Scribd
Upload
25 views21 pages

Lec2 pt2

1. This document discusses prospect theory and some open questions it presents. It summarizes an experiment showing people value losing an object more than gaining the same object (the endowment effect), explained by prospect theory but not expected utility theory. 2. It outlines several open questions with prospect theory, including whether narrow framing of individual gambles is used or integrated framing, the effect of time horizons, how to model risk preferences and reference points, dynamic inconsistency over time, and difficulties doing welfare analysis due to framing effects. 3. It previews discussing the costs of business cycles using prospect theory in the next lecture and assigning a problem set to address one of the open questions. Readings on

Uploaded by

toabhishekpal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views21 pages

Lec2 pt2

1. This document discusses prospect theory and some open questions it presents. It summarizes an experiment showing people value losing an object more than gaining the same object (the endowment effect), explained by prospect theory but not expected utility theory. 2. It outlines several open questions with prospect theory, including whether narrow framing of individual gambles is used or integrated framing, the effect of time horizons, how to model risk preferences and reference points, dynamic inconsistency over time, and difficulties doing welfare analysis due to framing effects. 3. It previews discussing the costs of business cycles using prospect theory in the next lecture and assigning a problem set to address one of the open questions. Readings on

Uploaded by

toabhishekpal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

14.

127 Behavioral Economics


(Lecture 2)

Xavier Gabaix

February 12, 2004

0.1

Cumulative PT

Remind from last lecture: for continuous gambles with distribution


f (x)
EU gives:
V =
PT gives:
V =
+

Z +

0
Z 0

Z +

u (x) f (x) dx,

u (x) f (x) 0 (P (g x)) dx

u (x) f (x) 0 (P (g x)) dx

Alternatively, we can write it as Riemann-Stieltjes integral


V =
+

Z +

Z 0

u (x) d (1 P (g < x))

u (x) d (P (g x))

This simplifies to PT for two outcome gambles. Indeed, it is selfevident in the Riemann-Stieltjes form.

The endowment eect a consequence of PT


Lab experiment, Kahneman, Knetsch, Thaler, JPE 1990.
Half of the subjects receives an MIT apple, and the other half
receives $10.
Then willingness to pay WTP for the apple is elicited from subjects
with money, and willingness to accept WTA is elicited from subjects
with mugs.

In EU we have W T P = W T A (modulo wealth eects, which are


small)

In simplified (linear) PT value getting an apple and lose $x is


V = u (apple) + u (x) = A x
(notethere are mental accounting ideas plugged in here that is we
process apple and money on separate accounts).
Thus, in PT, one accepts when
A x 0
so that
WTA =

A
.

In simplified (linear) PT value losing an apple and gaining $x is


V = u (apple) + u (x) = A + x

(note, once more time we process apple and money on separate accounts).

Thus, in PT, one pays when


A + x 0
so that
W T P = A.

Thus, PT gives stability to humane life, a status quo bias.

1.1

Endowment eect experiment with mugs

Classroom of one hundred. Fifty get the mug, fifty get $20.
One does a call auction in which people can trade mugs.
Trading volume rational expectation would be that the average
trading volume should be 12 50 = 25. Everybody has a valuation, and
probability that someone with valuation higher than the market price
is 12 .
If WTP<WTA then the trading volume is lower than 12 .
In experiments, the trading volume is about 14 .

1.2

Open questions with PT

1.2.1

Open question 1: Narrow framing

N independent gambles: g1, ..., gN


For each i do you accept gi or not?
In EU call ai = 1 if accept gi and ai = 0 otherwise. Your total wealth
is
W0 + a1g1 + ... + aN gN
and you maximize
max Eu (W0 + a1g1 + ... + aN gN ) .

a1,...,aN

In PT we have at least two possibilities


Separation: ai = 1 i V P T (gi).
Integrative: solve maxa1,...,aN V P T (a1g1 + ... + aN gN ).
Separation is more popular, but unlikely in for example in stock market,
or venture capital work.

KT dont tell whether integration or separation will be chosen. That


is one of the reasons PT has not been used much in micro or macro.

How to fix this problem?

Integration as far as possible subject to computational costs.


Natural horizon between now and when I need to retire.
Do what makes me happier, max (separation,integration) . That
would be an appealing general way to solve the problem.

Problem, each everyday gamble is small against the background


of all other gambles of life.
So, an EU maximizer would be locally risk neutral.
And also a PT maximizer would be locally risk neutral whenever
he or she accpets integrationist frame.

1.2.2

Horizon problem a particular case of the framing problem

Stock market.
Yearly values
standard deviation

1
T 2

= 20% per year where T ' 250days,

mean T = 6% per year.


Daily values
=

20%
1
250 2

6%
=
T

premium
Assume that a PT agent follows the rule: accept if Risk
>
St. dev.
k (PS1 asks to show existence of such an PT agent).

So, a PT agent with yearly horizon invests if


6%
> k
20%
A PT agent with daily horizon invests if

.024
=
' .01 << k

1.3
This is not even a debated issue, because people dont even know
how to start that discussion
Kahneman says in his Nobel lecture that people use accessible
horizons.

E.g. in stock market 1 year is very accessible, because mutual


funds and others use it in their prospectuses.
Other alternatives time to retirement or time to a big purchase.
or TV every day.
In practice, for example Barberis, Huang, and Santos QJE 2001
postulate an exogenous horizon.

1.2.3

Open question 2: Risk seeking

Take stock market with return R = + n with n N (0, 1) .


Invest proportion in stock and 1 in a riskless bond with return 0.
Total return is
R + (1 ) 0 = ( + n) .
Lets use PT with (p) = p. The PT value is
V =

Z +

u ( ( + n)) f (n) dn

Set u (x) = x for positive x and |x| for negative x.


Using homothecity of u we get
V =
=

Z +

|| u ( + n) f (n) dn

Z +
u ( + n) f
||

(n) dn

Thus optimal to equals 0 or + depending on sign of the last


integral.

Why this problem? It comes because we dont have concave objective function. Without concavity it is easy to have those bang-bang
solutions.

One solution to this problem is that people maximize V EU + V P T .

1.2.4

Open question 3: Reference point

Implicitly we take the reference point to be wealth at t = 0. Gamble


is W0 + g and
V P T = V P T (W0 + g R)
But how Rt evolves in time?
In practice, Barberis, Huang, and Santos QJE 2001 (the most courageous paper) postulate some ad hoc exogenous process. People gave
them the benefit of a doubt.

1.2.5

Open question 4: Dynamic inconsistency

Take a stock over a year horizon. Invest 70% on Jan 1st, 2001.
Its Dec 1, 2001. Should I stay invested?
If the new horizon is now one month, I may prefer to disinvest, even
though on Jan 1, 2001, I wanted to keep for the entire year.

By backward induction, Jan 1 guy should disinvest!

1.2.6

Open question 5: Doing welfare is hard

Why? Because it depends on the frame.

Take T = 250 days of stock returns gi N , 2 . Integrated


P

V PT (

gi) = V I and separated V P T = V S .

The cost of the business cycle (Lucas). Suppose c =average monthly


consumption. Assume simple consumption shocks:
ct = c + t
with normal iid t.
What is PT reference point? Take Rt = c = 0.

With PT integrated over one year


V PT

t = V P T

1
12 2 n

1
12 2 V P T

With segregated PT
V P T = 12V P T (n)
Which frame is better?

(n) < 0.

1.3

Next time

Lucas calculation of costs of business cycle. In practice people care


about business cycles, and election are decided on those counts.

Problem Set next time. One question try to circumvent one of


the problems.

Readings on heuristics and biases, the Science 74 KT article and


Camerers paper from the syllabus.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy