knowledge set of the supervisor)

Consider two mechanisms for team member selection:

1. Alternating draws (captain of each team choose members in alternating order)
2. Bidding for members (allow teams to bid for their team members; English auction: team that bids the highest wins
the member)

Preferences Efficiency Example:

Rank Team I Team II Rank Team I Team II 2 teams/4 new hires

Allison Brooke Allison David Both teams prefer B to C

Brooke Charles Charles Brooke Efficiency: A and C should work with Team I; B and D with Team
II
Charles David Brooke Charles
B is more valuable to Team II than to Team I
David Allison David Allison
Both teams prefer C to D, but C added value to Team I is greater
than to Team II

15
 Preferences
Rank Team I Team II
ALTERNATING DRAWS
Allison Brooke
Brooke Charles
Coin tossed and Team I wins
Charles David
Team I chooses A, Team II then chooses B, Team I chooses C, and
David Allison
Team II chooses D
Efficiency
Efficient allocation is achieved
Rank Team I Team II
Allison David Coin tossed and Team II wins
Charles Brooke Team II choose B, Team I choose A, Team II choose C, Team I choose
Brooke Charles D
David Allison Efficiency violated

Efficiency achieved depending on the outcome of coin toss

Not desirable property of any team selection scheme (it is not merely a feature of this example)
Allowing teams to choose members in an alternating fashion generally results in an inefficient allocation of team
members
 Preferences BIDDING FOR MEMBERS

Rank Team I Team II To attract a team member, team must give up some of its profit, which in
Allison Brooke turn affects team members' compensation
Brooke Charles According to the efficiency level, A's ability is best deployed in Team I
Charles David (contributes more to Team I's profit than to Team II's profit)
David Allison Therefore, Team I will be willing to bid a larger total amount of profits to
Efficiency acquire her than Team II
Rank Team I Team II When B is auctioned, the reverse is true
Allison David The same mechanism works for C and D
Charles Brooke Thus, A and C end up on Team I, and B and D on Team II
Brooke Charles
David Allison

Auctions generally result in an efficient allocation of resources (preferable to alternating draw methods)

Note: unnecessary to dictate the number of members that go to each team: if the value of acquiring the 3rd new worker is
higher to Team I than the value of a 2nd is to Team II, better that Team I have 3 new workers and Team II have only 1

Application of market metaphor inside the firm: market allocates resources well because individuals are motivated to bid
high when they have good uses for the resources
17
 PAYING FOR PERFORMANCE
 2008-2009: Financial crisis

 Incentives in many banks played a role (among several other causes)

 One of the main cause of bank failures was that they issued risky mortgages: when housing prices fell, may
mortgages went into default

 Many banks used incentive plans that encouraged mortgage loan officers to maximize the quantity of loan
issued, with inadequate attention to the risk of each loan

AIM

 Study pay for performance  one of the most motivational levers that a manger can pull

 Problem: employees tend to respond to incentives

 If incentive plan designed well: source of value creation

 If incentive plan designed poorly: source of value destruction

1
 Principal-Agent model
 Economic framework for analysing most incentives problems

 One individual (agent) acts on behalf of another (principal)

 Typically, P is the employer and A is the employee

 Problems arises because P and A may have different objectives  conflict of interest (e.g. firm owners want to
maximize firm value)

 Employee provides effort that affects firm value

 Effort: actions that employees can take and firm want to motivate (working harder, faster, longer, thinking more
carefully when making decisions, cooperating with colleagues, being helpful to customers…)

 Effort is costly to employees  employee exerts effort in the hope of earning rewards

2
 Employee’s contribution to firm value (ignoring cost):

! = !($)

 Firm profit attributable to employee:

! = ! $ − '()

 Conflict of interest: effort has a cost for employee (disutility of effort)

* = *($)

 Employees pay provides incentives if it depend on performance

'+ = ,$-./-0(12$ 0$(45-$

 If performance is measured perfectly

'+ = !

 However, in general

'+ = ! + 7 7 = ,$-./-0(12$ 0$(45-$0$18 $--/-

3
 Pay is a function of PM:

!"# = !"#(!&)

 Pay for performance is risky because performance can almost never be measured perfectly

 Is employees are risk adverse, there is an additional cost to the employee form working at the firm: the cost of
the riskiness of pay

 Certainty equivalent: amount a person is willing to pay to avoid the risk

1 !
() = - ./0
2

 Measure of risk is the variance of pay

 R = coefficient of absolute risk aversion (increasing with risk aversion)

 Employee net value from working for the firm:

1 !
!"# !& − ( 4 − - ./0
2
4
 Firm choose compensation plan Pay(PM) to maximize profit from employee with the constraint that total pay
must be at or above the employee’s labour market value

 Must compensate employees for effort cost C and for risk cost R generated by the incentive system

 Employee marginal cost from working harder:

∆"
!" =
∆%

 Employee marginal benefit from working harder:

∆'() ∆'() ∆'!

!# = = ×
∆% ∆'! ∆%

1. How does the PM vary with e? If it reflects e well, it generates good incentives

2. How does Pay vary with PM? If it does it strongly: incentives will be stronger

5
 C and R are cost that must be borne by the firm implicitly because they reduce the value of the job for the
worker

 worker will require higher total compensation as they increase

 Firm marginal cost from the employee working harder:

∆%
!" = ∆&

 Firm and employee have the same marginal cost from effort

 The source of conflict is that employee benefit (Pay) is generally not identical to the firm’s benefit

 Firm marginal benefit from the employer working harder:

∆)
!# =
∆*

 Which generally differs from MB of working harder for worker because evaluation does not reflect performance
perfectly or because pay does not reflect fully employee contribution

 These are the design challenges for incentives plan 6

 PERFORMANCE EVALUATION

 Most difficult part of any incentive scheme is performance evaluation

 If employees work in groups it is hard to disentangle who is responsible and who is not

 Employee’s performance may be partly due to luck

 Some contributions are difficult to quantify even when they are clearly observed (e.g. employee’s effects on group norms,
mentoring of junior staff, or customer satisfaction)

 Ideal performance measure should reflect employee’s total impact on firm value and nothing else

 If evaluation does not accurately reflect employee’s contributions, it may cause several negative outcomes:

 Employee might find that link between performance and pay is uncertain and require compensation for risk

 Employee might be poorly motivated or strongly motivated, but to do the wrong things

 Thus, even though evaluations are difficult and costly to do, they are a necessary component of a good reward
system
1
 Employee performance depends on, among other factors
 Innate abilities
 Accumulated skills or human capital
 Effort

 We focus on the use of evaluations to measure and motivate greater effort

 Evaluation might be quantitative or qualitative

QUANTITATIVE PERFORMANCE MEASUREMENT

 Examples: Hours worked or customer satisfaction

 Advantages:

 can be tied to compensation more easily (e.g., by computing a bonus using a formula)

 are often perceived as more objective than the use of judgment in evaluations

 We consider three properties that a performance measure should have

2
Risk Profile

 Every performance measure has some risk

 There are two types of risk:

1. Uncontrollable risk

 Corresponds to variation in performance that is beyond the employee’s ability to control (e.g. stock price)

 Because these variables cannot be controlled by the employee, but do influence the performance measure, they create risk
for the employee

2. Controllable risk

 Variation in the work environment that the employee has some ability to control

 With controllable risk the firm should increase the strength of incentives to motivate the employee (opposite with
uncontrollable risk)

 Therefore, the first question to ask about a performance measure is how much of the variation in this measure is
due to uncontrollable or controllable risk
3
Distortion

 Focused performance measures can distort incentives (e.g. custodian, measuring cleanliness of floors does not
motivate the custodian to be cost conscious; evaluating the CEO on accounting earnings is likely to cause the CEO
to take too much of a short-term focus)

 Distortion arises from unbalanced incentives across multiple tasks

 Virtually any performance measure will have some distortion

 Worth thinking carefully about them before putting too much weight on a measure for incentives

 Example of distorted incentives are:

 Performance measure that proxies for some tasks but not others, such as measuring quantity but ignoring quality
(intangible)

 Individual performance measure that reduce incentive to cooperate

 Performance measures that measure what just happened distort incentives for actions that have long term consequences
(new technology, brand name, employee training)

4
Manipulation

 Quantitative performance measures may be manipulated (e.g. custodian evaluated on pounds of trash hauled
away may motivate the custodian to bring trash to work)

 Manipulation occurs because actions are taken to improve the evaluation even when such actions do not improve
firm value

 In general firms often have to choose the scope of a measure

 narrow measures focus on fewer aspects of performance
 broad measures focus on more aspects of performance

Narrow Broad
(fewer factors) (more factors)

Less risk More Risk

More manipulated Less Manipulated
More distorted Less distorted

5
 SUBJECTIVE EVALUATION

 In many jobs, employees receive subjective rating once or twice a year (often on a scale of 1-5 or A-E)

Typical Performance Appraisal

Distribution
50% Concerns
40%  Grade inflation

30%  Managers reluctant to give lowest ratings because worried

20%
that this would reduce employee motivation
 Moreover, employees may pressure managers to change
10%
the rating which is unpleasant for the manager
0%
1 2 3 4-5

 Employees dislike subjective evaluations if they worry that evaluation reflects supervisor’s personal opinions and
biases and that the supervisor is playing favorites
 Reduce incentives, because the perceived link between effort and reward is weakened
 Despite all of these flaws, essentially every job uses subjective evaluations in important ways (hiring, promotion,
and termination decisions)
6
Why Use Subjective Evaluations?
 Improve consideration of risk in the evaluation
 A careful subjective performance evaluation defines what the employee is, and is not, responsible for

 Reduce Distortion and Manipulation

 Improve incentives for risk taking

 Easier for the manager to reward good results without simultaneously punishing mistakes
 Gives the manager greater flexibility to reward the up side without punishing the down side

 Give the incentive system more flexibility

 Incentive plans put in place at the beginning of the year may no longer be ideal if circumstances change
 Effective use of subjectivity means that changes to incentives are more likely to be accepted, because the employee
already expects judgment from the supervisor
 This makes it easier for the manager to tell the employee to change emphasis during the middle of one year, in effect
changing incentives
7
Why Use Subjective Evaluations?
 Expand Communication
 Best subjective evaluations occur implicitly every day as the manager works with the employee
 The manager monitors what the employee does and why, makes adjustments, and suggests improvements instead of
waiting until me end of the year to do the evaluation
 having these conversation all year long will improve the employee’s effectiveness and the working relationship with the
supervisor
 Improve training
 Subjective evaluation can provide lessons from the manager’s experience

8
Who Should evaluate Whom
 Decentralizing evaluations to an employee’s manager is risky because manager’s incentives will be imperfect
 Moreover, subjectivity makes favoritism and discrimination more likely
 Despite these problems, evaluations are not centralized because evaluations require knowledge that is very costly
to communicate to others
 Most evaluations are inevitably decentralized to the direct supervisor

 Obviously the firm has to worry about the incentives for the manager to implement the system in accord with the
firm’s interests instead of his own
 Supervisor’s incentives should be well designed
 Constraints might be placed on the evaluator
 Manager’s reputation can have substantial impact on how incentives play out in practice: if a manager has reputation
for fairness and careful evaluation, subjectivity will be easier to use effectively
 Firms should attempt to put judicious managers into positions where these considerations play a major role

9
Who Should evaluate Whom
 In addition, organizations usually put in place formal policies to establish greater fairness in the evaluation system
 Employees may have the right to ask to be re-evaluated if they disagree with their evaluation
 Supervisor’s boss may check evaluations
 Have multiple evaluators of an employee, reducing the likelihood that the ultimate evaluation is biased, because
different managers tend to have different biases
 Purpose of such oversight is to provide some incentives for the supervisor to do appraisals fairly

 Supervisor judgment can also distort employee incentives:

 Try to improve their evaluation not through effort, but through influencing the supervisor in other ways (spend time
and resources trying to lobby their boss)
 The worker may distort what he says to the manager and the quality of information the job may suffer
 Workers may have incentive to shade their reports in the direction of the manager’s initial opinion (Yes men
phenomenon)

10
’
 𝑥 =𝑒+𝜂
𝜂 (E 𝜂 = 0) 𝜈

𝑥̅ = 𝐸 𝑥 = 𝑒
 𝑤 = 𝛼 + 𝛽𝑥
𝑤 = 𝛼 + 𝛽(𝑒 + 𝜂)
𝛼 𝛽
𝑤 = 𝛼 + 𝛽𝐸 𝑥 = 𝛼 + 𝛽𝑒
E 𝜂 =0
𝜂 𝑤
 α>0 β=0
…

β
β

β=p and α < 0)

“ ”
α
 ’

𝑒
𝑈 𝑤 ,𝑒 = 𝑤 − 𝛿
2
𝛿

𝑀𝑈 = =1

𝑀𝑈 = = −𝛿𝑒

𝑢≥0

𝐸 𝜋 = 𝑝𝐸 𝑥 − 𝑤 = 𝑝 − 𝛽 𝑒 − 𝛼

𝑝−𝛽
𝑀𝑈 𝑤 + 𝑀𝑈 𝑒 = 0
𝑤 + (−𝛿𝑒) 𝑒 = 0

𝑤
= 𝛿𝑒
𝑒
𝑒∗ 𝑤 (𝑒 ∗ )
𝑤=𝛼 𝑤 =𝛼

𝑒∗ = 0
𝑤 = 𝛼 + 𝛽𝑥 𝑤 = 𝛼 + 𝛽𝑒

𝛿𝑒 = 𝛽

𝑒∗ =
 𝑒
𝑀𝑎𝑥 𝑈 𝑤 , 𝑒 = 𝑤 − 𝛿
2
𝑤 = 𝛼 + 𝛽𝑒

𝑒
𝑀𝑎𝑥 𝑈 𝑤 , 𝑒 = 𝛼 + 𝛽𝑒 − 𝛿
2

𝛽 − 𝛿𝑒

𝑒∗ =

𝛽>0
𝛽=𝑝
 ∗ ∗ ∗
𝑒∗
𝑈 𝑤 (𝑒 ), 𝑒 = 𝛼 + 𝛽𝑒 − 𝛿
2

𝑈[𝑤 (𝑒 ∗ ), 𝑒 ∗ ] ≥ 𝑢
𝑤 𝑒 ∗ = 𝛼 + 𝛽𝑒 ∗

𝛽
𝑤 𝑒∗ = 𝛼 +
𝛿

𝛽 𝛽
𝛼+ −𝛿 ≥𝑢
𝛿 2𝛿

𝛼+ ≥𝑢
 α and β

α and β
α β
β
 α
β,

α β

𝛽
𝛼=𝑢−
2𝛿

β
𝛼
β

𝐸 𝜋 = 𝑝𝑥 − 𝑤 = 𝑝𝑒 − 𝛼 − 𝛽𝑒 = 𝑝 − 𝛽 𝑒 − 𝛼

𝑒∗ =

𝛽
𝐸 𝜋 = 𝑝−𝛽 −𝛼
𝛿

𝛽
𝑝−𝛽 𝛽

𝛼 𝛽

𝛽
 β–

δ=1

1−

(2 − 𝛽)𝛽
𝛽

𝛽
𝛽
𝛽

𝛽
𝛽

𝛽=1

𝛽=1=

𝛽=𝑝
β–
α
𝐸 𝜋 = 𝑝 − 𝛽 𝑒∗ − 𝛼
𝛽 𝛽
𝐸 𝜋 = 𝑝−𝛽 − 𝑢−
𝛿 2𝛿
𝑝𝛽 𝛽 𝛽
𝐸 𝜋 = −𝑢+ −
𝛿 2𝛿 𝛿
𝑝𝛽 𝛽
𝐸 𝜋 = −𝑢−
𝛿 2𝛿
𝐸 𝜋 𝛽

𝑝 𝛽
− =0
𝛿 𝛿

𝛽=𝑝

α
β

𝑝
𝛼=𝑢−
2𝛿

𝑝𝛽 𝛽
𝐸 𝜋 = −𝑢−
𝛿 2𝛿

𝑝 𝑝
𝐸 𝜋 = −𝑢−
𝛿 2𝛿

𝑝
𝐸 𝜋 = −𝑢
2𝛿

𝑝
𝑢<
2𝛿
 𝑤 = 𝛼 + 𝛽𝑥
𝛽=𝑝

𝑝
𝛼=𝑢−
2𝛿

𝑝
𝑒∗ =
𝛿

𝑒∗ = 𝛼
 𝐸 𝜋 = 𝑝 − 𝛽 𝑒∗

𝛽
𝐸 𝜋 = (𝑝 − 𝛽)
𝛿
𝛽

𝛽<𝑝
𝛽 α

𝛽 𝛽=1=

𝛽
𝑀𝑎𝑥(𝑝 − 𝛽 )
𝛿
𝛽

𝛽 𝑝−𝛽
− + =0
𝛿 𝛿
𝑝
𝛽=
2
 𝛿
𝑈 𝑤 𝛽 , 𝑒∗ 𝛽 = 𝑤 𝛽 − 𝑒∗ 𝛽
2

𝛽 𝛽
𝑈 𝑤 𝛽 , 𝑒∗ 𝛽 = 𝛽−
𝛿 2𝛿

𝑝
𝑈 𝑤 𝛽 , 𝑒∗ 𝛽 = >0
8𝛿

𝛽=1
𝑤 = 𝛽𝑥
𝑝
𝛽=
2
𝛼=0

𝑝
𝑒∗ =
2𝛿
 𝑤

𝑒
𝑈 𝑤, 𝑒 = 𝑤 − 𝜆𝑉𝑎𝑟 𝑤 − 𝛿
2
λ
𝑤 = 𝛼 + 𝛽𝑥 𝑉𝑎𝑟 𝑤 = 𝛽 𝑉𝑎𝑟 𝑥 = 𝛽 𝜐

𝑒
𝑈 𝑤, 𝑒 = 𝛼 + 𝛽𝑒 − 𝜆𝛽 𝜐 − 𝛿
2

𝑒
𝑀𝑎𝑥 𝑈 𝑤, 𝑒 = 𝛼 + 𝛽𝑒 − 𝜆𝛽 𝜐 − 𝛿
2

𝑒∗ =
 𝑒∗
𝛼+ 𝛽𝑒 ∗ − 𝜆𝛽 𝜐 − 𝛿
2

𝑒∗ = 𝑈 𝑤 (𝑒 ∗ ), 𝑒 ∗ ≥ 𝑢

𝛽 𝛽
𝛼+ − − 𝜆𝛽 𝜐 ≥ 𝑢
𝛿 2𝛿

𝛽
𝛼+ − 𝜆𝛽 𝜐 ≥ 𝑢
2𝛿

𝛼+𝛽 ≥𝑢

𝜐 𝛼 𝛽
 𝛼 𝛽

𝛼
𝛼 𝛽

𝛽
𝛼+ − 𝜆𝛽 𝜐 = 𝑢
2𝛿

1 − 𝜆𝜐2𝛿
𝛼 =𝑢−𝛽
2𝛿
 𝛽
𝛽
𝛼

𝐸 𝜋 = 𝑝𝐸 𝑥 − 𝑤
𝐸 𝜋 = 𝑝𝑒 ∗ − 𝛼 − 𝛽 𝑒 ∗

𝑒∗ = 𝛼=𝑢− + 𝜆𝛽 𝜐

∗
𝑝𝛽 𝛽
𝐸 𝜋 = −𝑢− − 𝜆𝛽 𝜐
𝛿 2𝛿
 𝛽
𝛽

𝑝 𝛽
− − 2𝜆𝛽 𝜐
𝛿 𝛿
𝑝
𝛽=
1 + 2𝜆𝛿𝜈
𝛽

0<𝛽<𝑝
 PAY FOR PERFORMANCE WITH WAGE CONSTRAINTS

Baseline theory of optimal compensations suggests that properly designed incentive contracts can be very
effective in extract workers' effort

Benchmark limit case, when worker risk-neutral and no other constraints, optimal compensation scheme is
an extreme bonus scheme (franchising scheme: negative fixed payment up-front and bonus component
identical to the price)

In real life labour markets and real life organizations, it is very difficult to implement the franchising scheme:

Even if agents are risk neutral, payments from the workers to the firms are not realistic

Firms often have to offer a simple fixed wage independent of performance, sometimes negotiated outside
the firm control (minimum wage constraint or collective agreements)

What should the firm do when it is forced to pay a fixed wage independent of output?

How such a wage should be chosen?

1
 PAY FOR PERFORMANCE WITH WAGE CONSTRAINTS

We study the effect of

bonus scheme with a minimum guarantee
scheme in which a minimum wage is guaranteed but if workers do well they can obtain additional income

The main analytical novelty of this chapter is the emphasis on workers' heterogeneity

Two different effects emerge

Efficiency effects: improvements in performance and productivity of workers already in the firm
Sorting effects: change in the composition of the workforce in the firm

The second part of the chapter shows how a minimum guarantee with bonus scheme worked in a real life
case study (Safelite case, a window installer producer in the US that introduced a pay for performance
scheme with minimum wage guarantee)

2
 Heterogeneous Ability: The Set Up
Assume:
Risk neutral individuals and no uncertainty
Firm produce a homogenous product whose price is equal to 1
The quantity of output produced by an individual is x
Production depends on effort and ability
𝑥 𝑎 𝑒
Ability is a fixed individual parameter (differs across individuals) unobservable to the firm but known by the agent

For given effort, individual with larger ability produce a larger quantity of output

We also assume that the outside option faced by each individual grows with the ability of the agent
outside option= au
workers with larger ability have better outside option

3
 Utility depends positively on wage income and negatively on effort
Utility function for an individual with ability a' reads

𝑤 𝑖𝑓 𝑥 𝑎′
𝑈
𝑤 𝑖𝑓 𝑥 𝑎′

Indifference curve in such setting will refer to all the combination of x and w that gives the same utility to the
agent

Maximum output at zero effort

Each Individual has a maximum output level 𝑥 𝑎′ that can produce without exercising any effort
Beyond the zero effort output level 𝑥 𝑎′ , individuals need to exercise effort to produce more output
More able individuals have larger zero output level

4
 Indifference curves in the [w, x] space
Are fiat up to the maximum output at zero effort
Are upward sloping beyond x = a' (for obtaining more output, effort is needed and must be compensated with more
income)
Slope depends on ability: lesser able individuals have steeper indifference curves

Formal definition (indifference curve with utility level k):

𝑘 𝑖𝑓 𝑥 𝑎′
𝑤 𝛿 𝑥 𝑎
𝑘 𝑖𝑓 𝑥 𝑎′
2

Slope:

𝑑𝑤 0 𝑖𝑓 𝑥 𝑎′
𝑑𝑥 𝛿 𝑥 𝑎 𝑖𝑓 𝑥 𝑎′

5
Indifference curve for two different levels of utility k1 and k2 for two individuals with ability al and ah
Individual with ability al has a flat indifference curve up to x = al
Individual with ability ah has a flat indifference curves up to x = ah
Individual with low ability has steeper indifference curve:

𝑑𝑤 0 𝑖𝑓 𝑥 𝑎′
𝑑𝑥 𝛿 𝑥 𝑎 𝑖𝑓 𝑥 𝑎′
needs to be compensate with more wage to increase his output because more output (effort) is more costly for the
individual with low ability

6
 Fixed Wage with Minimum Output
The idea of a minimum level of output is simply to impose a minimum standard of quantity, so that if a
worker does not reach such minimum standard, the firm is not going to hire the worker, or the employment
relationship can be interrupted without cost
We label a job 𝑤, 𝑥 as a job that pays a fixed wage w and requires an output 𝑥

If the wage is outside the firm control, all it can do is to select a minimum quantity of output 𝑥 so that the
worker participation constraint is binding and operate if profits are positive
Assume that the firm pays a fixed wage w regardless of ability and requires some level of minimum output
𝑥 for working at the firm

7
Graphical analysis
Constraint
Wage w obtained only for quantity larger than 𝑥
Producing less than 𝑥 not allowed
Constraint is a step-function with the step at 𝑥

No employees want to produce more than 𝑥

𝑒∗ 𝑥 𝑎 → 𝑒∗ 0 𝑖𝑓 𝑎 𝑥

Individuals with high ability reaches utility k1 while individuals with low ability reaches utility k2
with k1>k2: an individual with larger ability must exercise less effort to produce 𝑥
larger level of utility for working at the firm

8
Yet there is another key dimension to consider: individuals with larger ability have also larger outside option,
and may not be interested in working in the firm
Thus need to solve the problem of which worker will self-select for the firm that we are analysing
Utility for an individual that works for this firm with ability a is

𝑤 𝑖𝑓 𝑎 𝑥
𝑈
𝑤 𝑖𝑓 𝑎 𝑥

The condition for working at a firm that pays according to the contract 𝑤, 𝑥 is

𝑤 𝑖𝑓 𝑎 𝑥
𝑎𝑢
𝑤 𝑖𝑓 𝑎 𝑥

More able individuals find easier to produce 𝑥 and thus need to exercise lower effort at the margin
For ability level larger than a = 𝑥 individuals do not even need to exercise effort to obtain 𝑥

9
The outside option of all individual increases linearly with a
This suggests that the condition above is satisfied for
𝑎 𝑎 𝑎

For individuals with utility<al and >ah, outside option (straight line
in the figure) larger than utility for working at the firm
Individuals in the positions al and ah are indifferent
Individuals between al and ah enjoy a pure economic rent (or a
positive surplus)
When a firm pays a job 𝑤, 𝑥 and there is heterogeneity,
a subset of workers enjoys a pure surplus by working at the
firm
There are three types of workers:
1. Not good enough for working at the firm (a<al)
2. Work at the firm (al<a<ah); Such workers enjoy a surplus inside the firm
3. Too good for working at the firm (a>ah) 10
NUMERICAL EXAMPLE
Consider a firm that it is forced to pay a fixed wage equal to w to all its workforce
Utility functions is
𝑤 𝑖𝑓 𝑎 𝑥
𝑈 𝛿 𝑥 𝑎
𝑤 𝑖𝑓 𝑎 𝑥
2
so that 𝛿 1
We also assume that 𝑢 ⁄ and that 𝑎 1, so that we initially keep ability constant
We label the output associated with the optimal effort (franchising) as 𝑥 and its value is
𝑥 𝑎 𝑒

𝛽 𝑝 1
𝑒 1
𝛿 𝛿 𝛿
𝑥 2
since 𝑎 𝛿 1

11
NUMERICAL EXAMPLE
What is the wage that the firm should fix? The wage must satisfy

𝛿 𝑥 𝑎 1 1
𝑤∗ 𝑎𝑢 1
2 2 2

This suggests that when ability is constant and equal to 1 a firm that pays 𝑤 1, 𝑥 2 behaves
efficiently

The profits of the firm are

𝜋 𝑝𝑥 𝑤 1

12
Assume heterogeneous workers and 𝑎 uniformly distributed between 𝑎 0 and 𝑎 4
To find which workers will choose to work at this firm that pays 𝑤 ∗ 1, 𝑥 2 the condition is
𝛿 𝑥 𝑎
𝑤∗ 𝑎𝑢
2
2 𝑎 𝑎
1
2 2
whose solutions are obtained by the following quadratic equation
𝑎 3𝑎 2 0→𝑎 1 and 𝑎 2
Individuals with a < 1 are not good enough for working in the firm
individuals with a > 2 are too good for working in the firm and will choose their outside option
Individuals in between the two values enjoy a rent (utility>outside option)

Heterogeneity implies that the firm that pays a fixed wage is no longer able to induce all workers to be just to
their outside option
Ex: Individual with ability 3/2
utility at the firm=7/8; outside option=3/4 rent=1/8
Firm profits are still 1
13
 PERFORMANCE PAY WITH A TWO-TIER WAGE SYSTEM
Consider a firm offering a minimum wage with minimum output attached 𝑤, 𝑥
The firm may implement a two tier wage system:
compensation structure that guarantees minimum wage to all workers &
give the option to have a bonus scheme to those particularly productive
We are interested in two dimension of the phenomenon
Incentive effects: How will change the performance of the individuals that with a fixed and egalitarian wage are
enjoying a rent
Sorting Effects: what will happen to the quality composition of the labor force inside the firm

The firm is currently paying with 𝑤, 𝑥̅

The firm is offering a performance pay 𝑤 that can be described by the general bonus scheme as
𝑤 𝛼 𝛽𝑥
All the individuals that produce more than 𝑥̅ can be eligible for the performance plan

14
 The fixed wage works as a step function
The constraint of a general bonus scheme is the
upward sloping line with intercept equal to α
The bold lines report the actual compensation
As long as the individual produces the minimum output
required, a wage equal to w is guaranteed, and only if the
pay with the bonus scheme is larger than such value the
bonus scheme is effective

The compensation is described by

0 𝑖𝑓 𝑥 𝑥̅
𝑐𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛
𝑀𝑎𝑥 𝛼 𝛽𝑥, 𝑤 𝑖𝑓 𝑥 𝑥̅

Where 𝑥̅ is output desired

15
Which individuals would choose to produce under the bonus scheme?
Let’s solve the problem for the quantity 𝑥 (but one-to-one relationship between effort and output, as
described by 𝑥 𝑎 𝑒)
Let’s find the quantity of output that maximizes the utility (call it 𝑥 )
𝑀𝑎𝑥 𝑈 𝑤 , 𝑥

𝛿 𝑥 𝑎
𝑈 𝑤 ,𝑥 𝛼 𝛽𝑥
2
FOC:

𝜕𝑈 𝑤 , 𝑥
𝛽 𝛿 𝑥 𝑎 0
𝜕𝑥
𝛽
𝑥 𝑎
𝛿
Individuals with different level of ability have different preferred quantity 𝑥

16
The key decision to be taken is for the individuals that under the performance scheme produce an amount
of output 𝑥 larger than the minimum
Assume that there is at least an individual with 𝑥 𝑥

Without bonus scheme individuals with different level of

ability choose the position C (no sense to produce more)
If the bonus scheme becomes available
Low ability individual chooses to remain at the point C (bonus
scheme associated to a lower indifference curve)
High ability individual moving to the point D increases utility

17
Two-Tier Wage System: Formal Solution
How ability influences the decision between bonus scheme and fixed wage?
An individual with ability level 𝑎 will choose the bonus scheme if
𝑈 𝑤 ,𝑥 ,𝑎 𝑈 𝑤, 𝑥̅ , 𝑎

Since 𝑥 𝑎, utility associated to the performance plan is

𝛽
𝑈 𝑤 ,𝑥 ,𝑎 𝛼 𝛽𝑎
2𝛿

individuals with larger ability level 𝑎 get larger utility from the bonus scheme
Utility associated to operating under the minimum guarantee (point C) is
𝑤 𝑓𝑜𝑟 𝑎 𝑥̅
𝑈 𝑤, 𝑥̅ , 𝑎 𝛿 𝑥̅ 𝑎
𝑤 𝑓𝑜𝑟 𝑎 𝑥̅
2
Utility associated to the minimum guarantee is at most w
Very able individual can obtain the minimum output x without any effort and therefore get utility w
Individuals with lower ability needs to exercise effort and reach a lower utility.
18
Two-Tier Wage System: Formal Solution
Since utility with the minimum guarantee has an upper bound level w, it is clear that there exist a threshold
level 𝑎 such that for all ability level above, individuals are better off in the bonus scheme
In formula:
𝑈 𝑤 ,𝑥 ,𝑎 𝑈 𝑤, 𝑥̅ , 𝑎 𝑓𝑜𝑟 𝑎 𝑎

Curved line (Fixed wage) reaches upper bound for ability

level beyond 𝑎 𝑥
PPP is upward sloping
Individuals beyond 𝑎 are better off in the bonus scheme
The most natural assumption is to have
𝑎 𝑎 𝑎

19
 Existing workforce partitioned into various categories
1. Individuals with ability level below al do not work in the firm
2. Individuals with ability 𝑎 𝑎 𝑎 operate under minimum wage
guaranteed even if bonus scheme is available: too costly, in terms of
effort, to produce in the bonus scheme range
3. Individuals with ability 𝑎 𝑎 𝑎 operate under the performance
plan if they have the option to do so: they were enjoying a rent under
the fixed scheme, but they would respond to the incentive given
(incentive effect)
4. Individuals with ability above ah are now interested in joining the firm:
the bonus scheme increases the fraction of the workforce for which it
is attractive to work at the firm
We can distinguish between
INCENTIVE EFFECTS: a minimum wage scheme with a performance option increases productivity of existing
workforce. This is a pure incentive effects, since a given workforce improves its average effort and productivity
SORTING EFFECT: A minimum wage scheme with a performance option increases the average ability of the
workforce 20
Two Tier System: An Example
Let's continue on the example of before
al=1 and ah=2
job [𝑤 2, 𝑥̅ 2]
Assume that α=-1/4 and β=1
Utility of the individual at the lowest minimum output required:

𝛿 𝑥̅ 𝑎 3
𝑈 𝑤 2, 𝑥̅ 2, 𝑎 1 𝑤
2 2
Utility of the individual with a bonus scheme:

𝛽 5
𝑈 𝑤 ,𝑥 ,𝑎 𝛼 𝛽𝑎
2𝛿 4

at the lowest support individual is better off with the fixed wage

For individuals at the upper support, 𝑈 𝑤 2, 𝑥̅ 2, 𝑎 2 2 while 𝑈 𝑤 ,𝑥 ,𝑎 2 9/4

individuals at the upper support prefer the bonus scheme
21
The individuals that is indifferent is the solution to

𝛿 𝑥̅ 𝑎 𝛽
𝑤 𝛼 𝛽𝑎
2 2𝛿

which becomes
2𝑎 4𝑎 1 0
whose solution above 𝑎 1 is

2
𝑎 1
2
Individuals sort in the following way
individuals with 𝑎 𝑎 do not join the firm
individuals with 𝑎 𝑎 𝑎 choose minimum wage
individuals with 𝑎 𝑎 𝑎 choose piece rate
individuals with 𝑎 𝑎 are now interested in joining the firm

22
This particular type of sorting depend crucially, among other things, on the parameter 𝛼 which is the implicit
cost for the job
The two conditions that we need for this type of sorting are
𝑈 𝑎 ,𝑤 ∗ 𝑈 𝑎 , 𝑃𝑃𝑃

𝑈 𝑎 , 𝑃𝑃𝑃 𝑈 𝑎 ,𝑤 ∗

The first condition requires 𝛼 sufficiently large while the second condition requires 𝛼 sufficient low
The implicit cost of the job must be
sufficiently high that low ability individual do not find it convenient
sufficiently low that some individuals would want to swap to PPP

23
 PAY FOR PERFORMANCE AT SAFELITE
Safelite is the largest car windows installer in the US
In 1994, it changed the salary structure of its car window operation
Up to January 1994 the salary structure was based on a hourly rate paid independently of the number of
windows installed (a minimum amount of production was required in order to continue to be employed at
Safelite)
Between 1994 and 1995 the salary structure changed, and it become linked to the number of car windows
actually installed
Goal: increase productivity without losing any workers
Two options
1. (homogenous workforce): Safelite may increase the minimum amount of output required by the firm , and leave the
salary structure based on a hourly wage (possibly increasing the hourly salary).
2. (heterogeneous workforce): Introduce a compensation based on output with minimum guarantee so that more able
individuals can work harder without penalizing less able individuals

24
 Safelite implemented option number 2, and introduced a plan that was called PPP (Performance Pay Plan)
It could rely on a very sophisticated and digitally based accountancy system allowing management to
monitor in real times who installed what component
Since PPP has been introduced over a period of 19 months across all branches, most workers have been
observed under both regimes (fixed time input and PPP) (ideal for studying the effect on individual
productivity)

Theoretical Predictions
The setting of the Safelite changes fits perfectly with the introduction of a two tier regime
There are thus two key theoretical predictions linked to the introduction of the PPP plan
INCENTIVE EFFECT. Output (or effort) of the workers should increase (or not fall) as the company moves to a PPP
system. Further, if there are able individuals in the workforce, the effort increases
SORTING EFFECT. The average quality of the workforce increases: lesser able individuals stay with the firm, in
addition it is possible to attract better individuals into the firm
Average productivity (as a consequence of the two previous effects) should thus increase
25
RESULTS

Average number of window installed (units per worker per day) increases by 0.54 units (20 percent
increase)
Variance of individual output increases
Labor cost for unit of output falls

Concern: At the same time as Safelite introduced the PPP there was an exception increase in demand
effect could be the result of the demand pick up
After controlling for this possible demand effect (time dummies in the basic average regression) the results
show that the increase in productivity increases by 44%
If there was a demand effect, such effect worked in the opposite direction 26
Possible Interpretations

The pure incentive effect

To obtain an estimate of the pure incentive effect, calculate the effect only for the individuals
operating under both regimes
The increase in productivity is equal to 22% (50% of the total increase in productivity)

Sorting effect
Calculate the effect of the increase in productivity for the individuals hired when the PPP was
already introduced
Data suggests that the productivity of the new hires (controlling for tenure or seniority) is much
larger. In other words, new hires are more productive

27