Vous êtes sur la page 1sur 14

lOMoARcPSD|2507885

Summary Notes Personnel Economics, lecture 1-7

Toegepaste micro-economie (Erasmus Universiteit Rotterdam)

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

Notes Personnel Economics


Lecture 1

Motivating workers:
- Pay-for-performance
- Threat of dismissal
- Seniority-based pay
- Non-monetary rewards: praise, respect, days off, impact.

Setting hiring standards


Is the best strategy to hire the best or to hire the cheapest people? Neither rule can be optimal for
all firms. If low-skilled workers are the cheapest, and every firm only hires low-skilled workers,
the high demand will increase the wage of low-skilled workers, and at a point they are not the
cheapest workers anymore. In equilibrium there is a wage spread such that no firm wants to
switch. The allocation of high-skilled/low-skilled labor depends on the firm’s production
technology.

Firm’s profits: π = Q - Nwi


N = Q / Qi  π = Q(1 – wi/Qi) profits are the highest when hiring workers with the lowest
cost per hour, not necessarily low-skilled workers.

Cheap labor is not necessarily low-cost labor. High-productivity labor is not always the most
profitable labor.

With asymmetric information while hiring workers, managers can:


- Guess the best choice: not very effective.
- Guess the relevant information, and perform analysis using these estimates.
- Experiment: can be very costly and time-consuming, not everything can be made clear by
an experiment.

Lecture 2

Adverse selection problem: Firms cannot readily observe an applicant’s productivity.

Ways to deal with adverse selection:


1 Screening by requiring credentials: Credentials should signal a worker’s ability in a
particular job. Credentials should not be too expensive or too difficult to obtain. However, it
should also not be too easy.
Trade-off: requiring more credentials increases the probability to get a suitable worker, but it
necessitates a higher wage offer or placing more ads.

2 Offering contingent contracts: Pay for performance. Piece-rate pay may result in self-
selection. The problem is: exact measurement of a worker’s output is often costly. Some firms
may not choose piece-rate pay, they attract low-skilled workers, but save on measurements costs.

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

Firms that offer piece-rate pay attract more high-skilled workers, but also pay higher wages and
bear greater monitoring cost.
3 Setting a standard and give a reward if standard is met. For example: initial probation
period. Pay little during probationary period to keep out the unskilled workers, and a lot after
probation to attract the skilled workers.

These three methods deal with adverse selection and can make sure the high-skilled worker is
hired, but it also includes two extra costs: higher wages, compared to low-skilled workers, and
measurement costs.

Lecture 3

Apart from adverse selection problems, performance pay can also reduce moral hazard
problems: it motivates workers to work hard.
Assumption: no adverse selection problem, all people are identical.
Moral hazard problem: behavior cannot be observed.
The optimal bonus is a commission of 100%

Moral hazard problems with commission < 100%


1. Taxi driver goes home too early (MR = MC at a lower quantity than with 100%
commission rate.
2. Taxi driver and customer can deceive the cab company.
Possible solution: monitoring, but it is imperfect and costly, 100% commission rate seems
the easiest solution.

Formal analysis
Workers like money (w) and dislike effort (e). c(e) is the cost of effort, c’(e) > 0, c”(e) > 0.
U = w – c(e)
c(e) = ½ θe2
q = e q is observable, e is unobservable
w = a + bq
U = V if worker decides to work somewhere else
p = price of product, given
π = pq – w
U = w – ½ θe2 ≥ V Participation constraint
Approach: backward induction

1. Determine how the workers behave under different wage schemes, taking as given that he
is willing to participate.
2. Derive the level of fixed compensation necessary to attract the worker for all wage
schemes.
3. Use these results to find the wage scheme that maximizes profits.

Step 1 – Worker’s behavior


Max. U = w – ½ θe2
Substitute a + bq for w

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

Max. U = a + bq – ½ θe2
Substitute e for q
Max. U = a + be – ½ θe2
dU/de = b – θe = 0
e* = b/θ

Step 2 – Participation constraint


U = a + be – ½ θe2
e* = b/θ
U = a + b2/θ – ½ θ(b/θ)2
U = a + ½ b2/θ
a + ½ b2/θ ≥ V
a + ½ b2/θ = V
a = V – ½ b2/θ

Step 3 – Finding the profit-maximizing wage scheme


π = pq – w = pq – a – bq
q=e
π = (p – b)e – a
π = (p – b)b/θ – (V – ½ b2/θ)
Max. π = (p – b)b/θ –V + ½ b2/θ
dπ/db = -b/θ + (p – b)/θ + b/θ = 0 1st part: wage costs ↑, profits ↓ 2nd part: effort ↑, profits ↑ 3rd
part: if b ↑, a can be reduced, profits ↑
(p – b)/θ = 0
p = b 100% commission rate

By setting the bonus equal to the full marginal product, the worker fully internalizes the effect of
his actions on the firm’s profits.
Efficiency: all mutually beneficial actions are taken!

π = (p – b)e – a = -a
-a = ½ p2/θ – V

A negative a is unnecessary when b is only given above a certain level of q. At the margin: bonus
should equal full marginal product.

Problems with performance pay


1. Moral hazard on employer’s side: manipulating figures or favoring some workers  pay-
for-performance usually based on sales/production rather than profits.
2. Other moral hazard problems may become more severe.
3. Risk-averse workers: pay-for-performance also makes wages dependent on uncontrollable
shocks: weather, business cycle, and luck. Most people don’t like volatility in earnings.
Volatility should be compensated. Trade-off between efficiency (high incentive pay) and
insurance (low incentive pay). Lower pay-for-performance also implies more volatile
profits. Risk is shifted, not eliminated.

Counterarguments
3

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

- When ‘firms’ are the shareholders, they can diversify risk by portfolio choice.
- When ‘firms’ are entrepreneurs, they may be better able to deal with risk as, usually, they
are richer than workers.
- Insofar as volatility is due to worker’s luck, firms can pool risks of different workers.

Lecture 4 – Education

Capital: assets that yield income and/or other useful outputs over longer periods of time.
Human capital: assets embodied in people, skills, knowledge, health.
However, human capital is not easily tradable. Some part of human capital is in the
characteristics of a person, the largest part is education.
Investment in human capital:
1. Early pre-school learning
2. Formal education
3. On-the-job training
4. Health improvements

1 Early pre-school training


Social skills and motivation: self-discipline, reliability, communicative skills. Partly innate, partly
determined by upbringing. These skills affect performance in formal education and have high
pay-offs in the labor market. However, government intervention is controversial, it is generally
the view that this is the task of the parents, not of the government.

2 Formal education
Benefit of education: increases future earnings. On average 10% higher wage for additional year
of schooling after controlling for ability. However, there are diminishing marginal returns.
Costs of education: tuition, books, and foregone earnings (largest (implicit) cost).

Optimal schooling choice


There are 2 period. Period 1: school or work. Period 2: work
J = unskilled wage. K = skilled wage. W > J C = tuition r = interest rate
Schooling is optimal when – C + K/(1 + r) > J + J/(1 + r)
(K – J)/(1 + r) > J + C Return must exceed direct and indirect costs of education.

More years of schooling is optimal, until:


T
K −J
R=∑ t tt =C +J 0
t =1 (1+r )

Diminishing marginal return of education. Increasing marginal cost of education.

3a General on-the-job training


Trainings is also useful for other firms, so current employer does not want to pay full training
costs if there is a high probability that the employee leaves the firm. If the employee leaves the
firm, he bears the costs and other firms receive the benefits.

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

If the employee will also benefit from the training, he should also pay (part of) the training costs.
In practice, we see payback clauses and low starting wages, to maintain the employee for the
firm.
Reciprocity (returning the favor) is quite important. Although reciprocity is self-reported, there is
evidence to assume honesty.

3b Firm-specific on-the-job training


Training only increases worker’s productivity at specific firm. Worker’s productivity outside the
firm remains the same.

Who should pay for training?


There is no need to raise the wage after the training since firm-specific on-the-job training is
useless to other firms. Therefore, all of the benefits from training pass to the firm. And, if so, so
should the costs. The firm earns a surplus after the training: the worker’s productivity is higher
than the worker’s wage. This surplus is not a true profit: it results from the training, which
entailed a sunk costs. The surplus is therefore called a quasi-rent.

The worker starts earning less than outside option. Later on, worker earns more  steep tenure-
earnings profile. The firm first makes a loss on the worker. Later on, it makes a profit. Clearly,
both firm and worker have an incentive to remain together  turnover rates are likely to be low.
External events may induce separation of the worker and the firm. At least one of the parties will
be worse off.
Investment in firm-specific training is therefore most attractive when turnover is likely to be low.
Firm will only invest when worker is likely to stay. Worker will only accept low starting-wage
when firm is likely to keep the worker.

4 Health
Physical fitness increases earnings.
Does beauty also pay off?
Possible reasons:
- Good looks may be valued by co-workers or customers
- Good looks may signal good health
- Good looks may improve self-confidence and hence productivity
- Discrimination by employers or fellow workers.

When attractiveness is based on rating by interviewer, beauty has a bonus of 10-15%, for both
men and women.

Lecture 5 – Promotions

Tournament theory
- Number of prizes/promotions to be awarded is usually fixed in advance.
- Relative performance matters, rather than absolute performance.

B: boss’s earnings
W: worker’s earnings B > W

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

Workers are in a ‘rat race’. Chance of being promoted increases in work pace, but no additional
promotion opportunities are created.
There are limits to worker’s effort: exhaustion, worker’s willingness to participate in the
tournament (participation constraint).

Firm’s optimal salary structure


- Bigger spread (B – W) increases effort.
- Pay higher W to ensure willingness to participate.

Luck may also affect a worker’s performance


- Production uncertainty
- Measurement error
- Manager’s bias
 Tends to reduce effort

A larger wage spread may help offset the effect of noise on effort; not necessarily at a higher cost
for the firm.

Formal analysis
B=W+Z earnings after promotion
W earnings without promotion
Z promotion bonus
qi = ei + εi
Promotion if: qi > qj
ei + εi > ej + εj
ei – ej > εj – εi

Simplified:
Promotion chance: pi = ½ + π(ei – ej)
π: measure of luck π = 0: only luck π → +∞: no luck
High π: stronger link between change of promotion and actual promotion.
Disadvantages: does not rule out pi < 0 or pi > 1; there is also a weak link between the two effort
levels.

Assumption: no salary in period 1.

Steps
1. Determine worker’s effort
2. Derive level of fixed compensation
3. Find wage scheme that maximizes profits

Step 1 Determine worker’s effort


U = pi(W + Z) + (1 – pi)W – ½ θei2
U = piZ + W – ½ θei2 substitute: pi = ½ + π(ei – ej)
U = [½ + π(ei – ej)]Z + W – ½ θei2
Maximizing:
6

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

dU/dei = πZ – θei = 0
ei* = πZ/θ

Step 2 Derive level of fixed compensation


Workers are identical: ei = ej pi = pj = ½ in equilibrium
EU = ½ Z + W – ½ θe2
e = πZ/θ
U = V when worker elsewhere
½ Z + W – ½ θ(πZ/θ)2 ≥ V
W = V – ½ Z + ½ π2Z2/θ Participation constraint

Step 3 Find wage scheme that maximizes profits


Effort is worth R to the firm
Profits = 2R – Z – 2W
Maximize profits subject to participation constraint: W = V – ½ Z + ½ π 2Z2/θ and worker’s effort
choice: e = πZ/θ

Profit = 2RπZ/θ – Z – 2[V – ½ Z + ½ π2Z2/θ]


dProfit/dZ = 2Rπ/θ – 1 – 2[-½ + π2Z/θ]
(1) (2) (3) 1: effort 2: cost of Z 3: cost of W
2Rπ/θ – 1 + 1 – 2π2Z/θ = 0
2Rπ/θ – 2π2Z/θ = 0
Z = R/π
e = πZ/θ = π(R/π)/θ = R/θ Noise does not affect the level of effort, noise is neutralized
θe = R Marginal cost to worker = marginal benefit to firm

When luck plays a larger role in promotions, the optimal promotion bonus will be bigger so as to
nullify its negative effect on effort. Expected wage need not rise: higher bonus permits a
reduction in base salary.

Advantage of promotions
Measurement
- Sometimes it is easier to measure relative performance than absolute performance.
- Performance pay → manager must determine all of the workers’ performance.
- Promotions → manager only has to rank the workers.

Common-luck effect
- Relative compensation eliminates the effect of common luck (the weather or the state of
the economy) on rewards.
- Performance pay → common good luck increases all of the workers’ rewards.
- Promotions → all of the workers’ performance is better due to common good luck,
rewards are unaffected!

Manager’s moral hazard


- With promotions, manager is forced to give a fixed number of rewards. Therefore,
manager’s mood or attitude matters less.
- So, a mean manager must give a reward to some workers.
7

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

- And nice managers are forced to differentiate, rather than to give all workers the same
reward.

Problems with promotions


Risk-averse workers
- Performance pay → pay dependent on own performance; many ‘small’ rewards.
- Promotions → pay dependent on relative performance; few ‘big’ rewards. Workers may
face more risk (but recall elimination of common luck-effect). Risk-averse workers need
to be compensated for this!

Too fierce competition


Reducing performance of competitors: refusing to help, share knowledge, cooperate or even
undermine coworker’s activities. This is clearly costly for the firm.

Lecture 6 – Team production

With team production, there is a trade-off between synergy effects (complementarity in


production, specialization, knowledge transfer) and free-rider effects (one bears the full cost, but
the benefits are shared).

Q = e1 + e2
wi = ½ PQ
Ui = wi – ½ θei2
P = 10
θ=1

Scenario 1: effort is verifiable


Uteam = U1 + U2 = 10Q – ½ e12 – ½ e22
Uteam = 10(e1 + e2) – ½ e12 – ½ e22
dUteam/de1 = 10 – e1 = 0
e1* = 10
dUteam/de2 = 10 – e2 = 0
e2* = 10
Q = 10 + 10 = 20
wi = ½*10*20 = 100
Ui = 100 – ½ 102 = 50

Scenario 2: effort is not verifiable


Wi = ½ 10(e1 + e2)
U1 = ½ 10(e1 + e2) – ½ e12
dU1/de1 = 5 – e1 = 0
e1* = 5 = e2*
Q = 5 + 5 = 10
wi = ½*10*10 = 50
Ui = 50 – ½ 52 = 37.5 < 50

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

When sharing output value with N people, one gets only 1/Nth share of the benefits, while
bearing the full cost. Workers free-ride on another’s effort.
The free-rider problem is more severe when a team consist of more workers.
U1 = (1/N)(e1 + (N – 1)ei) – ½ e12
dU1/de1 = (1/N)10 – e1 = 0
e1 = (1/N)10 the larger N, the less effort will be exerted.
Larger firms also induced free-riding in punishing shirkers, but this is also costly for the co-
worker. However, he only obtains 1/Nth of the increase in productivity of the other worker. This
leads to an additional free-riding problem.

How to resolve free-rider problem? (Not all work)


1. Different sharing rule. Does not work.
2. Hierarchy. May or may not work.
3. Relying on teammate’s kindness. May work, but tricky.
4. Peer pressure, norms and altruism. May work.
5. Reputation. May work under some conditions.
6. Very high team bonuses. Works in theory.

2 Hierarchy
w1 = w2 = 100 if Q = 20 w1 = w2 = 0 otherwise.
A third party (boss) is needed.
There are 2 equilibria: (e1, e2) = (0, 0) or (10, 10). A boss is needed to resolve the free-rider
problem

4 Peer pressure, norms and altruism


Norm may develop to exert high effort. Psychological cost of the norm: guilt (internal), and
shame social punishment (external). Friendship may also help.

5 Reputation
Repeated interaction may resolve free-ride problem, while keeping equal split. Usually
problematic if there are other potential partners who cannot observe past behavior.

6 Very high team bonuses


Again, a boss is needed.
Suppose boss commits to pay total value of team output Q to each worker. This solves the free-
rider problem, since the worker fully internalizes the effect of his effort on team’s interest.
However, workers must pay a large sum in advance to the boss.
f = 10Q(N – 1)/N

Research in defense of teams


- 14 percent increase in productivity on average!
- A few teams realized higher productivity than productivity of highest ability member →
Team productivity effect!
- Teams with greater spread in ability are more productive → learning in teams.
- Surprisingly, some high-ability workers earn less after (voluntarily!) joining a team.
- Team work apparently yields non-pecuniary benefits of less boring work, more social
interaction, and income smoothing.
9

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

- High-ability workers may, in addition, receive respect and joy from leading teams.

Lecture 7 – Seniority based pay and benefits

Wages increase with tenure. Even though experience is beneficial to productivity, ability
deteriorates. However, wages do not decline with age. Why do employers want to keep more
expensive, but less productive workers?
1. Selecting the right type of worker (adverse selection problem).
2. Firm-specific investment (hold-up problem).
3. Incentive reason (moral hazard problem).
Problem: workers in dead-end position (close to retirement, or not very productive) are difficult
to motivate. Solution: pay a low initial salary, and pay a high salary later on. This is effective for
retaining workers. They have a low incentive to leave the firm, because then they forgo a higher
wage.

Simple model
A worker can either work or shirk.
Probability that shirking is detected = π
When caught: worker gets no wage and is fired.

1 Old worker, in the last year of his working life.


Working: U = W – C
Shirking: U = (1 – π)W + π*0
Old worker works when W > C/π
The smaller the probability of detection, the higher the wage should be.

2 Young worker
V = discounted value of keeping the job after this period. For old worker: V = 0
For the young worker, V > 0 if
- Worker prefers to keep working for the same employer.
- Or when he needs good references when searching for a new job.

Working: U = W – C + V
Shirking: U = (1 – π)(W + V) + π*0
Young worker works when W > c/π – V

This leads to the wage profile with high-tenure earnings. A low base salary for young workers, a
high base salary for old workers.
When V > 0, younger workers are easier to motivate than older workers.

Over all periods, the contract must ensure that:


- Worker is willing to work for the firm.

10

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

- The firm does not make a loss.


- Worker has an incentive to work in all periods.
A rising tenure-wage profile may ensure this (Figure 11.1).

Problems with steep wage-tenure profile:


- Firm may go bankrupt
- Firm may fire older workers → reputation
- Old workers may want to work too long → mandatory retirement
Mandatory retirement is illegal in some countries (US), but a clever design of pension plans may
be a substitute.

Benefits
Advantages of offering benefits as compared to cash grants:
- Economies of scale in buying the benefits.
- Tax arbitrage opportunities: Benefit is counted as a cost for the firm, but (in some
countries) not as income for the worker.
- Benefits may reduce employees’ cost of effort (θ).

Disadvantages of offering benefits as compared to cash grants:


- Overconsumption of freely provided benefits, value to workers may be smaller than the
cost to the firm.
- Workers differ in their valuation of the benefits. For some groups, level of benefits is too
low, while for other groups the level of benefits is too high. However, this is also an
opportunity: benefits as a screening device to solve adverse selection problem.

Lecture 8 – Nonmonetary compensation

Positive job attributes: interesting work, nice colleagues, flexibility, and autonomy.
Negative job attributes: risk of injury, stress, and commuting.

Compensating wage differentials


Workers with not so nice jobs get compensated better. The firm that offers a not so nice job is
thus penalized, therefore they have an incentive to make the job more enjoyable. Wages will
reflect job attributes, this is most likely to work in competitive labor markets with few
restrictions.
Compensating wage differentials are often hard to observe. Employees with higher wages usually
enjoy better working conditions. It is thus not right to simply compare two employees, the ones
with higher wages are usually higher skilled. If job attributes are normal consumption goods, we
would expect people with higher incomes to have more enjoyable jobs. To discern compensating
wage differential for a particular attribute, we need to compare comparable employees in
different jobs.

Heterogeneous preferences
The wage differential reflects the preferences of the marginal worker.
Regulation of risky jobs tend to reduce welfare, since workers are forced to take a safer job,
which pays less.

11

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

However:
1. Government may know better than workers about the true risks and true costs (but then,
why not just tell them?).
2. Externalities, e.g. when collectively financed medical costs (but then, why not condition
premium on occupation?).
3. If people are not rational, there is room for government intervention.
However, government intervention does not always improve the situation, misguided intervention
may even worsen it.

Worker power
Research: employees with more power have higher productivity and higher job satisfaction. Self-
employed people have lower earnings, but are happier, the autonomy seems to play an important
role.
Governments and trade unions sometimes impose some degree of worker involvement in
decision-making on the firm through employee councils, sometimes enforced by collective labor
agreements.

Cost of employee councils


- Time of manager and workers.
- Workers may exploit information to their advantage: they may require high wages or
exert low effort, knowing that their position is safe.

Better informed workers may also be a benefit: Communicating bad news is easier through a
well-informed works council. Profit-maximizing firms may therefore have an incentive to
empower workers.

Simple model
Good state of nature with probability p
Bad state of nature with probability 1 – p
Workers can demand high or low wages, but do not know state of nature, but firm does know.

Payoffs
Workers claim
State High wage Low wage
Good Uh, πm Ul, πh
Bad U, π Ul, πm

Worker’s payoff: Uh > Ul > U


Firm’s payoff: πh > πm > π

Suppose the workers trust the firm’s statements.


Workers will set low (high) wage when the firm claims that the state is bad (good). → Firm will
always claim “state is bad”!
Therefore the firm cannot be trusted.

12

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)


lOMoARcPSD|2507885

Workers will always claim low wages if Ul > pUh + (1 – p)U


Workers will always claim high wages if pUh + (1 – p)U > Ul
p > (Ul – U) / (Uh – U)

Firm’s expected profits if p < p* (workers always claim low wages):


E(π) = pπh + (1 – p)πm

Firm’s expected profits with information disclosure:


E(π) = pπm + (1 – p)πm = πm There is no incentive to empower workers.

Firm’s expected profits if p > p* (workers always claim high wages):


E(π) = pπm + (1 – p)π
Firm’s expected profits with information disclosure:
E(π) = pπm + (1 – p)πm = πm Expected profits are always higher when workers are empowered.

When workers are aggressive claimers, firm can better commit to always disclose information.
Employee council can be a means to credibly communicate information.

Communication from workers to management


A works council can also facilitate communication in the other direction. Workers may
communicate truthfully to managers when they have some power over the way the information is
used.

Suppose two current workers earn some rent: U = w – c > V


Suppose the workers have discovered a way to perform task with one worker. If they tell the firm,
one worker will be fired.
E(U) when telling: ½ V + ½ U < U If U > V, there is no incentive to tell the firm.

Suppose the workers are empowered and can claim dismissal pay b for the fired worker.
E(U) when telling: ½ (V + b) + ½ U < or > U
b ≥ U – V is sufficient.

Firm’s profits when workers are not empowered:


π = Z – 2w
Firm’s profits when workers are empowered:
π=Z–w–b
Firm is better off if b < w

Workers who want empowerment


- Older workers under seniority-based incentives.
- Workers who have (implicitly) paid for firm-specific training.
- Union workers
- When U > V.

13

Distributing prohibited | Downloaded by Nakata Aleica Wilson (kaithteen@gmail.com)

Vous aimerez peut-être aussi