Skip to content

First-Order Linear Difference Equation


Definition

A first-order linear difference equation with constant coefficient is

xt+1=a xt+gt,t=0,1,2,…,x_{t+1}=a\,x_t+g_t,\qquad t=0,1,2,\ldots,

where a≠0a\ne0 is a real constant and (gt)t≥0(g_t)_{t\ge0} is a given forcing sequence. It is homogeneous if gt=0g_t=0 for all tt. For every initial value x0x_0 the unique solution is

xt=atx0+∑s=0t−1a t−1−s gs.x_t=a^t x_0+\sum_{s=0}^{t-1}a^{\,t-1-s}\,g_s.

Interpretation

The next state is the current state scaled by aa plus an external input. The solution formula is obtained by substituting the equation into itself: the initial value is multiplied by aa once per period, and the input gsg_s received at time ss has been multiplied by aa for the remaining t−1−st-1-s periods.

Two paths from x_0 = 10 to the dashed equilibrium x* = 6: for a = 0.5 the values fall monotonically, for a = -0.5 they jump above and below 6 while closing in.

The figure shows two solutions with constant input, both starting at x0=10x_0=10 and both with the equilibrium (constant solution) x∗=6x^*=6, dashed. The blue path is the convergent example xt+1=0.5xt+3x_{t+1}=0.5x_t+3, that is xt=6+4(0.5)tx_t=6+4(0.5)^t, which falls monotonically (10,8,7,6.5,…10,8,7,6.5,\ldots). The red path is xt+1=−0.5xt+9x_{t+1}=-0.5x_t+9, which has the same equilibrium 9/1.5=69/1.5=6; here xt=6+4(−0.5)tx_t=6+4(-0.5)^t alternates (10,4,7,5.5,…10,4,7,5.5,\ldots). Both transients shrink by the factor ∣a∣=0.5\lvert a\rvert=0.5 per period.

Move aa, bb and x0x_0: for ∣a∣<1|a|<1 the path settles on the dashed equilibrium x∗=b/(1−a)x^*=b/(1-a), monotonically when a>0a>0 and alternating around x∗x^* when a<0a<0. At a=1a=1 with b≠0b\ne0 there is no equilibrium and the path grows linearly.

With a constant input gt=bg_t=b and a≠1a\ne1, the constant x∗=b/(1−a)x^*=b/(1-a) is the unique equilibrium, and the deviation xt−x∗x_t-x^* satisfies the homogeneous equation. Hence xt=x∗+at(x0−x∗)x_t=x^*+a^t(x_0-x^*): an equilibrium plus a transient. The size of ∣a∣\lvert a\rvert decides whether the transient dies out (see stability of a scalar difference equation), and the sign of aa decides whether the path is monotone (a>0a>0) or alternates around x∗x^* (a<0a<0).

The case a=1a=1 is different: there is no equilibrium when b≠0b\ne0, and xt=x0+btx_t=x_0+bt grows linearly. Forgetting this case, and dividing by 1−a=01-a=0, is a common error.


Key Results

  • Homogeneous case: xt=atx0x_t=a^t x_0, a geometric sequence.
  • Constant input bb, a≠1a\ne1: xt=b1−a+at(x0−b1−a)x_t=\dfrac{b}{1-a}+a^t\Big(x_0-\dfrac{b}{1-a}\Big); by the geometric series formula, ∑s=0t−1at−1−sb=b 1−at1−a\displaystyle \sum_{s=0}^{t-1}a^{t-1-s}b=b\,\dfrac{1-a^t}{1-a}.
  • Constant input bb, a=1a=1: xt=x0+btx_t=x_0+bt.
  • If ∣a∣<1\lvert a\rvert<1 and gt=bg_t=b, then xt→b/(1−a)x_t\to b/(1-a) for every x0x_0; if ∣a∣>1\lvert a\rvert>1, the path diverges unless x0=x∗x_0=x^*.
  • a=−1a=-1 gives a two-cycle xt=x∗+(−1)t(x0−x∗)x_t=x^*+(-1)^t(x_0-x^*) around x∗=b/2x^*=b/2 (a constant path if x0=x∗x_0=x^*).

Examples

Convergent case. xt+1=0.5xt+3x_{t+1}=0.5x_t+3, x0=10x_0=10. The equilibrium is x∗=3/(1−0.5)=6x^*=3/(1-0.5)=6, so xt=6+4(0.5)tx_t=6+4(0.5)^t. Check: x1=0.5⋅10+3=8=6+4⋅0.5x_1=0.5\cdot10+3=8=6+4\cdot0.5, and xt→6x_t\to6.

Savings account. A balance earning 5%5\% per period with a deposit of 100100 at the end of each period satisfies Bt+1=1.05Bt+100B_{t+1}=1.05B_t+100, B0=0B_0=0. Here x∗=100/(1−1.05)=−2000x^*=100/(1-1.05)=-2000, so Bt=2000(1.05t−1)B_t=2000(1.05^t-1). Check: B1=100B_1=100 and B2=2000⋅0.1025=205=1.05⋅100+100B_2=2000\cdot0.1025=205=1.05\cdot100+100. The equilibrium is negative and unstable: it only describes the path, not a long-run level.

Time-varying input. xt+1=2xt+tx_{t+1}=2x_t+t, x0=0x_0=0: the formula gives x3=22⋅0+21⋅1+20⋅2=4x_3=2^2\cdot0+2^1\cdot1+2^0\cdot2=4, matching the recursion x1=0x_1=0, x2=1x_2=1, x3=2⋅1+2=4x_3=2\cdot1+2=4.


Prerequisites



Sources

Aliases: First Order Linear Difference Equation, first-order linear recurrence, linear difference equation of order one, linear first-order difference equation Keywords: x_{t+1} = a x_t + g_t, x_t = a^t x_0, equilibrium b/(1-a), steady state, forcing term Symbols: sum

Last updated: 17 October 2026 21:53