You are on page 1of 3

Discounting:

Discrete versus Continuous Compounding


Yves Zenou
December 5, 2006

Discrete compounding

Suppose that someone has X dollars to invest currently at an annual interest


rate r. Then the amount of money received at the end of one year would
be V = X (1 + r). Therefore it is equivalent to say that the present value
of amount V to be received in one years time is X = V / (1 + r). We can
generalize this result for t periods and thus, in discrete time, the present
value of present value P Vt of an amount V to be received t periods from
now is given by:
V
(1)
P Vt =
(1 + r)t
where r is the interest rate per period and compounding occurs at the end
of each period.
Observe that for r > 0, the denominator (1 + r)t becomes larger as
t becomes larger, and thus P Vt gets smaller. In other words, receiving
a certain sum in the future has a lower present value the longer one has
to wait for the payment. This is natural since the further in the future
one receives the fixed amount V , the less one would need to invest now
to replicate that future payment. For this reason, economists refer to the
discounting of future benefits and the value 1/ (1 + r) is referred to as the
discount rate, or discount factor. Moreover, (1 + r)t grows without bound
as t +, and so P Vt 0 as t +.

Continuous compounding

Imagine now that compounding occurs n time a year. By continuous compounding, we then mean that interest is compounded instantaneously, i.e.

Research Institute of Industrial Economics, Box 55665, 102 15 Stockholm, Sweden,


and GAINS, Universite du Maine, France. E-mail: yves.zenou@riie.se

n +. Thus, if someone has X dollars to invest currently at an annual interest rate r with compounding n times a year, the amount of money
received at the end of t years would be V = X (1 + r/n)n t . Thus the appropriate discount formula for finding the present value P Vt of a future amount
V when discounting occurs n time per year is:
V
P Vt =
n t
1 + nr

Lemma 1 We have:

1 s
1+
=e
s+
s
lim

(2)

Proof. Let us first remind the LHpital rule. Suppose that f and g
are continuous and dierentiable functions, on an open interval containing
x = a, except possibly at x = a, and that
lim f (x) = + and lim g(x) = +

xa

Then, if limxa

f 0 (x)
g 0 (x)

xa

has a finite limit, or if this limit is + or , then:


f (x)
f 0 (x)
= lim 0
xa g(x)
xa g (x)
lim

Moreover, this statement is also true in the case of a limit as x a ,


x a+ , x , or as x +.
Let us introduce a new variable: w 1/s. Then s = 1/w. Upon
substituting these values for w into (2), and noting that as s +, w 0,
we have:

1 s
= lim (1 + w)1/w
lim 1 +
s+
w0
s
Now let us introduce a dependent variable
y = (1 + w)1/w
If we take the logarithm of both sides of (3), we have:
i
h
log y = log (1 + w)1/w
=

1
log (1 + w)
w
log (1 + w)
w

Thus,
log (1 + w)
w0
w

lim log y = lim

w0

(3)

which is an indeterminate form of the type 0/0. By the LHpital rule, we


have:
log (1 + w)
1/ (1 + w)
= lim
=1
lim log y = lim
w0
w0
w0
w
1
So we have shown that log y 1 as w 0, the continuity of the exponential
function implies that elog y e1 as w 0, and this implies that y e as
w 0. In other words, we have shown that:

1 s
= lim (1 + w)1/w = e
lim 1 +
s+
w0
s

Lemma 2 We have:

r n t
= ert
1+
n+
n
lim

Proof. Let us calculate

r n t
1+
n+
n
lim

First, introduce a new variable s n/r. Then r/n = 1/s and n = s r. Upon
substituting these values for r/n and n into the formula above, and noting
that as n + so does s +, we have:

r t

1 srt
1 s
r n t
1+
= lim 1 +
= lim
lim 1 +
n+
s+
s+
n
s
s
Ignoring for a moment the term r t, we see from Lemma 1 that:

1 s
lim 1 +
=e
s+
s
It follows that

r t

1 s
lim
1+
= ert
s+
s

and the result follows by changing back the variable to s n/r.


As a result, if interest is compounded continuously, the amount (X dollars) needed to generate V dollars in t yearss time is determined by:
V = X ert
and so the appropriate discounting formula under continuous discounting is
given by:
V
(4)
P Vt = rt = V ert
e
3

You might also like