Scribd
Upload
176 views8 pages

Arrow Pratt Theorem

This document discusses different measures of risk aversion. It defines local absolute risk aversion as -u''(x)/u'(x), where a more negative value indicates greater risk aversion. Global risk aversion compares two utility functions u1 and u2, and says u2 exhibits greater risk aversion if it is a concave transform of u1. Pratt's Theorem shows these three measures - local absolute risk aversion, global risk aversion, and risk premium - are equivalent. The document also discusses constant absolute risk aversion and relative risk aversion.

Uploaded by

Lalit Sapkale
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
176 views8 pages

Arrow Pratt Theorem

This document discusses different measures of risk aversion. It defines local absolute risk aversion as -u''(x)/u'(x), where a more negative value indicates greater risk aversion. Global risk aversion compares two utility functions u1 and u2, and says u2 exhibits greater risk aversion if it is a concave transform of u1. Pratt's Theorem shows these three measures - local absolute risk aversion, global risk aversion, and risk premium - are equivalent. The document also discusses constant absolute risk aversion and relative risk aversion.

Uploaded by

Lalit Sapkale
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/ 8

Measuring Risk Aversion

Local Risk Aversion

De…nition: Given a twice-di¤erentiable Bernoulli util-


ity function u(¢); the Arrow-Pratt measure of absolute
risk aversion at x is de…ned as:

¡u00 (x)
rA = u0 (x)
.

For two individuals, 1 and 2, with twice-di¤erentiable,


concave, utility functions u1(¢) and u2(¢), respectively,
person 2 is more risk averse than person 1 at the level
of income x i¤

¡u001 (x) ¡u002 (x)


r1A = u01(x)
< u02 (x)
= rA2

This measure allows us to compare attitudes towards


risky situations whose outcomes are absolute gains or
losses from current wealth x.
NOTE: this is only a local measure of risk aversion. For
a di¤erent level of x; the comparison could switch.

12
Local Risk Aversion (Cont.)

Why not just use u00(x)? Two reasons:

1) No equivalence with other reasonable measures.

2) The AP measure is invariant to a¢ne transforma-


tions of the utility function, but u00 is not.

Consider º and suppose U represents º.


The associated Bernoulli utility function is u(¢).

Take the family of utility functions À(x) = ¯u(x)+°:


All these represent the same preferences.

We have À0(x) = ¯u0(x) and À 00(x) = ¯u00(x).


00
The AP is then ¡u (x)
u (x) , and is therefore the same for
0
any
function in this family.

13
Global Risk Aversion

Given two twice-di¤erentiable Bernoulli utility func-


tions u1(¢) and u2(¢); individual 2 is globally more risk
averse than individual 1 if and only if there exists a
concave function ª(¢) such that

u2(x) = ª(u1(x)):

That is, u2(¢) is a concave transform of u1(¢).

14
Risk Premium and Certainty Equivalent

Consider two individuals with utility functions u1(¢)


and u2(¢): Individual 2 is more risk averse than indi-
vidual 1 if and only if:

c(F; u2) · c(F; u1) for every lottery F (¢):

Since ½ = EV ¡ CE, equivalently individual 2 is more


risk averse than individual 1 when 2’s risk premium is
higher:

½(F; u2) ¸ ½(F; u1) for every F (¢):

15
Pratt’s Theorem:

The three previous measures of risk aversion are all


equivalent, given twice-di¤erentiable utility functions.

That is, the property that a utility function u2 is a


concave transform of a utility function u1 is equivalent
to 2 having a greater A-P measure of absolute risk
aversion for all x:

rA(x; u2) ¸ rA(x; u1) for all x 2 [a; b];

and equivalent to 2 having a greater risk premium:

½(F; u2 ) ¸ ½(F; u1) for any F (¢):

16
Risk Aversion and Wealth

De…nition: The Bernoulli utility function u(¢) a ex-


hibits decreasing (constant) (increasing) absolute risk
aversion if rA (x; u) is a decreasing (constant) (increas-
ing) function of x.

e.g. consider two di¤erent wealth levels w1 > w2.


The set of possible outcomes involves a monetary
payment x.
A person’s utility function u exhibits decreasing
absolute risk aversion (DARA) i¤

rA (w1 + x; u) < rA(w2 + x; u):

17
Exponential Utility and CARA

Economists often model utility by an exponential func-


tion

u(x) = ¡e¡¸x where ¸ > 0:

This utility function exhibits constant absolute risk aver-


sion (CARA).

u0(x) = ¸e¡¸x u00(x) = ¡¸2e¡¸x ,

rA = ¡u00=u0 = ¸.

18
Relative Risk Aversion

De…nition: Given a twice-di¤erentiable Bernoulli util-


ity function u(¢); the coe¢cient of relative risk aversion
at x is de…ned as:

¡u00 (x)
rR = x u0(x) = xrA.

We can write it as follows:

¡u00 (x) du0 (x) x du0 (x) dx


rR = x u0(x) = dx
¢ 0
u (x)
= =
u0 (x) x
= %¢u0=%¢x.

If rR is decreasing with respect to x we say that an


individual exhibits decreasing relative risk aversion.
She becomes less risk averse with regard to gambles
that are proportional to her wealth as her wealth in-
creases.

19

You might also like