Intro Heterogeneous Agents Model

Models with Heterogeneous Agents
Introduction
Wouter J. Den Haan
London School of Economics
c by Wouter J. Den Haan
August 10, 2015
Overview
Simple example & key issues
Aiyagari model
Does it matter?
Avoiding complexity
Overview
"Simple" model with heterogeneous agents

understanding the complexity of these models
role of aggregate uncertainty
role of incomplete markets
Solving the Aiyagari model

basic numerical techniques (refresher)
Does heterogeneity matter?
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Overview continued
Avoiding complexity
heterogeneity only within the period
partial equilibrium
are two agents enough?
Other models with heterogeneity

New Keynesian model
Multiplicity & domino eects due to tax externality
macro model with search frictions
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
First model with heterogeneous agents
Agents are ex ante the same,

but face dierent idiosyncratic shocks
=) agents are dierent ex post
Incomplete markets
=) heterogeneity cannot be insured away
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Individual agent
Subject to employment shocks:

i,t 2 f0, 1g
Incomplete markets
only way to save is through holding capital
borrowing constraint ki,t+1 0
Other models
Overview
Aiyagari model
Does it matter?
Aggregate shock
zt 2 fzb , zg g
zt aects
1
2
aggregate productivity
probability of being employed
Avoiding complexity
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Laws of motion
zt can take on two values
i,t can take on two values
probability of being (un)employed depends on zt
transition probabilities are such that
unemployment rate only depends on current zt
thus
ut = ub if zt = zb &
ut = ug if zt = zg
with ub > ug .
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Firm problem
1 1
Lt
) zt Kt L1t
rt = zt Kt
wt = ( 1
These are identical to those of the rep. agent version
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Government
t wtl(1
u(zt ) = wt u(zt )
u(zt )
t =
l(1 u(zt ))
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Individual agent
max
fci,t ,ki,t+1 gt=0
E t=0 t ln(ci,t )
s.t.
ci,t + ki,t+1 = rt ki,t + (1
t )wt li,t + wt (1
ki,t+1 0
this is a relatively simple problem

if processes for rt and wt are given
i,t ) + (1
)ki,t
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Individual agent - foc
1
ci,t
Et
0 = ki,t+1
1
ci,t+1
1
ci,t
ci,t + ki,t+1 = rt ki,t + (1

ki,t+1
(rt+1 + 1
Et
)
1
ci,t+1
( rt + 1 + 1
t )wt li,t + wt (1
i,t ) + (1
)ki,t
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
What aggregate info do agents care about?
current and future values of rt and wt

the period-t values of rt and wt
only depend on aggregate capital stock, Kt , & zt
!!! In most models, prices also depend on other characteristics
of the distribution
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
What aggregate info do agents care about?
the future values, i.e., rt+ and wt+ with > 0 depend on
future values of mean capital stock, i.e. Kt+ , & zt+
=) agents are interested in all information that forecasts Kt

=) typically this includes the complete cross-sectional
distribution of employment status and capital levels
(even when you only forecast futures means like you do here)
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Equilibrium - rst part
Individual policy functions that solve agents max problem

A wage and a rental rate given by equations above.
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Equilibrium - second part

A transition law for the cross-sectional distribution of capital,
that is consistent with the investment policy function.
ft+1 = (zt+1 , zt , ft )
ft = cross-sectional distribution of beginning-of-period capital
and the employment status after the employment status has
been realized.
zt+1 does not aect the cross-sectional distribution of capital
zt+1 does aect the joint cross-sectional distribution of capital
and employment status
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Transition law & timing
ft & zt =) ftend-of-period
ftend-of-period & zt+1 =) ftbeginning-of-period
+1
ft+1
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Transition law & timing
Let gt be the cross-sectional distribution of capital

(so without any info on employment status)
Why can I write
gt+1 = g (zt , ft )?
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Transition law & continuum of agents
gt+1 = g (zt , ft )
ft+1 = (zt+1 , zt , ft )
Why are these exact equations without additional noise?
continuum of agents =) rely on law of large numbers to
average out idiosyncratic risk
are we allowed to do this?
Overview
Aiyagari model
Does it matter?
Recursive equilibrium?
Questions
1
Does an equilibrium exist?

1
If yes, is it unique?
Does a recursive equilibrium exist?

1
2
If yes, is it unique?
If yes, what are the state variables?
Avoiding complexity
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Recursive equilibrium?
Jianjun Miao (JET, 2006): a recursive equilibrium exist for following
state variables:
usual set of state variables, namely
individual shock, i,t
individual capital holdings, ki,t
aggregate productivity, zt
joint distribution of income and capital holdings, ft
and cross-sectional distribution of expected payos
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Unique?
Heterogeneity =) more reasons to expect multiplicity

my actions depend on what I think others will do
heterogeneity tends to go together with frictions
and multiplicity more likely with frictions
e.g. market externalities
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Wealth-recursive (WR) equilibrium

WR equilibrium is a recursive equilibrium with only i,t , ki,t , zt ,
and ft as state variables.
(Also referred to as Krusell-Smith (KS) recursive equilibrium)
Not proven that WR equilbrium exists in model discussed here
(at least not without making unveriable assumptions such as
equilibrium is unique for all possible initial conditions)
Kubler & Schmedders (2002) give examples of equilibria that
are not recursive in wealth
(i.e., wealth distribution by itself is not su cient)
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Wealth distribution not su cient - Example
Static economy
two agents, i = 1, 2, two commodities, j = A, B
Utility:
ln qA + ln qB
Endowments in state I: 1,A = 2,A = 1; 1,B = 2,B = 1
Endowments in state II:
1,A = 2,A = 1; 1,B = 2,B = 10/9
Normalization: pA = 1
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Wealth distribution not su cient - Example
State I:
equilibrium: pB = 1; q1,A = q2,A = 1; q1,B = q2,B = 1
wealth of each agent: = 2
State II:
equilibrium: pB = 0.9; q1,A = q2,A = 1; q1,B = q2,B = 10/9
wealth of each agent: = 2
Thus: same wealth levels, but dierent outcome
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
How to proceed?
Wealth distribution may not be su cient!

For numerical analysis less problematic: It typically leaves stu
out
After obtaining solution, you should check whether the
approximation is accurate or not
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
How to proceed?
For now we assume that a wealth recursive equilibrium exists

(or an approximation based on it is accurate)
This is still a tough numerical problem
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
If a wealth recursive equilibrium exists
Suppose that recursive RE for usual state space exists

si,t = fi,t , ki,t , st g = fi,t , ki,t , zt , ft g
Equilibrium:
c(si,t )
k(si,t )
r(st )
w ( st )
(zt+1 , zt , ft )
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Alternative representation state space
Suppose that recursive RE for usual state space exist

si,t = fi,t , ki,t , st g = fi,t , ki,t , zt , ft g
What determines current shape ft ?

zt , zt
zt , zt
zt , zt
zt , zt
1 , ft 1
or
1 , zt 2 , ft
1 , zt 2 , zt
1 , zt 2 , zt
or
3 , ft 3
3 , zt
or
,
4 ft
or
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
No aggregate uncertainty
st = lim fzt , zt
n !
1,
, zt
n , ft n g
Why is this useful from a numerical point of view?

when zt is stochastic
when zt is not stochastic (case of no aggregate uncertainty)
Overview
Aiyagari model
Does it matter?
Avoiding complexity
No aggregate uncertainty
State variables
lim fzt , zt
n !
1,
, zt
n , ft n g
If
1
2
zt = z 8t and
eect of initial distribution dies out
then st constant
distribution still matters!
but it is no longer a time-varying argument
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Aggregation
Statement:
The representative agent model is silly,
because there is no trade in this model,
while there is lots of trade in nancial assets in reality
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Aggregation
Statement:
The representative agent model is silly,
because there is no trade in this model,
while there is lots of trade in nancial assets in reality
Problem with statement:
RA is justied by complete markets
which relies on lots of trade
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Complete markets & exact aggregation
economy with ex ante identical agents

J dierent states
complete markets =) J contingent claims
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Complete markets & exact aggregation
max
ci,t ,b1i,t+1 ,
,bi,t+1
h
(ci,t )1
+ Et v(b1i,t+1 ,
1
J
J
, bi,t+1 )
s.t. ci,t + qj bi,t+1 = yi,t + I (j )bi,t

j=1
j=1
bi,t+1 > b with b < 0
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Euler equations individual
qj (ci,t )
= ci,t+1
prob(j)
8j
This can be written as follows:

ci,t =
prob(j)
qj
1/
ci,t+1
8j
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Aggregation
Aggregation across individual i of
prob(j)
qj
ci,t =
1/
8j
ci,+1
gives
Ct =
1/
prob(j)
qj
Ct + 1
8j,
which can be rewritten as

qj (Ct )
= Ct + 1
prob(j)
8j
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Use equilibrium condition
In equilibrium:
aggregate consumption equals aggregate income or
contingent claims are in zero net supply
Thus
qj (Yt )
= Yt + 1
prob(j)
8j
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Back to represenative agent model

Idential FOCs come out of this RA model:
max
C,B1+1 ,
,B+1
h
( Ct ) 1
+ Et v(B1t+1 ,
1
J
s.t.Ct +
j=1
j
j
qj Bt+1
, Bt + 1 )
= Yt + I (j )Bt
j=1
Bt+1 > b with b < 0
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Back to model with heterogeneous agents
(For now) no aggregate risk

Aiyagari model
We simplify the standard setup as follows:
Replace borrowing constraint by penalty function
=) going short is possible but costly
workers have productivity insteady of unemployment shocks
i,t with E[i,t ] = 1
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Individual agent
max
fci,t ,ki,t+1 gt=0
E t=0 t ln(ci,t )
ci,t + ki,t = rt ki,t
1
0
exp( 0 ki,t )
s.t.
1 + wt i,t + (1
)ki,t
2 ki,t
1
First-order condition
1
+ 1 exp( 0 ki,t )
ci,t
2 + Et
ci,t+1
( rt + 1 + 1
) = 0
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Penalty function
advantage of 2 term:
supppose k and r are steady states of rep agent model
if
2 = 1 exp( 0 k )
then steady state of this model is same
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Equilibrium
Unit mass of workers, Lt = 1
Competitive rm =) agent faces competitive prices
wt = (1 ) Kt L1t = (1
rt = Kt 1 Lt = Kt 1
) Kt
No aggregate risk so
Kt = K
How to nd the equilibrium K?
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Algorithm
Guess a value for r
This implies values for Kdemand and w
Solve the individual problem with these values for r & w
Simulate economy & calculate the supply of capital, Ksupply
If Ksupply < Kdemand then r too low so raise r, say
rnew = r + (Kdemand
Iterate until convergence
Ksupply )
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Algorithm
Using
rnew = r + (Kdemand
Ksupply )
to solve
Kdemand (r) = Ksupply (r)
not very e cient
Value of may have to be very low
More e cient to use equation solver to solve fro r
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Use Dynare to solve indiv. policy rule
Specify guess for r in mother Matlab le

Make r parameter in *.mod le
In mother Matlab le write r using
save r_file r
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Use Dynare to solve indiv. policy rule
In *.mod le use
load r_file
set_param_value(r,r)
instead of
r = 0.013;
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Simulate yourself using Dynare solution
1
2
Use values stored by Dynare or

Replace Dynares disp_dr.m with my alternative
this saves the policy functions exactly as shown on the screen
asa matrix
in a Matlab data le dynarerocks.mat
under the name decision
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Does heterogeneity matter?
Important to distinguish between

(i) theoretical results
(ii) their quantitative importance
Examples
no aggregation in presence of incomplete markets
Arrows impossibility theorem
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Does incompleteness/heterogeneity
matter?
Take model with
innitely-lived agents
no complete markets
e.g. agents can only borrow/lend through a safe asset
=) no aggregation to RA model possible

But in many models eects small
why does innitely-lived agent assumption matter?
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Does incompleteness/heterogeneity
matter?
Eects often small for
asset prices
aggregate series
except possibly some impact on means
Eects much bigger for

individual series, e.g. VAR(ci,t ) >> VAR(Ct )
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
heterogeneity only within the period

partial equilibrium
two agents?
Avoiding complexity
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Avoiding complexity
Lesson learned above:

incomplete asset markets dont do much in many environments
This implies you should

either use more interesting environment
or use complete asset markets
This does NOT imply you should eliminate heterogeneity from

your models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Only heterogeneity within period
Household with heterogeneous members within the period:

members are on their own and face frictions. E.g.
cannot transfer funds to each other
cannot transfer information
At the end of period:

all members bring this periods revenues to household who
makes savings decision
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Partial equilibrium
Which of the following two would you prefer?

General equilibrium asset pricing model that generates
unrealistic asset prices
Partial equilibrium model that uses realistic asset prices as
exogenous processes
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Partial of general equilibrium?
What about follwing example

Government sets interest rates
!!! Government cannot set current rt nor rt+1 .
Suppose it sets Et [rt+1 ]. E.g.,
Et [ rt + 1 ] = ( 1
r )r + r rt + r,t
Government supplies capital to implement this.
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Partial of general equilibrium?
These expenditures are nanced by lump sum taxes.

State variables are
ki,t
i,t
zt
Kt but no higher-order moments
Et [rt+1 ] or ???
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Small number of agents
Consider following endowment economy

Type 1 agent receives z1,t
Type 2 agent receives z2,t
average endowment zt
zt = 0.5z1,t + 0.5z2,t
agents smooth idiosyncratic risk by trading in safe bonds
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Small number of agents

h
i
qt Et ci,t+1
ci,t
bi,t+1
ci,t
bi,t+1
i
h
qt Et ci,t+1 = 0
b
b1,t+1 + b2,t+1 = 0
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Idiosyncratic risk
You want to study eect of idiosyncratic risk.

Suppose agent 1 repeatedly gets the bad shock
Dierence with model with lots of types?
here: lots of agents always get same shock at same time
so what?
types.
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
To document the quantitative importance of the interest rate e!ect, I plot in

Fig. 3 the impulse response function of the interest rate when the same agent
receives the low-income realization for several periods. The graph plots the
interest rate for three levels of the borrowing constraint parameter and a
Idiosyncratic risk and interest rate
Fig. 3. The response of the interest rate to idiosyncratic shocks. This graph plots the realization of
the interest rate in the economy with two types when the agent receives the low-income realization
for several periods. The straight line drawn with the same style indicates the interest rate in the
corresponding economy with a continuum of types. The parameter of risk aversion is equal to three
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Heterogeneity in other models
Standard New Keynesian model

Simple static model with tax externality
Standard model with search friction
multiple steady states
multiple solutions
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
New Keynesian models
Calvo devil =) heterogeneous price dispersion

Standard approach:
only focus on aggregates
focus on linearized solution
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Disaggregate results in NK models
Suppose
all rms start with same price (for simplicity)
consider monetary tightening
Aggregate:
downturn because of sticky prices
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Disaggregate results in NK models
Firms that are not constrained by Calvo devil: pi #
their aggregate demand " because pi /P #

their aggregate demand # because aggregate demand #
total eect can easily be "
But empirical evidence suggests decline across dierent sectors
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Asymmetry in New Keynesian models
Suppose commodities are perfect substitutes

Monetary tightening:
rms that are not constrained by Calvo devil: pi #
=) rms constrained by Calvo devil sell 0
=) same outcome as fully exible case
=) Y = 0
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Asymmetry in New Keynesian models
Suppose commodities are perfect substitutes

Monetary stimulus:
Firms that are constrained by Calvo devil: pi = 0
=) rms not constrained by Calvo devil: pi = 0
=) same outcome as xed P case
=) Y < 0
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
The true New Keynesian models
Conclusion:
True New Keynesian models are much more interesting than the
linearized version the profession is obsessed with
Is the true NK model also more realistic?
Overview
Aiyagari model
Tax externality
Static model
N dierent skill levels
zk , k = 1,
,N
z1 = z
zk+1 = zk +
unemployed get benets
Does it matter?
Avoiding complexity
Other models
Overview
Tax externality
animated picture
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Search model
Consider the following model

unit mass of workers
workers need to search to nd a job
employers post vacancy to nd worker
productivity of matched pairs distributed i.i.d
so each period a new draw
Other models
Overview
Aiyagari model
Does it matter?
Key decision
Given value of i,t is it better to

1
2
produce or
quit and enjoy leisure?
Avoiding complexity
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Equation for cut-o value

The cut-o value i given by
0 = i + G
G : continuation value of ending period in match

does not depend on i,t (i.i.d. assumption)
does depend on i
W : continuation value of ending period not in match

also depends on i
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Solution for cut-o value

We are looking for a solution to
0 = i + G ( i )
W ( i )
Unique solution if
( i + G ( i ) b
i
W ( i ))
>0
8 i
but typically we have

(G ( i ) W ( i ))
< 0 for some i
i
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Unique steady state

2
1.5
1
0.5
0
-0.5
1
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Multiple steady state case
2
1.5
1
0.5
0
-0.5
1
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Reasons for multiplicity
expectations about the stability of future matches

as in example above
market activity could aect revenues

as in static example with tax externality
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Tax externality and multiplicity
Easy to get two steady states

Low (high) taxes =)
Surplus high (low) =)
Job destruction low (high) =)
Unemployment rate low (high) =)
Taxes indeed low (high)
Other models
Overview
Aiyagari model
Does it matter?
Multiple what?
Multiple steady states ; multiple solutions
Avoiding complexity
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Transition dynamics I
12%
to high
unemployment
expectations
matter
10%
8%
6%
to low
unemployment
4%
2%
-5
20
45
70
95
33
Other models
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Transition dynamics II
ut+1
negative expectations
positive expectations
45o
ut
34
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Why is it hard to get this published in AER?
What aspect of distribution determines whether this is

quantitatively important?
How do you get data on this?
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
References
Aiyagari, S. R., 1994, Uninsured Idiosyncratic Risk and Aggregate Saving, Quarterly
Journal of Economics.
classic paper using a model with capital.
Algan, Y., O. Allais, W.J. Den Haan, and P. Rendahl, Solving and Simulating Models
with Heterogeneous Agents and Aggregate Uncertainty.
Survey article available online.
Den Haan, W.J., 2001, The Importance of the Number of Dierent Agents in a
Heterogeneous Asset-Pricing Model, Journal of Economic Dynamics and Control.
Shows that models with two agents are problematic models to study risk of
individual agents.
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Den Haan, W.J., 2007, Shocks and the Unavoidable Road to Higher Taxes and Higher
Unemployment, Review of Economic Dynamics.
Matching model with heterogeneous agents with multiple steady states and
(sometimes) multiple equilibria.
Fuerst, T.S., 1991, Liquidity, loanable funds, and real activity, Journal of Monetary
Economics.
One of the rst articles to use the trick of the happy household with
heterogeneity within the period, but not across periods.
Huggett, M., 1993, The risk-free rate in heterogeneous-agent incomplete-insurance
economics, Journal of Economic Dynamics and Control.
classic paper using a model with bonds.
Krusell, P., and A.A. Smith, 1997, Income and Wealth Heterogeneity in the
Macroeconomy, Journal of Political Economy.
classic paper using a model with capital and aggregate risk.
Overview
Aiyagari model
Does it matter?
Avoiding complexity
Other models
Kubler, F., and K. Schmedders, 2002, Recursive Equilibria in Economies with

Incomplete Markets, Macroeconomic Dynamics.
discusses (non)existence and uniqueness.
Lucas, R.E. Jr., 1990, Liquidity and Interest Rates, Journal of Economic Theory.
The article that introduces the idea of using the happy household with
heterogeneity within the period, but not across periods.
Miao, J., 2006, Competitive Equilibria of Economies with a Continuum of Consumers
and Aggregate Shocks, Journal of Economic Theory.
discusses existence of a recursive equilibria in an environment like the one in
Krusell and Smith (1997).

Intro Heterogeneous Agents Model

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Intro Heterogeneous Agents Model

Hochgeladen von

Copyright:

Verfügbare Formate

Models with Heterogeneous Agents

August 10, 2015

Simple example & key issues

"Simple" model with heterogeneous agents

Solving the Aiyagari model

Does heterogeneity matter?

Simple example & key issues

Other models with heterogeneity

Simple example & key issues

First model with heterogeneous agents

Agents are ex ante the same,

Simple example & key issues

Subject to employment shocks:

Simple example & key issues

Simple example & key issues

Simple example & key issues

These are identical to those of the rep. agent version

Simple example & key issues

Simple example & key issues

fci,t ,ki,t+1 gt=0

ci,t + ki,t+1 = rt ki,t + (1

this is a relatively simple problem

Simple example & key issues

Individual agent - foc

ci,t + ki,t+1 = rt ki,t + (1

Simple example & key issues

What aggregate info do agents care about?

current and future values of rt and wt

Simple example & key issues

What aggregate info do agents care about?

=) agents are interested in all information that forecasts Kt

Simple example & key issues

Equilibrium - rst part

Individual policy functions that solve agents max problem

Simple example & key issues

Equilibrium - second part

Simple example & key issues

Transition law & timing

Simple example & key issues

Transition law & timing

Let gt be the cross-sectional distribution of capital

Simple example & key issues

Transition law & continuum of agents

Simple example & key issues

Does an equilibrium exist?

Does a recursive equilibrium exist?

Simple example & key issues

and cross-sectional distribution of expected payos

Simple example & key issues

Heterogeneity =) more reasons to expect multiplicity

Simple example & key issues

Wealth-recursive (WR) equilibrium

Simple example & key issues

Wealth distribution not su cient - Example

Simple example & key issues

Wealth distribution not su cient - Example

Thus: same wealth levels, but dierent outcome

Simple example & key issues

Wealth distribution may not be su cient!

Simple example & key issues

For now we assume that a wealth recursive equilibrium exists

Simple example & key issues

If a wealth recursive equilibrium exists

Suppose that recursive RE for usual state space exists