Problem Set #2Answer Key Economics 808:Macroeconomic Theory
Fall 2004
1Linear di?erence equations
See the Excel ?le di?eq.xls on the web site.
2A nonlinear di?erence equation
a )See the computer ?le.
b )We ?nd a ?xed point by solving the equation x ∞=0.5x 0.5∞for x ∞
.This gives us two ?xed points,x ∞∈{0,0.25}.
c )In general,f (x )=0.25
x 0.5
,so f (0)=∞1and f (0.25)=0.5.Zero is an unstable ?xed point,meaning that if the system starts near zero,it will tend to move away from zero.In contrast,the 0.25is a stable ?xed point,implying that if the system starts near 0.25it will tend to move towards 0.25.In addition,because f (0.25)>0,it will move monotonically towards 0.25.
d )Th
e simulated time series (in the computer ?le)shows x t monotonically converging to 0.25,just as the local analysis predicted.
3Another nonlinear di?erence equation
a )See the computer ?le.
b )We ?nd a ?xed point by solving the equation x ∞=ax ∞(1?x ∞)for x ∞.In general there are two solutions:
x ∞∈{0,a ?1
a }
When a =1.1,the two ?xed points are 0and 0.09.c )In general f (x )=a (1?2x ).This means
f (0)=a f (a ?1a
)=2?a
1
If you prefer,lim x →0f (x )=∞
1
The Central University of Finance and Economics
China Center for Human Capital and Labor Market Research
Advanced Macroeconomics
Problem Set 2
ECON808,Fall20042 In this case,f (0)=1.1and f (0.09)=0.9.Zero is an unstable?xed point and0.09is a stable ?xed point.
d)The simulated time series(in the computer?le)shows that x t converges monotonically to the stable?xed point.
e)See the computer?le.
f)The function f has two?xed points,0and0.655.
g)In this case f (0)=2.9and f (0.655)=?0.9.Zero is an unstable?xed point,and the0.655is stable.Since f (0.655)<0,this implies that starting near0.655,the system will converge to0.655 with oscillations.
h)The simulated time series(in the computer?le)shows that x t moves towards the stable?xed point,but oscillates,as predicted by the local analysis.
i)This plot is in the computer?le.
j)The?xed points of f are zero and0.748.
k)In this case f (0)=3.97and f (0.748)=?1.97.Both?xed points are unstable,implying that starting near either,the system will move away.
l)The simulated time series shows that x t appears to jump around randomly with no clear pattern even though each value is a deterministic function of the previous value.
The time series you see here is an example of a chaotic system.Chaos only appears in systems governed by deeply nonlinear di?erence equations.In fact,the way you get“random”numbers from a computer is through successive application of a nonlinear di?erence equation to some“seed”value(say,the current date in seconds since Jan1,1970).