Title: Review
1Review
- Given the scenario below, the firm should hire
- A. High school educated workers
- B. College educated workers
- C. A mix of both
isoquant
Qhs
isocost
Qcollege
2Chapter 2
3Opinion poll
- Do applicants to highly skilled jobs tend to be
qualified for these jobs? - A. Yes, almost always
- B. No, they are often not qualified
4Opinion poll
- Do applicants to highly skilled jobs tend to be
qualified for these jobs? - A. Yes, almost always
- B. No, they are often not qualified
- CNN Article Many employers consider more than
half of all applicants to be unqualified. - Adverse Selection
5Reading for next class
- http//www.bepress.com/cgi/viewcontent.cgi?article
1097contextev - Or
- http//www.econ.ucsb.edu/babcock/EconomitVoice.pd
f
6Chapter 2 Motivation and Themes
- In the last chapter, we looked at whether firms
should hire high or low skill workers. But - Say, the firm has decided it wants skilled
workers - How does it go about hiring them?
- Big problem Adverse selection
7Hiring the Right People
- How do you find the most able people for a given
job? - High wages?
- Suppose worker prod in all firms is equal and
they know productivity but firms do not - Say firm pays the average wage
- Will it get workers who are
- A. Better than average
- B. Less than average
- C. Average
8Adverse Selection
- If firm offers 20/hr then only workers with
lt20/hr productivity are apt to apply - This is called Adverse Selection
Distribution of Productivities
Density
W
9Solutions
- Pay to screen workers
- Often impractical
- Offer probation wages
- Require credential
10Probation wage
- If you cannot observe a workers true
productivity, then - Use low initial wage to keep low-productivity
workers away. - Offer big raise to those who are not fired after
first period (to attract productive workers)
11Model of Probation Wage
- Assume
- 2 periods, interest rate0.
- Unskilled wage (outside option) is WU
- Skilled wage (outside option) is WS WU G
- Your firm pays W1 in period 1, W2 in period 2
- How do you choose W1 and W2?
- Goal 1 Attract skilled
- Must offer SUM over 2 periods that at least
matches what skilled worker could get elsewhere - W1W2gt2WS
- For simplicity set W1W22WS
- Goal 2 Repel Unskilled
- W1ltWU
- For simplicity W1WU
12Optimal Strategy
- Goal 1 Attract skilled
- Must offer SUM over 2 periods that at least
matches what skilled worker could get elsewhere - W1W2gt2WS
- For simplicity set W1W22WS
- Goal 2 Repel Unskilled
- W1ltWU
- For simplicity W1WU
- Just substitute in
- W1W22WS
- WUW22WS
- W22Ws-Wu
- W22(WuG) Wu
- W2Wu2G
- Proposition 1 As premium G earned by skilled
workers in the general market rises, the gap
between period 1 and period 2 wage must rise.
13Graphical Representation
- In practice, could offer W1WU-e, W2Wu2G e
WSG
WS
WS-G WU
Period 1
Period 2
14Extension
- Assume P prob unskilled worker doesnt get
detected - Goal 1 Attract skilled
- W1W22WS,
- Goal 2 Repel Unskilled
- W1PW2 (1-P) WUlt2WU
- Substitute in
- W12WS- W2
- (2Ws-W2)PW2(1-P) WUlt2WU
- (P-1) W2lt2WU-(1-P) WU- 2WS
- W2gt2WU-(1-P) WU- 2WS
- (P-1)
- W2gtWUPWU-2WS
- (P-1)
- W22WS-WU-PWU
- (1-P)
- GW2-W1
- GW2-(2WS- W2)
- G2W2-2WS
- G2 (2WS-WU-PWU) -2WS
- (1-P)
- G4WS-2WS(1-P)-2(1P) WU
- (1-P)
- G2WS2WSP-2(1P) WU
- (1-P)
- G2 (1P) (WS-WU)
- (1-P)
15Interpretation
- G2 (1P) (WS-WU) (1-P)
- If P goes up what happens to G?
- A. G?
- B. G?
- C. Cant tell
- P??G?
16Punchline
- P??G?
- Proposition 2
- As the probability that less productive workers
will not be detected rises, the probationary wage
must fall and the wage paid after the end of the
probationary period must rise. - Intuition
- If unskilled workers now have a CHANCE to get the
W2 payoff, must reduce the wage they now would
get for sure in period 1 and increase wage paid
in period 2, to lower the overall expected
payment they get. (Increase in period 2 wage is
necessary to keep skilled workers happy).
17Question 1
- I am hiring workers. Im paying a probation wage
in the first period. I want only skilled workers
to apply, but Im still getting some unskilled
workers. - Should I pay a lower probation or higher
probation wage? - A. Lower
- B. Higher
WSG
WS
WS-GWU
Period 1
Period 2
182) Education credential as signal
- Mechanism
- Find an observable trait tied to productivity,
e.g. education - Credential (or signal) should be harder for
low-productivity workers to acquire - If wage gap between those with and without the
credential is small, only most able will find it
worth the effort to obtain.
19Hazing
- http//www.youtube.com/watch?v2BLV1vdqzxgfeature
related
20Opinion Poll
- If you had to choose, which would you rather
have? - A. All the knowledge you gained in college, but
no college degree - B. The college degree, but none of the knowledge
you gained
21Signaling
- Assumptions for Signaling Theory of Spence
- 2 ability types, high and low.
- VMPhigtVMPlo
- Asymmetric information
- Firms dont know ability type
- Workers do
- More able workers have lower costs of schooling
- Lower effort costs, enjoy school more, find it
easier, have more time for work or other things
22Equilibrium Notion
- Firms best-respond to workers, workers
best-respond to firms. - Both end up correct in their beliefs about what
other will do. - Example Equilibrium Demonstrations in former
Soviet Union - You get shot if you show up and no one else does.
- You overthrow the government if everyone shows
up. - I believe no one else will show up, so I dont
show up. - Everyone else believes no one will show up
- No one shows up.
- Equilibrium. Beliefs are not falsified
- Example Not an equilibrium Roommates
- I believe you will do the dishes, so I dont do
them - You believe I will do the dishes, so you dont do
them - No one does the dishes
- Not self-sustaining. Beliefs will be falsified
23Spence Signaling Equilibrium - Graphically
- Firm strategy
- Pay VMPH to those who get s yrs of schooling
- Belief these will be high types
- Pay VMPL to those who dont
- Belief these will be low types
- Workers strategy
- Get s yrs if high type
- Get 0 yrs if low type
- Belief firms will pay VMPH to those with s yrs,
VMPL to others
- CHltCL
- VMPHProd of high types
- VMPLProd of low types
CL
VMPH
CH
VMPL
s
24Spence Signaling Equilibrium - Graphically
- Is there s that makes both workers and firms
strategies a best choice, and their beliefs
correct?
- CHltCL
- VMPHProd of high types
- VMPLProd of low types
CL
VMPH
CH
VMPL
s
s
25Spence Signaling Equilibrium - Graphically
- Why does s work?
- Low types get
- VMPL-0 if no schooling
- VMPH-C0 if s years
- These values are equal
- Deters low skill types from s yrs (degree)
- High types get
- VMPL-0 if no schooling
- VMPH-C1 if s years
- VMPH-C1gtVMPL
- So high types get s yrs
- Firm is right. Workers are right. Beliefs not
falsified
- This is called a separating equilibrium.
CL
VMPH
CH
C0
VMPL
C1
s
s
26Spence Signaling Equilibrium - Graphically
- Not an equilibrium, because firms beliefs
falsified
- What if s chosen as cut-off by firms?
- Do low types choose to get s years of schooling?
- A. Yes
- B. No
- C. Cant tell
- Yes, because get more utility than from VMPL.
- Effort cost not high enough to deter them
- Firm would end up paying low types VMPH
- But their prod is VMPL
- Firm loses money, goes out of business
CL
VMPH
CH
VMPL
s
S
27Spence Signaling Equilibrium - Graphically
- What if s is chosen as cut-off by firms?
- Do low types choose to get s years of
schooling? - A. Yes
- B. No
- C. Cant tell
- No -- get more utility from VMPL.
- Do high types get s years of schooling?
- Yes -- get more utility from (VMPH-effort cost).
- Many equilibria exist
- S was the lowest s that still resulted in a
separating equilibrium
CL
VMPH
CH
VMPL
s
S
28Spence Signaling Equilibrium - Graphically
- What if s is chosen as cut-off by firms?
- Do low types choose to get s years of
schooling? - No, because get more utility from VMPL.
- Do high types s years of schooling?
- A. Yes
- B. No
- C. Cant tell
- No, because get more utility from VMPL.
- Nobody gets schooling
- Pooling Equilibrium
- (Uninteresting case)
CL
VMPH
CH
VMPL
s
S
29Notes - Spence Signaling Equilibrium
- Here, schooling does not increase productivity
- Only gives evidence of pre-existing ability
- Schools as social evil
- Expensive signal. Resources thrown away.
- Is there a better way?
- Should we outlaw discrimination on the basis of
education?
CL
VMPH
CH
VMPL
s
s
30Is Schooling Only a Signal?
- Counterarguments
- Vocational aspects of education increase
productivity - Surgeon
- Engineer.
- Wouldnt firms/workers both be better off with
contingent contract? - Worker accepts lower wage in period 1, firm
agrees to raise wage if worker high ability - Cheaper for everyone than wasting 4 years in
college!
CL
VMPH
CH
VMPL
s
s
31Empirical Evidence
- Goes both ways
- Wolpin
- Self-employed acquire just as much ed as
employees. - Why, unless it increases prod?
- Who do they impress??
- Weiss
- Found dropouts and high school grads equally
productive, but lower quit rates for grads. - Implies signaling.
CL
VMPH
CH
VMPL
s
s
32When is education more likely to raise prody?
- When specific skills taught map closely and very
directly to skills used on the job - Engineering
- Law
- Medicine
- MBA
33Question
- In which setting is college degree more apt to
capture productivity gains achieved through
education and less apt to be purely a signal of
pre-existing ability? - A. Brain surgeon
- B. Car salesman
34Signaling Extenions Fraternities and Hazing
- Frats provide benefits
- But need members to provide public goods
- Participate enthusiastically
- Organize parties, etc.
- Worry about free-riders
- Cant tell if applicants are free-riders (F) or
committed (C) - Use hazing H to get applicants to reveal types
- CFgtCC
CF
VMPC
CC
VMPF
H
H
35Signaling Extensions - Gangsters
- Your run a crime gang
- Two types Mob types and FBI informants
- Potential gangsters know their types but you do
not - Cost of committing violent crime higher for FBI
types - Require v violence before you will hire worker
CFBI
VMPM
CM
VMPFBI
V
V
36How many years of schooling above should firms
require for the high wage job?
Clo
VMPhi
Chi
VMPlo
s
1
2
3
4
5