ECN115 G. Renshaw ch.20: Difference & Differential Equations

ECN115
Lecture 10
G. Renshaw ch.20
Difference & Differential

equations
• linear difference equations with constant coefficients

• difference equations & market equilibrium
• the cobweb model of supply and demand
• first order linear differential equation with constant coefficients
• differential equations and dynamic stability/instability of a market
1
First order difference equations
First order difference equations: time varies discretely; value of y in
time period t, written as y t , depends on its value in one previous period.
(a) Homogeneous equations. Example: y t  (1.1)y t 1

A linear, homogeneous difference eqn. Its solution is the underlying
functional relationship between y and t.
That function (the solution) is: y t  A(1.1)t .
This is the general solution.
A is an arbitrary multiplicative constant. It is there because we only
have info about changes in y.
If we are given some info (called initial conditions) about levels we
can solve for A.
For example if we are told y1  99 we can substitute this into general
solution and get 99  A(1.1) , with solution A = 90.
1
Thus y t  90(1.1) is the unique solution.

t
Generalizing: the differential equation y t  by t 1 has general

solution y t  A( b ) .
t
Unique solution can be found, given initial conditions. (rule 20.1)

(b) Non−homogeneous equations. General form y t  by t 1 c .
Additive constant, c.
General solution is then y t  A( b )  1cb .

t
As with homogeneous equations, we can solve for A if we are given

some initial conditions, and thus find unique soln.
If for example we know that y has the value y0 when t = 0, we can

substitute that into the general solution and get y 0  A  c ,
1b
so A  y 0  1cb . (rule 20.2).
difference equations
5
Qualitative analysis
Look at general solution y t  A(b )t  c .
1b
What happens to y as time passes (t increases)?
Four cases:
(1) b > 1. Then (b )t increases without limit.

If A > 0, y increases without limit.
If A < 0, y decreases without limit. Monotonic divergence.
(2) 0 < b < 1.

Then ( b )t gets smaller and smaller, and y approaches 1cb .
Monotonic convergence.
(3) −1 < b < 0.
Then ( b )t gets smaller and smaller in absolute value, and alternates
between positive and negative as t is even or odd.
c
So y approaches 1b . Oscillatory convergence.
(4) b < −1.

Then ( b )t gets larger and larger in absolute value, and alternates
between positive and negative as t is even or odd.
Oscillatory divergence. (rule 20.2)
Non-homogeneous difference equations
8
Time paths
example 20.1
example 20.2
9
Time paths
example 20.3
example 20.4
10
Economic application:
the cobweb model
A supply/demand model with supply a function of previous period's
price:
S D
qt  hpt 1  j and qt   mpt  n (linear fns)
In each period, supply and demand must be equal, so qtS  qtD .

From these 3 equations we can derive:
  j n
pt    h  pt 1  m (1st order, non−homogeneous difference eqn.)
 m
t
  j n
Using rule 20.2, general solution is: pt  A h   m h
 m
the cobweb model of supply and demand
t
 h
Depending on the value of   , the behaviour of p through time will
 m
follow one of the 4 cases above.
For stability in this market, we want p to converge towards some long run
equilibrium value.
 
(1)   h  > 1.
 m
Monotonic divergence; p increases or decreases without limit. The market
is unstable.
 h
(2) 0 <  < 1.
 m j n
Monotonic convergence; p approaches mh . The market is stable
 h
(3) −1 <   < 0.
 m j n
Oscillatory convergence; p approaches mh but in any period is
always either above or below the limit. Stable. (see fig. 20.5)
 h
(4)   < −1.
 m j n
Oscillatory divergence; in any period, p is always above or below mh
and gets further and further away. Unstable. (See fig 20.6)
Cobweb diagram for Cobweb diagram for

5 10
pt  1( )t  8, showing oscillatory Pt  1.222( )t  5.778 , showing
8 8
convergence of pt oscillatory divergence in pt
14
Which of these 4 is most likely? The slope of the demand function is −m,
so assuming the demand function slopes downward, −m < 0, so m > 0.
 
We also have h > 0 if the supply function slopes upward. Then   mh  < 0
and we have either case (3) or case (4).
If h < m (that is, the absolute slope of the supply fn is less than the
absolute slope of the demand fn) we have case (3); the convergent
cobweb of fig 20.5.
If h > m (that is, the absolute slope of the supply fn is greater than the
absolute slope of the demand fn) we have case (4); the divergent cobweb
of fig 20.6
differential equations
First order differential equations: y and time (t)

vary continuously. Rate of change of y given by
dy
dt
(a) Homogeneous equations. General form:
dy
dt
 by , (b constant). Linear, homogeneous.
General solution y  Aebt where A arbitrary.
(Check: ddyt  bAebt  by ).
(b) Non−homogeneous equations General form:
dy bt c
dt
 by  c . General solution y  .
Ae 
b
dy
[Check: dt  bAebt  b( y  cb )  by  c
differential equations
Given initial conditions we can solve for A in

both cases (a) and (b) If for example we know
that y = y(0) when t = 0, we can substitute that
information into the general solution y  Aebt  cb
and get y (0)  A  c , so A  y (0)  c . (rule 20.4).
b b
Linear first order homogeneous
18
Linear first order non-homogeneous
19
Time paths - qualitative analysis
Monotonic increasing
A > 0, b > 0
Convergent from above

A > 0, b < 0
Monotonic decreasing
A < 0, b > 0 Convergent from below
A < 0, b < 0
20
Economic application: lagged price
adjustment
Supply and demand functions:
qt S  hpt  j and qt D  mpt  n (linear, no lags)
In long run equilibrium, qt S  qt D and therefore
hpt  j  mpt  n . Solving this for pt gives the long
n j
run equilibrium price, p*, as p *  m h
However, we assume demand and supply not
necessarily equal at all times. When demand
greater (less) than supply, price rises (falls).
Dynamic stability of a market
20.35
22
adjustment
dp
This adjustment process modelled as  k (q D  q S )
dt
(where k is a positive parameter).
By substituting the supply and demand
functions into the price adjustment equation we
obtain the differential equation:
dp
 k ( h  m ) p  k ( n  j )
dt
Applying rule 20.4 the general solution is
k (n j )
p  Ae k (hm )t  k (hm )  Ae k (hm )t  p*
adjustment
If we know that p = p(0) when t = 0, we can
substitute that into our solution and get
p(0)  A  p* so A  p(0)  p* .
So our unique solution is

p  ( p(0)  p* )ek (hm)t  p*
Given p  ( p(0)  p* )ek (hm)t  p*, there are 4 cases
to be considered.
Two cases arise according to whether ( p(0)  p* ) is
positive or negative (that is, whether the initial
price p(0) is above or below the long run
equilibrium price p*).
Within each of those 2 cases, there are 2 cases
according to whether h + m is positive or
negative.
See fig 20.8.

There, cases 1(a) and 1(b) arise when h + m < 0.
Here the market is dynamically unstable.
Cases 2(a) and 2(b) arise when h + m > 0.

The market is dynamically stable.
Figure 20.8: Cases 1(a) and 1(b): a dynamically unstable market;
cases 2(a) and 2(b): a dynamically stable market
Stability/Instability
conditions
28
 Study:
from G. Renshaw’s ch. 20,
 Attempt:
all relevant progress exercises
 for next week’s class/seminar prepare the so-called tutorial

exercises
 for next week’s teamwork prepare & upload the so-called

project exercises AS PART OF THE SINGLE DOCUMENT
WITH ALL 30 PROJECT EXERCISES!!!
29

ECN115 G. Renshaw ch.20: Difference & Differential Equations

Uploaded by

Copyright:

Available Formats

ECN115 G. Renshaw ch.20: Difference & Differential Equations

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ECN115 G. Renshaw ch.20: Difference & Differential Equations

Uploaded by

Copyright:

Available Formats

ECN115

Difference & Differential

• linear difference equations with constant coefficients

(a) Homogeneous equations. Example: y t  (1.1)y t 1

Thus y t  90(1.1) is the unique solution.

Generalizing: the differential equation y t  by t 1 has general

Unique solution can be found, given initial conditions. (rule 20.1)

General solution is then y t  A( b )  1cb .

As with homogeneous equations, we can solve for A if we are given

If for example we know that y has the value y0 when t = 0, we can

(1) b > 1. Then (b )t increases without limit.

(2) 0 < b < 1.

(4) b < −1.

In each period, supply and demand must be equal, so qtS  qtD .

Cobweb diagram for Cobweb diagram for

First order differential equations: y and time (t)

Given initial conditions we can solve for A in

Convergent from above

So our unique solution is

See fig 20.8.

Cases 2(a) and 2(b) arise when h + m > 0.

 for next week’s class/seminar prepare the so-called tutorial

 for next week’s teamwork prepare & upload the so-called

You might also like