Chapter 5: Difference Equations

5.1. Definitions and concepts

In the continuous-time context, the pattern of change of a variable y is embodied in the
derivative of y’ (t), Y” (t). Etc. the time change involved in these is infinitesimal in magnitude.
While time is fixed instead, we can take to be a discrete variable. So that the variable t is allowed
to take only integer values, the concept of the derivative obviously will no longer be appropriate.
Then, the pattern of change of the variable y must be described by so-called “difference,” rather
than by derivatives or differentials, of y ¿). Accordingly, the techniques of differential equations
will give way to those of difference equations.
When we are dealing with discrete time, the value of variable y will change only when the
variable t changes from one integer values to the next, such as ¿ t=1(denoting period one) to t=2
(denoting period two), etc. then we may simply regard y as having one unique value in each time
period. In view of this interpretation, the discrete-time version of economic dynamics is often
referred to as period analysis. When the time changes from one-integer value to the other,
economic variable changes from one integer value to another value.
The change from continuous time to discrete produces no effect on the fundamental nature of
dynamic analysis, although the formulation of the problem must be altered. Our dynamic
problem is still to find a time path from some given pattern of change of a variable y over time.
However, the pattern of change should now be represented by the difference quotient Δy / Δt ,
which is the discreste-time counterpart of the derivative dy /dt . however, that t can now take only
integer values: thus, when we are comparing the values of y in two consecutive periods, we must
have Δt =1. for this reason, the difference question Δy / Δt can be simplified to the expression Δy
; this is called the first difference of y . the symbol Δ , meaning difference, can accordingly be
interpreted as a directive to take the first difference of (y).
The expression Δy can take various values, of course, depending on which two consecutive time
periods are involved in the difference-taking. By adding a time subscript to y, the first difference
more specified as follows:
Note: The order of a difference equation is determined by the greatest number of periods
lagged. A first order difference equation expresses a time lag of one period; second order
difference equation a two period time lag and so on.

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 1

 Chapter 5: Difference Equations

Example1:
The change in y as t changes ¿ t ¿ t +1 iscalled the first difference of y .
Δy
i .e . = Δ yt= y t +1− y t [ first order ]
Δt
Most of the time increment (Δ)is omitted, and we write like this
I t=a ( y t−1− y t−2 ) [ second order ]
Qt =a+ b p t−1 [ first order ]
y t +3−9 y t +2 +2 y t +1+ 6 y t=8 [ third order ]
y t +1− y t =2 (First order)
y t +1= y t + 2 (First order)
y t +1−0.9 y t =0 (First order)
y t +1=0.9 y t (First order)
Analogous to differential equations, difference equations can be ether linear or non-linear,
homogeneous or non-homogeneous, and of the first, second, or higher orders.
5.2. First order difference equation
In solving a differential equation, our objective was to find a time path y (t ). As we know, such a
time path is function of time which is totally free from any derivative (or differential)
expressions and which is perfectly consistent with the given differential equation as well as with
its initial conditions.
Iteration methods:
The first order difference equation describes the pattern of change of y between two consecutive
periods only. Hence, once the difference equation is specified and an initial value y 0is given, it is
possible to find y, from the equation. Similarly, once y 1is found, y 2will be immediately obtained,
and so on, b y repeated application (iteration) of the pattern of change specified in the difference
equation. The results of iteration will enable us to inter a time path or the variable under
consideration.

Examples 2:
1) Find the solution of a difference equation Δ y t =2 assuming aninitial value of
y=15 ⇒ y 0=15
solution:
since Δ y t = y t +1− y t , Δ y t =2⇒ y t +1= y t +2

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 2

 Chapter 5: Difference Equations

Then by successive substitutions of t=0 , 1 ,2 , 3 , e . t . c we obtain

t=0 : y 1= y 0 +2
t=1 : y 2= y 1 +2=( y 0+ 2 ) +2= y 0 + 4= y 0 +2(2)
t=2 : y 3= y 2 +2= [ y 0+ 2(2) ]+2= y 0 +6= y 0 +3(2)
t=3 : y 4 = y 3 +2=[ y 0 +3 (2) ] +2= y 0 +8= y 0+ 4 (2)
. . . .
.
. . .
For any period t=t , y t = y 0 +t (2)
⇒ y t =15+2 t ( this is the time path of the variable ) .
Example 3:
2 Find the solution of the difference equation y t +1=b y t given that the initial value of y is y 0.

Solution:

Given y t +1=b y t

At t=0 : y 1=b y 0
2
At t=1: y 2 =b y 1=b ( b y 0 )=b y 0
At t=2: y3 =b y 2=b ( b2 y 0 )=b3 y 0
. . .
. . .
t
At t=t : y t =b y 0 (The solution of the difference equation)

3. Find the solution of the difference equation:

y t +1=0.9 y t
At t=0 : y 1=0.9 y 0
2
At t=1: y 2 =0.9 y 1=0.9 ( 0.9 y 0 )=0.9 y 0
At t=2: y3 =0.9 y 2=0.9 ( 0.9 2 y 0 ) =0.93 y 0
. . .
. . .
t
At t=t : y t =( 0.9) y 0 (The solution of the difference equation)

4. Solve the homogeneous differential equation:

m y t+ 1−n y t=0
n
This can be transformed in to the form of y t +1= y t
m

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 3

 Chapter 5: Difference Equations

n
At t=0 : y 1= y
m 0

( )
2
n n n n
At t=1: y 2 = y 1= y 0 =( ) y 0
m m m m

( ) ( )y
2 3
n n n n
At t=2: y3 = y = ( ) y0 =
m 2 m m m 0

. . .
. . .
t
n
At t=t : y t =( ) y 0 (The solution of the difference equation)
m

The general solution for difference equation:

Given the first order difference equation: y t +1 +a y t =c , where a and c are two constants. Its
general solution will consist of the sum of a complimentary function ( y ¿¿ c )¿ and a particular
integral ( y p), which is any solution of the complete non-homogeneous equation and a
complementary function Yc, which is the general solution of the reduced equation.
Definition: the y p component again represents the inter-temporal equilibrium level of y and y c
represents the deviation of the time path from that equilibrium level. For the time path to be
dynamically stable, y c must approach zero as t approaches to infinity.
From the above given difference equation:
 y c -is the general solution of the reduced (homogeneous) from equation.
y t +1 +a y t =0
 y p-is the general solution of the complete or non-homogeneous equation.
y t +1 +a y t =c
A. First find y c for the homogeneous equation:
Given: y t +1 +a y t =0 ⇒ y t +1=−a y t
From our experience of iteration methods of solving the difference equation, the general solution
will be of the form y t =A bt (see the above examples in iterative method) where,
A b ≠ 0 , for otherwise y t becomes a horizontal straight line on t-axis. In that case the
t

complimentary solution will be y c = A b t .

t t +1
y t =A b , y t +1= A b , substitute in the reduced form equation
y t +1 +a y t =0

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 4

 Chapter 5: Difference Equations

t +1 t
Ab +a A b =0
t t
A b b+a A b =0
t
A b (b+ a)=0
t
A b =0 And b+ a=0⇒ b=−a
t
y C = A (−a)
B. Next, find the particular integral for the non-homogeneous equation:
Given y t +1 +a y t =c , we can choose any solution of particular integral ( y p). Thus, we take the
simplest form of the trial solution: y t =k (a constant). Then, since y maintains the same constant
value over time, we must also have y t +1=k . Then, substitute these values in to the complete
equation:
c
y t +1 +a y t =c ⇒ k +ak =c ⇒ k ( 1+a )=c ⇒ k =
1+a
c
Hence, the particular solution becomes y p=k ⇒ y p= , a ≠−1.
1+a
c
Note that since is a constant, a stationary equilibrium is indicated in this case. Since y pis
1+ a
undefined at a=−1, we need to find some other solution for the non-homogeneous equation. So,
let us try a solution of the form y t =kt , which indicates a moving equilibrium, y t =kt implies
y t +1=k (t+1).
Substituting these values in the complete equation, we obtain
y t +1 +a y t =c ⇒ k ( t+1 )+ a ( kt )=c
⇒ k ( t +1 ) +kt=c , since a=−1
⇒ k ( t +1+t )=c ⇒ k =c thus , y p=kt ⇒ y p =ct , for a=−1
C. Finally, adding y p∧ y c we arrive at the general solution
t c
y t = y c + y p ⇒ y t= A (−a) + … … . ( 1 ) General solution when a ≠−1.
1+a
t t
y t =A (−a) +ct ⇒ y t =A (−a) + ct … … … . ( 2 ) General solution whena=−1.
Eliminating the arbitrary constant, the definite solution will be written as:
0 c c
From (1) at t=0 , y 0= A (−a) + ⇒ A= y 0 −
1+ a 1+ a

(
y t = y 0−
c
1+ a )t
(−a) +
c
1+ a
( 1 ' ) [ Definite solution, when a ≠−1 ]

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 5

 Chapter 5: Difference Equations

From (2) at t=0 , y 0= A+ c ( 0 )= A

y t = y 0 +ct ( 2' ) [ Definite slution , when a=−1 ]
Example 4:
7
Solve the first order difference equation y t +1−5 y t =1 ,( y 0= ).
4
Solution:
 The complementary solution (Yc): - solution for reduced form of y t +1−5 y t =0. let
t t+ 1
y t =A b , y t +1= A b , by substitution, A bt +1−5 A b t=0
t t
A b b−5 A b =0
A b (b−5)=0 , ⇒b=5
t

Thus, the complementary solution (Yc) is Yc= A (5)t

 Particular solution: the solution of general equation of y t +1−5 y t =1.

Let y t =k , y t +1=k , then by substitution: k −5 k=1 ⇒k =

−1
4
−1
Therefore, the particular integral solution is Yp=
4
t 1
The general solution is Yt =Yp+ Yc ⇒ Yt =A (5) −
4
The definetized value of the general solution when t=0 :

Y 0= A− = ⇒ A= + =2
1 7 7 1
4 4 4 4
Thus, Y t =¿
Example 5:
Given the difference equation, y t =−7 y t −1 +16∧ y 0=5 ,find the time path.
Solution:
This equation re-written as y t +7 y t−1=16 , ⇒ y t +1 +7 y t =16 ;( y 0 =5)
 The complementary solution (Yc): the general solution of reduced form of y t +1 +7 y t =0.
Let, y t =A bt , y t +1= A b t+ 1, by substitution, A bt +1 +7 A b t=0
t t
A b b+7 A b =0
A b (b+7 )=0, ⇒b=−7
t

Thus, Yc= A(−7)t

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 6

 Chapter 5: Difference Equations

 The particular integral solution (Yp): the general solution of complete equation of the form
y t +1 +7 y t =16.
Let, Yt =k , y t +1=k
Then, substitute these values in the general equation:
y t +1 +7 y t =16
k +7 k=16
8 k =16 ⇒ k =2, thus, the y c =2
The general solution is Yt =Yc+Yp=A (−7)t +2
The definite solution is t=0 , Y 0=5 , A(−7)0+ 2=5
A+2=5 ⇒ A=3

Therefore, the definite solution is Yt =3 ¿

Activities:
1. Solve the following difference equation using iteration methods.

2. For each of the following difference equations find Yc, Yp and the definite solution.

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 7

 Chapter 5: Difference Equations

Second order difference equations

The economic model in the preceding chapter involve difference equations that relate Pt and Pt −1
to each other. As the p value in one period can uniquely determine the p value in the next, the
time path of P becomes fully determinate once an initial value P0 is specified. It may happen,
however, that the value of an economic variable in period t (say, Y t ) depends not only on Y t −1
but also on Y t −2. Such a situation will give rise to a difference equation of the second order.
Strictly speaking, a second-order difference equation is one that involves an expression Δ 2 y t ,
called the second difference of Y t , but contains no differences of order higher than 2. The symbol
2 2 2
Δ , the discrete-time counterpart of the symbold y /dt , is an instruction to “take the second
difference” as follows:
2
Δ y t =Δ ( Δ Y t )=Δ (Y t +1−Y t )
2
Δ y t =( Y t +2−Y t +1) −(Y t+ 1−Y t )
2
Δ y t =Y t +2−2 Y t +1 +Y t
Thus, a second order difference of Y t is transformable into a sum of terms involving a two-period
time lag.
Economic Applications of first order difference equations
 The cobweb Model
To illustrate the use of first-order difference equations in economic analysis, we shall cite two
variants of the market model for a single commodity. The first variant, known as the cobweb
model, differs from our earlier market models in that it treats Qs as a function not of the current
price but of the price of the preceding time period.
Consider a situation in which the producer’s output decision must be made one period in advance
of the actual sale-such as in agricultural production, where planting must proceed by an
appreciable length of time the harvesting and sales of the output. Let us assume that the output
decision in period t is based on the then-prevailing price Pt. since this output will not be available
for the sale until period (t+1), however, Pt will determine not Qs ,t +1. Thus, we now have a
“lagged” supply function: Qs ,t +1=S(Pt ) or equivalently, by shifting back the time subscripts by
one period. Qst =S (Pt −1)
When such a supply function interacts with a demand function of the form Qdt =D(Pt )
interesting dynamic price patterns will result.

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 8

 Chapter 5: Difference Equations

Let, given the non-lagged demand function Qdt=α −βPt and lagged supply function of the form
Qst =−γ +δ Pt −1, where∂ , β , γ , δ > 0. The market clearing condition exists when Qdt=Qst . Then
by substitution, the model can be reduced to a single first-order difference equation as follows:
β P t +δ Pt −1=α + γ .
Therefore, in order to solve this equation, it is desirable first to normalize it and shift the time
subscripts ahead by one period[ alter t ¿(t +1) ]. The result to be:
δ α+γ
Pt +1 + Pt =
β β
This equation is first-order linear difference equation with constant coefficient and constant term
of “a” and “c” respectively. since δ∧β are both positive, and the constant coefficient is differ
from -1, the solution for complementary function is : Pc=A ( b )t , where b=−a .

( )
t
−δ
Pc=A
β

( )
t
α+γ −δ α +γ
In addition, the particular solution is Pp= . Therefore, Pt =Pc + Pp= A + .
β +δ β β+ δ

( )( )
t
α+γ −δ α+γ
The definite value of the time path is Pt = P0 − + .
β +δ β β +δ
α +γ
Three points may be observed about this time path. In the first place, the expression , which
β+ δ
constitutes the particular integral of the difference equation, can be taken as the inter-temporal
α+γ
equilibrium price of the model: P= .
β +δ
Since it is a constant, it is a stationary equilibrium. Substituting P into our solution, we can

( )
t
−δ
express the time path Pt alternatively in the form of: Pt =( P0−P ) + P.
β
Two points may be observed in regard to this time path:
i. The significance of the expression ¿ ¿) : (same as A in A bt )
a) Its sign tells us whether the time path commences (starts) above or below equilibrium
(mirror effect)
b) Its magnitude helps us to determine how far above or below the time path is (scale
effect)

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 9

 Chapter 5: Difference Equations

−δ
ii. Lastly, there is the expression ( ¿, which corresponds to the b component of Ab t. From
β
our model specification that β∧δ all are positive. Then, there can arises three possible
varieties of oscillation pattern in the model.
The oscillation pattern will be:
 Explosive (when δ > β ¿ - the interaction of demand and supply will produce an explosive
oscillation
 Uniform (when δ=β ¿- the interaction of demand and supply will produce uniform
oscillation
 Damped (when δ < β ¿ - the interaction of demand and supply produce damped oscillation
/convergent
The above condition explained by the following graphical expression:
When δ > β , given an intial price Po assume price above the equilibrium price, we can follow the
arrowhead and read off on the S curve that the quantity supplied in the next period (period 1) will
be Q1. In order to clear the market, the quantity demanded in period 1 must also be Q 1, which is
possible if and only if price is set at the level of P 1 (see downward arrow). Now, via the S curve,
the price P1 will lead to Q2 as the quantity supplied in period 2, and to clear the market in the
latter period, price must be set at the level of P2 according to the demand curve. Repeating this
reasoning, we can trace out the prices and quantities in subsequent periods by simply following
the arrowheads in the diagram, thereby spinning a “cobweb” around the demand and supply
curves. By comparing the price levels, P0, P1, P2, we observe in this case not only an oscillatory
pattern of change but also a tendency for price to widen its deviation from P as time goes by.
With the cobweb being spun from inside out, the time path is divergent and the oscillation
explosive.
On the other hand, in the case of diagram b, whereδ < β the spinning process will create a cobweb
which is centripetal. From P0, if we follow the arrowheads, we shall be led ever closer to the
intersection of the demand and supply curves, where P is. While still oscillatory, this price path
is convergent.

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 10

 Chapter 5: Difference Equations

AU, SBE: DEPARTMENT OF ECONOMICS: Lemi T.Page 11