Coherent

arbitrariness
Stable preferences in an unstable world

Pricing?
How do we decide if we should buy a cup
of coffee (for $3)?
Ideally an internal representation of coffee
will be compared to price.
If value > price ‫ ؞‬buy
If value < price ‫ ؞‬don’t buy

How about new products?

How much is this sound worth?

How come the the sound is so different

from coffee?
Sound should be simpler…

Some intuitions

Can one map the pleasure of a

chocolate-bar to money?

Is this mapping immediate & direct?

How about chocolate vs. ice-cream?

 The questions

How do people decide about the

pricing-values of goods?
Novel hedonic stimuli

Do people learn over time?

& What

Fundamental Values
Economic theories assume underlying
“fundamental” values.
Rarely, however, is it possible to measure
fundamental values.
Virtually all tests of economic predictions examine
the effect of changes in circumstances on valuation.

Such results are necessary but not sufficient

condition for fundamental valuation.

A psychological perspective
Sensitivity to anchors
Kahneman & Tversky, 1974

Context effects
Simonson & Tversky, 1992

‫؞‬People often have an imperfect

understanding of their own values

How do people “create” value?

Based on internal evaluations?
Based on external cues?
If both when…

If people don’t have fundamental

values, how is it that observed behavior
is so consistent.
 A toy example
180,000
160,000
140,000
120,000
Variance

100,000
80,000
60,000
40,000
20,000
0
se

s
ue

og
es

s
ng

er
oe
ou
ho
Bl

D
ap
So

Sh
M
lS

k
ce

sp

al
ng

e
el

ll

ew
Fa

in

W
Ki

Si
Sm

Sh

Experiment 1 Procedure

Do people have fundamental values?

2 (Anchor) by 2 (order) by 3 (duration) by 3
(replications)

Experiment 1 Procedure
Introduction
Subjects listen to sound

Hypothetical question
No anchor
10¢
50¢

Bidding for real X 9

Getting real payoffs
Increasing 10 sec, 30 sec, 60 sec
Decreasing 60 sec, 30 sec, 10 sec

Experiment 1 Results I
High-(50c) Low-(10c) No-Q
6
5
4
Ratios

3
2
1
0
Ratio 60/10 Ratio 30/10 Ratio 60/30
 Experiment 1 Results II
WTH over time

No-ExPrice Low-ExPrice High-ExPrice

100
90
80
70
60
50
40
30
20
10
0
T1 T2 T3 T4 T5 T6 T7 T8 T9

Experiment 1 Conclusions
Subjects do not have internal value for
novel hedonic stimuli
Once a response is made, other
responses follow

Coherent arbitrariness

Real products
Mean value Low ss# High ss# Increase
Trackball 16.25 10.38 21.52 107%
Keyboard 32.47 21.81 42.03 93%
$9 wine 15.80 11.62 19.55 68%
$82 wine 22.89 17.42 27.76 59%
Design book 18.81 14.15 23.00 62%
Belgian Chocolates 13.31 10.04 16.24 62%

High dependency between the prices of the

2 wines and 2 computer accessories

Tom Sawyer

 Poetry I
Poetry, charge Poetry, pay

10
9
8
7
6
5
4
3
2
1
0
1 min 5 min 10 min

Poetry II

Duration
-

Decision making exp

DM study, charge DM study, pay

14
12
10
8
6
4
2
0
1 min 5 min 10 min

“The model”
Self herding

Evaluating self preferences is hard

Behavior is a signal for preferences
Frequent behavior is taken as a strong
signal
 Experiment 2 Procedure

Will the forces of the market correct

arbitrariness?
2 (Anchor) by 2 (order) by 3 (duration) by 3
(replications)
Anchors were 10¢ and 100¢
Groups (Markets), not individuals

Experiment 2 Procedure
Introduction
Subjects listen to sound

Hypothetical question
No anchor
10¢
100¢

Bidding for real X 9

Getting real payoffs
Increasing 10 sec, 30 sec, 60 sec
Decreasing 60 sec, 30 sec, 10 sec

Experiment 2 Results I
Bids and wins over time
Bid Low-ExPrice Bid High-ExPrice Win Low-ExPrice Win High-ExPrice

200
180
160
140
120
100
80
60
40
20
0
T1 T2 T3 T4 T5 T6 T7 T8 T9

Experiment 2 Results II
Standard Deviation of WTA

250
200
150
100 B B
B B
50 B
B
B
B B

0
T1 T2 T3 T4 T5 T6 T7 T8 T9
Trial Number
Experiment 2 Conclusions
Arbitrary values are not “corrected”
over time

Arbitrary values are not “corrected” in

marketplaces

Learning to arbitrary values?

Experiment 3
Known random anchors
Use subjects’ own SS#

Larger magnitude
100 sec, 300 sec, 600 sec

Same basic procedure as in Exp 1

Plus ranking of small annoying tasks

Experiment 3 results I
Low Anchor High Anchor

10
9
8
WTA ($)

7
6
5
4
3
2
1
0
100 sec. 300 sec. 600 sec.

Experiment 3 results II
Increasing Decreasing

10
9
8
600 sec.
WTA ($)

7
6 600 sec.
300 sec.
5
4 300 sec.
3 100 sec.
2
100 sec.
1
0
First Sound Second Sound Third Sound
 Experiment 4
What is the role of the first anchor?
Information?
Changes utility or mapping?

3 (sounds) by 2 (order) by 3 (Anchor)

Anchors were 10¢, 50¢ and 90¢
Duration was constant at 30 sec.
Type of sound changed
Fourth trail was the same as trial#1 + vise

Experiment 4 Procedure I
Introduction

Subjects listen to sound

Bidding for real

Getting real payoffs

Hypothetical question

10¢ / 50¢ / 90¢

90¢ / 50¢ / 10¢ Vise

Why the vise?

{
{
=
>

$ Vise Vise
=
>

Experiment 4 Results WTA

WTA
WTA-1 WTA-2 WTA-3
100
90
80
70
60
50
40
30
20
10
0
Increasing Decreasing
 Experiment 4 Results II

Significant effect of order

No effect for anchors

No effect for Vise/Sound choice

Experiment 4 Conclusions
The first trial had a large role

The first response determines future

responses

Not an inference process

Using the anchor as information

Utility theory interpretation

Traditional view on tradeoffs

100
90 D>$
80
D=$
70 D<$
60
50
40
30
20
10
0
10 30 60

Duration

Utility theory interpretation

“Incomplete preference” view on
tradeoffs
100
90
80
70
60
50
40
30
20
10
0
10 30 60

Duration
 Utility theory interpretation
Initial response and its consequences
100
90
80
70
60
50
40
30
20
100 10
90 0
80 10 30 60
70
60 Duration
50
40
30 100
20 90
10 80
70
0 60
50
10 30 60 40
30
Duration 20
10
0
10 30 60
Duration

Utilities or mapping?
Leftovers
Experiment 3 (large payoffs & Random anchor)
Rank order of small annoying tasks
Blood test, missing a bus, dropping your
ice-cream etc.
Experiment 4 (3 different anchors)
Choice of sound vs vise

No effect for anchoring!

Utilities vs. mapping

experiment
Anchoring 4 stimuli
A 1 sec shock from a 9V battery
A bad drink (OJ & vinegar)
Drinking from a new bedpan
Eating a spoon of butter

4 Values ($0.5, $1, $2, $4)

The procedure
Anchoring
Pair-wise preference
Y/N Question
OR Pair-wise preference

WTA
No measurement

Time
 Results Anchoring

Bedpan: [R2 = 0.46, F = 17.96, p < 0.01]

Butter: [R2 = 0.28, F = 5.8, p < 0.02]
Shock: [R2 = 0.34, F = 8.8, p < 0.01]
Drink: [R2 = 0.29, F = 6.23, p < 0.02]

Results effects on pair-wise

Are the pair-wise preferences influenced by
the ratios of the pair-wise anchors?
Shock/bedpan (R2 = 0.046)

Bedpan/butter (R2 = 0.023)

Bedpan/drink (R2 = 0.002)

Butter/shock (R2 = 0.002)

Drink/Shock (R2 = 0.011)

Mapping or utilities --
summary
So far evidence for mapping

Questions
What makes mapping difficult?
Money? (sound vs. drink ‫ ؞‬no)
Abstract attributes?
Distance?

Conclusions
The coherence of the market seem to
reflect the psychology of relative
valuations
People do not seem to have
fundamental values even for simple
experiences
Is money a bad idea?
 Possible Applications /
economics I
Financial markets (Shiller, 1987)
“Who would know what the value of the Dow Jones Industrial
Average should be? Is it really "worth" 6,000 today? Or 5,000
or 7,000? or 2,000 or 10,000? There is no agreed-upon
economic theory that would answer these questions. In the
absence of any better information, past prices (or asking prices
or prices of similar objects or other simple comparisons) are
likely to be important determinants of prices today.”
- From: Shiller, Robert J. "Investor Behavior in the October 1987 Stock Market
Crash: Survey Evidence." Cambridge, MA: National Bureau of Economic Research,
1987.

Market reactions are sensitive to performance relative to

expectations (IBM is doing X better than expected) and to other
relative changes (IBM is buying back X stocks)

Possible Applications /
economics II
Labor markets Bewley (1998).
“Non-union companies seemed to be isolated
islands, with most workers having little systematic
knowledge of pay rates at other firms. Pay rates in
different non-union companies were loosely linked
by the forces of supply and demand, but these
allowed a good deal of latitude in setting pay”
- From: Bewley, Truman. "Why Not Cut Pay?" European Economic Review 42, 1998,
pp. 459-490.

Well being within a company is related to relative

pay to others and to former levels

Possible Applications /
economics III
contingent valuation
“valuations of any particular quantity [of a good]
would be sensitive to its relative position within
the range selected for valuation, but insensitive to
which range is chosen, resulting in insensitive (or
incoherent) values across studies using different
quantity ranges” Frederick and Fischhoff (1998,
p. 116)
- From: Frederick, S., and B. Fischhoff. "Scope Insensitivity in Elicited Values." Risk Decision
and Policy 3, 1998, pp. 109-124.

BUT, within subjects evaluations are coherent

Possible Applications /
economics IV
Criminal deterrence (Ross, 1973)
People seem sensitive to policy changes in
deterrence (perhaps only short term)

People seem insensitive to absolute levels of

probability of punishment

Reactions to crime also have a very strong

cultural effects
- Based on: Ross, H. L. "Law, Science and Accidents: The British Road Safety Act of 1967."
Journal of Legal Studies 2, 1973, pp. 1-78.
Marketing implications
First price has long lasting implications

Products are comparative in nature

Understanding what a product will be
compared to is important

Product lines and product extensions