/ Garden

Discount factors

Note on the time value of money and discount factors.

What is a discount factor? Why does it exist?

In introductory finance courses, we learn that money today is worth more than money tomorrow due to the time value of money. The discount factor is the quantification of this concept.

Suppose we have a utility of consumption,

U(c0,c1,,cT)U(c_0, c_1, \dots, c_T)

where ctc_t is consumption at time tt. Given an investment opportunity that pays xx at time TT and costs pp at time tTt\leq T; buying nn units of this investment will give us a payoff of nxnx at time TT and cost us npnp at time tt. We will buy the investment if it increases our utility, i.e. if

U(c0,,ctnp,,cT+nx)>U(c0,,ct,,cT).U(c_0, \dots, c_t - np, \dots, c_T + nx) > U(c_0, \dots, c_t, \dots, c_T).

Adding and subtracting U(c0,,ct,,cT)U(c_0, \dots, c_t, \dots, c_T)

U(c0,,ctnp,,cT+nx)U(c0,,ctnp,,cT)>U(c0,,ct,,cT)U(c0,,ctnp,,cT).U(c_0, \dots, c_t - np, \dots, c_T + nx) - U(c_0, \dots, c_t - np, \dots, c_T) > U(c_0, \dots, c_t, \dots, c_T) - U(c_0, \dots, c_t - np, \dots, c_T).

if and only if

xU(c0,,ctnp,,cT+nx)U(c0,,ctnp,,cT)nx>pU(c0,,ct,,cT)U(c0,,ctnp,,cT)np.moremoremoremoremoremox\frac{U(c_0, \dots, c_t - np, \dots, c_T + nx) - U(c_0, \dots, c_t - np, \dots, c_T)}{nx} > p\frac{U(c_0, \dots, c_t, \dots, c_T) - U(c_0, \dots, c_t - np, \dots, c_T)}{np}.moremoremoremoremoremo

Taking the limit as n0n\to 0,

xU(c0,,ctnp,,cT+nx)cT>pU(c0,,ctnp,,cT)ct.x\frac{\partial U(c_0, \dots, c_t - np, \dots, c_T + nx)}{\partial c_T} > p\frac{\partial U(c_0, \dots, c_t - np, \dots, c_T)}{\partial c_t}.

In fact, this holds with equality as well, since we can always find a small enough nn such that the inequality holds. Rearranging,

p=xU(c0,,ctnp,,cT+nx)cT/U(c0,,ctnp,,cT)ct.p = x\frac{\partial U(c_0, \dots, c_t - np, \dots, c_T + nx)}{\partial c_T} \bigg/ \frac{\partial U(c_0, \dots, c_t - np, \dots, c_T)}{\partial c_t}.