Transcript Games Results

Great Leap ‘Forward’
wishful thinking, poor incentives, hungry
Table 1.3: How to cripple an agricultural economy: Statistics during China’s “Great Leap Forward”
End (1959)
Start (1957)
Forecasted output
195 MM tons
525 MM tons (1958)
Grain output
195 MM tons
170 MM tons
Grain imports
2.1 MM tons
3.9 MM tons
Grain procured from
communes
46 MM tons
64 MM tons
Grain kept by rural areas
273 kg/capita
193 kg/capita
Rural per-capita calorie
consumption
2100 calories/day
1500 calories/day (1960)
% of provinces where workers
can not exit communes
20% (1955)
60% (1957)
Centralization, I
Table 1.2: Failure rates in public and private expeditions to the North Pole and Northwest Passage (Karpoff, 2001)
Private (n=56)
Public (n=35)
Crew deaths (%)
8.0
6.2
Number of ships (number lost)
1.63 (.53)
1.15 (.24)
% with scurvy
46.7%
13.2%
Unintended consequences
•
•
•
•
•
•
...A law enforcement source in Chicago said police see some evidence of soldiers working
with gangs here. Police recently stopped a vehicle and found 10 military flak jackets inside. A
gang member in the vehicle told investigators his brother was a Marine and sent the jackets
home, the source said. (from Sun-Times)
"We're lowering our standards," [Defense Department gang detective Scott] Barfield said.
"A friend of mine is a recruiter," he said. "They are being told less than five tattoos is not an
issue. More than five, you do a waiver saying it's not gang-related. You'll see soldiers with a
six-pointed star with GD [Gangster Disciples] on the right forearm."[....]
Of particular concern are reports that the Folk Nation, consisting of more than a dozen gangs
in the Chicago area, is placing young members in the military in an effort to gather
information about weapons and tactics, said FBI Special Agent Andrea Simmons, who is
based in El Paso, Texas.
"Our understanding is that they find members without a criminal history so that they can join,
and once they get out, they will have a new set of skills that they can apply to criminal
enterprises," Simmons said. "This could be a concern for any law enforcement agency that
has to deal with gangs on a daily basis.“
According to the Tribune, “nearly every one of the cases that we have looked into, it is a
young man or woman who thought that the symbol looked cool," said Christopher Grey,
spokesman for the Army's Criminal Investigation Command. "We have found some people
even get gang tattoos not really knowing what they are, or at least that they have not had any
gang affiliation the past."
BEM 146: Some simple games
• Cognitive hierarchy approach
–
–
–
–
Iterative
easy to compute
Captures individual differences
Explains when Nash succeeds and fails
• Nash equilibrium
– Players’ maximize, beliefs are accurate (no surprise
when results are announced)
– End of a learning process
• Quantal response equilibrium (Palfrey, Goeree)
– Nash+stochastic choice
The thinking steps model
• Discrete steps of thinking
• Step 0’s choose randomly
K-step thinkers know proportions f(0),...f(K-1)*
Normalize g(h)=f(h)/ h=0K-1 f(h) and best-respond
A j(K)=m (sj,sm) (Pm(0) g(0) + Pm(1) g(1)+... Pm(K-1)g(K-1))
logit probability P j(K)=exp(Aj(K))/ hexp(Ah(K))
• What is the distribution of thinking steps f(K)?
*alternative:
K-steps think others are one step lower (K-1)
Poisson distribution of thinking steps
Working memory bound  f(k)/f(k-1)1/k
 f(K)=tK/et K!
84 games: median t=1.65
Heterogeneous (“spikes” in data)
Steps > 3 are rare (Keynes, Binmore, Stahl et al)
Steps can be linked to cognitive measures
Poisson distributions for
various t
frequency
•
•
•
•
•
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
t=1
t=1.5
t=2
0
1
2
3
4
number of steps
5
6
Dominance-solvable game (Nash,
1,1; CH .73,1)
Top(.86)
ROW
Bottom
COLUMN
Left(.95)
Right
30, 20
10, 18
20, 20
20, 18
Dominance-solvable game results
L
R
Nash
CH
Data
06
Data
07
T
30, 20
10, 18
1.00
.73
.81
.86
B
20, 20
20, 18
.00
.27
.19
.14
Nash
1.00
.00
CH
.89
.11
Data 06
.95
.05
Data 07
.95
.05
Keynes’s “beauty contest analogy”
• Professional investment may be likened to those newspaper
competitions in which the competitors have to pick out the six
prettiest faces from a hundred photographs, the prize being awarded
to the competitor whose choice most nearly corresponds to the
average preferences of the competitors as a whole; so that each
competitor has to pick, not those faces which he himself finds
prettiest, but those which he thinks likeliest to catch the fancy of the
other competitors, all of whom are looking at the problem from the
same point of view. It is not a case of choosing those which, to the
best of one's judgment, are really the prettiest, nor even those which
average opinion genuinely thinks the prettiest. We have reached the
third degree where we devote our intelligences to anticipating what
average opinion expects the average opinion to be. And there are
some, I believe, who practise the fourth, fifth and higher degrees.
(Keynes, GTMEI)
relative
frequencies
Beauty contest results (Expansion,
Financial Times, Spektrum)
average 23.07
0.20
0.15
0.10
0.05
0.00
numbers
22
100
50
33
num be r choice s
97
89
81
73
65
57
49
41
33
25
17
9
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
1
predicted frequency
0
(2/3) of average (BEM 07 ave.28)
Table: Data and estimates of t in pbc games
(equilibrium = 0)
data
steps of
subjects/game
mean
std dev
thinking
game theorists
19
21.8
3.7
Caltech
23
11.1
3.0
newspaper
23
20.2
3.0
portfolio mgrs
24
16.1
2.8
econ PhD class
27
18.7
2.3
Caltech g=3
22
25.7
1.8
high school
33
18.6
1.6
1/2 mean
27
19.9
1.5
70 yr olds
37
17.5
1.1
Germany
37
20.0
1.1
CEOs
38
18.8
1.0
game p=0.7
39
24.7
1.0
Caltech g=2
22
29.9
0.8
PCC g=3
48
29.0
0.1
game p=0.9
49
24.3
0.1
PCC g=2
54
29.2
0.0
mean
1.56
median
1.30
CDF of bids for $10 with n=2 bidders (bid2) and
n=5 (bid5)
Nash: Bid $10 CH: 1-step bid $5 (n=2), $8 (n=5)
Thinking steps in entry games
• Entry games:
Enter or stay out ($.50)
Prefer to enter if n(entrants)<c (earn $1);
not enter n(entrants)>c (earn 0)
All choose simultaneously
• Experimental regularity in the 1st period:
Close to equilibrium prediction n(entrants) ≈c
“To a psychologist, it looks like magic”-- D. Kahneman
‘88
Thinking steps in entry games
How entry varies with capacity (c) ,
experimental data and thinking model
1
0.9
0.8
% entry
0.7
0.6
entry=capacity
experimental data
0.5
t=1.25
0.4
0.3
0.2
0.1
0
0
0.1
0.2
0.3
0.4
0.5
capacity
0.6
0.7
0.8
0-Step
and 1-Step
Entry
0-Step
+ 1-Step
+ 2 Step
Entry
100
100
90
90
Percentage Entry
Entry
Percentage
80
80
70
70
60
60
50
50
Capacity
Capacity
Capacity
Capacity
0-Level
0+1
Level
0+1+2
0+1 Level
1-Level
2-Level
``
40
40
30
30
20
20
10
10
00
00
10
10
20
20
30
30
40
40
50
50
6060
Percentage
PercentageCapacity
Capacity
70
7070
80
8080
90 100
100
9090
100
BOS: Canonical mixed-motive
game
FL
FH
COLUMN
ML
MH
0, 0
10, 30
30, 10
0, 0
Results 07 (male-female)
COLUMN
ML
MH (.56)
FL
0, 0
10, 30
FH (.50) 30, 10
0, 0
Matching pennies (Nash .5, .33;
CH .28, .55)
Left
0, 10
Top (.63)
ROW
Bottom (.37) 20, 0
COLUMN
(.43)
Right (.57)
10, 0
0, 10
Private information: Hidden
information & hidden action
• Hidden information (“adverse selection”)
–
–
–
–
–
–
Cannot measure pre-contract information
E.g.: Acquire-a-company problem
Betting game (Groucho Marx theorem)
Coin auction
Insurance market failure
movie “cold opening”
• How to overcome?
–
–
–
–
Measure
Exclude (insurance)
Screening or signaling
Efficient? (e.g. jockeys)
Betting game & Groucho Marx
theorem
1's payoffs
2's payoffs
STATE
A
B
C
D
+32 -28 +20 -16
-32 +28 -20 +16
1 learns (A,B) or (C,D)
2 learns (A), (B,C) or (D)
Should they bet?
• Figure ?: Professor Rafael Robb: Guilty of hubris and
murder or neither?
• A possibly ironic touch comes at the very end of the AP
story (“Penn Professor Charged in Wife’s Slaying”, jan 8,
2007):
• Penn officials said earlier that they had arranged for
someone else to teach Robb's graduate seminar in
game theory this semester.
Insurance company exclusions
•
Air traffic control
Building, moving
Chemical/rubber manufacturing
Circus or carnival work
Concrete or asphalt work
Crop dusting
Firefighting
Furniture and fixtures manufacturing
Lumber work, including wood chopping,
timber cutting and working in a sawmill
Migrant labor
Oil well or refinery work
Police work
Roofing
Sandblasting
Sports, semi-pro or professional
Stockyard work, with or without
butchering
Stables, all employees
Stunt work
Telecom installation
Transportation and aviation
Tree climbing
Tunnel work
War reporting
Window work at heights exceeding three
stories
• Lipitor (cholesterol)
Zocor (cholesterol)
Nexium
(heartburn, ulcers)
Prevacid
(heartburn, ulcers)
Advair (asthma)
Zoloft (depression)
Singulair (asthma)
Protonix
(heartburn, ulcers)
• Hidden action (“moral hazard”)
– Cannot enforce choice of post-contract action
– E.g., trust games
– Air traffic control
Moral hazard in air traffic control?
Table ? : Increases in diagnoses after rule changes in
“Second Career” program for air traffic controllers
(Staten and Umbeck, 1982)
Pre-Second
Career
Program
Change
incidence
Post-Second
Career
Program Change
incidence
% change
Respiratory
1.9
1.5
-21
Muscles
.5
.4
-20
Ear,nose,throat
6.7
8.0
+19
Abdominal
16.7
20.4
+22
Eye
5.6
7.4
+32
Bones and joints
2.3
4.6
+39
Cardiovascular
22.1
32.5
+47
Neuropsychiatric
10.9
27.2
+150
Disorder
• 2nd Career program
changes in ’74
• 100% of pay if
injury is “disabling”
+ 2 yrs job training
• Could choose own
MD or psychologist
• Increase in “system
errors” (<3mi or
1000 vertical ft)
• No increase in near
misses
How to avoid hidden info & action?
• Monitoring
– WaWa stores, undercover retail checkers
• Reputations
– Internet! (Ebay, dontdatehimgirl.com)
• 3rd party assurance (“social collateral”)
• Honor code! (Caltech)
• Make moral people
– Socialization?
– Early-childhood nutrition (Adrian Raine Mauritius
study) reduces ASPD?
PROFILE
1
.
CITY
killeen ,
VIEWS
0
displaying 1 to 5 of 18581 of alleged cheaters
Velazquez, Joel "Slim, patches,"
NEW
added: Thursday, 1:47:00 PM
2
.
Long Beach State, Texas A&M,
11
Miamisburg, USA
33
Petersburg,
23
Virginia Beach, Indianapolis,
31
Beecroft, Bart "bbuzz68"
NEW
added: Thursday, 1:27:00 PM
3
.
Ross, Jonathan "Johnny, JR, Scooby"
NEW
added: Thursday, 1:19:00 PM
4
.
Harris, Marion "Tyke"
NEW
added: Thursday, 1:06:00 PM
5
.
Green, James "MIKE" "Mike, Whitechocotate"
NEW
added: Thursday, 12:50:00 PM
Theories of human nature
• Crucial for organizational design
• Are people good & need opportunity
(Theory Y) or bad & need constraint
(theory X)? (a la Maslow hierarchy)
• Models:
– States vs traits (sorting good from bad)
– Differences in social preferences
• “self-interest seeking with guile” (opportunism) as
limiting case (or ASPD?)
– Social image & moral wriggle room
States vs traits
• Behavior due to situational “states” or personal
(immutable) “traits”?
• Attribution theory (Kelley, Nisbett-Ross):
– (Western) tendency to overestimate effect of global
traits & ability to do trait inference (e.g. interviews)
– E.g. question-answer study
– E.g. Asian vs Caucasian “brains vs work” in
educational success
– Self-serving tendency to blame state for bad
outcome, claim trait credit for good outcome (annual
reports, oil company executive pay)
Social preferences
•
•
•
•
Will sacrifice money to help/hurt others
Dictator game
Ultimatum game
One view (inequality-aversion):
– Dislike envy & guilt
– Prefer equal shares
• Another view
– Rawlsitarian (like $, minimum, total)
• Best guess?: 40% selfish, 50% conditional
cooperators, 10% “saints”
The right view of human nature?
(from Dawes and Thaler 1988)
Ultimatum vs dictator “games” (Forsythe et al 1994)
NB: Dictator games are “weak situations”, more variance
This is your brain on unfairness: Areas that are differentially
active facing unfair offers (1-2) versus fair offers (4-5)
(Sanfey et al 04 Science)
Ultimatum offer experimental sites
Ultimatum offers across societies
(mean shaded, mode is largest circle…)
A subject complaining post-experiment
(Zamir, 2000)
Social image
• People do not care directly about others
• People care about how others perceive them
(2nd-order belief)
– Few large anonymous donations
– Choose $9 or Play $10 dictator game
–  Information avoidance
•
•
•
•
•
Dana-Weber-Kuang/Feiler studies
High-risk people avoiding an AIDS test
Cross the street to avoid a homeless person
“Plausible deniability”
HP Chief Ethics Officer (?): “I shouldn’t have asked”