You are on page 1of 5

SIAM REVIEW (C) 1983 Society for Industrial and Applied Mathematics

Vol. 25, No. 2, April 1983 0036-1445/83/2502-0007 $01.25/0

THE CONCEPT OF ELASTICITY IN ECONOMICS*


Downloaded 03/26/14 to 140.120.135.222. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php

YVES NIEVERGELT"
Abstract. The following pages introduce the notion of elasticity and present its main mathematical and
graphical properties, illustrating them on real economic cases, such as the consumption of marijuana and the
demand for hospital beds.

1. Introduction. Economists often use the notion of elasticity of a function, which


they define to be (see [4, pp. 116-117])
(AF)/F x F(x + Ax) F(x)
(1) riF(X) lim lim
-.o (Ax)/x ax- o F(x) Ax
where F is usually a positive function defined on all or part of the positive real axis *+. In
this situation, OF(X) exists if and only if the derivative F’ (x) does, and then
X
(2) riF(X) F’(x).
F(x)
Thus, riF(X) measures the ratio of the relative change in F(x) to the corresponding relative
change in x, whereas F’(x) measures the ratio of their absolute changes.
This article explains how the concept of elasticity relates to real life, shows how to
visualize ri on logarithmic graph paper, and presents a few calculus rules governing
elasticities. This may provide useful suggestions to those whose interests lie in fields other
than economics, but who teach calculus to business students. First, we illustrate the above
definition by the two kinds of functions most frequently encountered in the classroom and
in practice.
Example 1.1. Power functions of the type F" *+ *+, x A x" have a constant
elasticity, equal to c (and conversely, by (9) below):
x x
(3) riF(X) F’(x) A o x"-1 o.
F(x) A. x"
Example 1.2. Consider an affine function D" ]0, b/m[ *+, p w- -m p + b with
positive coefficients m and b. Then (see Fig. 1)"
p P p OP
(4a) rio(P) D’(p)= (-m)=
D(p) -mp +b p- b/m AP
Since -mp D(p) b, one can also write:
p mp D(p) b BQ
(4b) rio(p) D’ (p)
D(p) D(p) D(p) OQ
2. Application. Suppose that p represents the unit price of some commodity, say in
dollars per item, and q D(p) the quantity sold at price p in one time period, say in items
per week. (Economists call D a demand curve.) Denote P the inverse function of D, so that
p P(q); then rio determines the maximum of the total revenue function TR as follows:

*Received by the editors March 26, 1982, and in revised form June 17, 1982. The work of this author was
partially supported by a grant from the Swiss National Research Fund.
’Department of Mathematics, University of Washington, Seattle, Washington 98195.

261
262 CLASSROOM NOTES

Pb I Price
Downloaded 03/26/14 to 140.120.135.222. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php

nD(p) -1

rD(p) >-1

Quantity
0 qo q

MR(q) 0 MR(q) 0

FIG. 1. The domain of D lies or, the vertical axis, whereas TR and MR have theirs on the horizontal
axis.

since TR(q):= p q P(q) q (price times quantity), differentiating with respect to q


yields:

TR’-P+q.P’=P.
+.. =P. + op =P. +
r P
Consequently, the marginal revenue MR(q) TR’(q) is positive where ro(p) < and
negative where nv(P) > (with p P(q)). For instance, if D is an affine function as in
Example 1.2, then total revenues reach their maximum at the midpoint (qo, P0), where
ro(po) by (4a), and therefore TR’(q0) 0 (see Fig. 1). Students will enjoy working
out the following two real cases:
Example 2.1. Belinfante and Davis found that the demand for record albums was
q D(p) -88.3 p + 1821, so that p -0.01133 q + 20.62. Hence, total reve-

q0 1821 / 2
-
nues will be largest with a price tag Po--20.62/2, i.e. $10.31 per album, and then
910 records will be sold.
Example 2.2. Hogarty and Elzinga [3] estimated the price elasticity of demand for
beer to be a constant r/o 1, with q D(p) 123/p (where p is in cents per can and q
in cans per adult per day). In this case, total revenues do not depend on price:
TR(q) TR oD(p)= p. D(p)= 123 (per adult per day) regardless of price and
quantity.
3. Visualizing r. Plotting a function F against logarithmic scales produces a curve
whose .slope equals the elasticity/’r of F: performing the changes of coordinates
(5) 4:u--e and ff:z-*ln(v)
and forming the composite (which amounts to changing scales)
(6) f:= p o Fo q,
CLASSROOM NOTES 263

one obtainsf’ (’ o (F o 4)) (F’ o q) 4/. Hence,


Downloaded 03/26/14 to 140.120.135.222. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php

e
(7) f’(u)- F’(e") e
F’(eu) rlr(eU),
F(e u) F(eu)
so that the slope off at u is indeed the elasticity of F at x e u. An alternative argument
runs as follows"
x dF
(1/F(x)) dF d[Log(F(x))]
n (x)
F(x) (1/x) dx
dx d [Log (x)]
(8)
lim Log [F(x + Ax)] Log IF(x)]
ax-.O Log (x + Ax) Log (x)
which equals the slope offat u In (x). Observe that both arguments hold with any base
for the logarithm, e.g. l0 or e. Of course, the slope off must be measured against linear
scales to yield r/F. This gives an easy way to locate the values of x where F(X) 1" find
wherefhas slope 1.
The next example will catch students’ attention.
Example 3.1. Nisbet and Vakil [6] proposed two demand curves for marijuana
among UCLA students (quantities in "lids" per month, prices in dollars per "lid"; one
"lid" equals one dried ounce)"
(A) q D1 (p) 15.4 p-l.O3, i.e. p 14.8 q-0.987,
(B) q D:(p):= -0.225 .p + 3.74, i.e.p -4.44 q + 16.6.

S/ounce
20

15 tangent line
D

i0
(q*,p*) ’\, (qo,Po)

ounces/month
1
0.i 0.2 0.3 .4 .5 .6 .8 1.0 2.0 3.0 4. 5. 6. 8. i0

FIG. 2. Dz(p) -0.225 p + 3.74, on logarithmic scales. (q*,p*) (1.49, 10) shows the prevailing
market situation, while (qo, Po) (1.87, 8.30) indicates the optimal point, where the tangent line to D2 (drawn
in) has slope 1.
264 CLASSROOM NOTES

D has constant elasticity -1.013 and therefore would simply appear as a straight line
with slope -1.013 on logarithmic scales. At the going price of $10 per ounce, D has
Downloaded 03/26/14 to 140.120.135.222. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php

elasticity 1.51, the slope of the tangent to p o D_ o 4 at (q*, p*) (1.49, 10). Sliding a
straight edge with slope until it is tangent to the curve shows that D has elasticity
at (qo, P0) (1.87, 8.30), where revenues would be maximal (see Fig. 2).
4. Calculus with elasticities. The elasticities rir and rig of tWO given positive functions
F, G" *+ *+ follow rules similar to those of elementary calculus:
(a) riFG OF + riG,
(b) F/G riF riG,

(d) ri a- riF,
(e) riFoG (riF G) (chain rule),
rig

where a and X are constants. Let us prove (a) and (e) as examples:
X X
(a) riFG (F. G)’- (F’G + F. G’)= riF + riG,
F.G F.G
X X
(e) riFoG (F o G )’ (F’ o G) G’
FoG FoG
G
.(F, oG).X .G’ (riF o G) riG.
Fo G
The proof of (b) is similar to that of (a), while (c) and (d) constitute particular cases of
(e).
A function F can also be recovered from its elasticity as follows: since riF
(x/F) dF/dx then dF/F (riF/X) dx, and an integration yields
In IF(x) f,x), dx + In
J X
(9)
F(x) K. exp .frir(X)x dx]
for some constant K. Here is an application of this formula.
Example 4.1. Cullis, Forster and Frost [2] studied the demand for inpatient
treatment, whose elasticity can be written rio(S)=-2.19. S, where S denotes the
number of available hospital beds, and D that of deaths and discharges per year. Equation
(9) then yields
S
(10) D(S) K.exp [f-2.19. S
dS K e -2"’9"s

(K was estimated to be about 5.17.)


5. Analogy with an elastic rubber band. Perhaps the reader wonders whether there is
a connection between economic elasticity and physical elasticity. It turns out that there is
a formal similarity between the price elasticity of demand, written no (d Log (D))/
(d Log (p)), and the tensor a that describes the stress undergone by an isotropic and
homogeneous elastic rubber band of volume V subject to a deformation gradient tensor X:
k. T dLog(Z)
()
V d Log (X)
CLASSROOM NOTES 265

where Z represents the partition function and q -kT. Log (Z) the Helmholtz free
energy. (See [5, p. 62] and [7, p. 64].) Of course, the present paper does not tell how to
Downloaded 03/26/14 to 140.120.135.222. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php

stretch a dollar bill.

REFERENCES
[1] A. BELINFANTE AND R. R. DAVIS, JR., Estimating the demand for record albums, Rev. Business and
Economic Research, XIV/2 (winter 1978-1979), pp. 47-53.
[2] J. G. CULLIS, D. P. FORSTER AND C. E. B. FROST, The demand for inpatient treatment: some recent
evidence. Appl. Economics, 12 (1980), pp. 43-60.
[3] T. F. HOGARTV AND K. G. ELZ|NGA, The demand for beer, Rev. Economics and Statistics, LIV/2 (1972),
pp. 195-198.
[4] E. MANSFIELD, Microeconomics, Theory and Applications, 3rd. ed., W. W. Norton, New York, 1979.
[5] F.D. MURNAGHAr,Finite Deformation of an Elastic Solid, John Wiley, New York, 1951.
[6] C.T. NISBET AND F. VAKIL, Some estimates ofprice and expenditure elasticities of demandfor marijuana
among U.C.L.A. students, Rev. Economics and Statistics, LIV/4 (1972), pp. 473-475.
[7] C.J. THOMPSON, Mathematical Statistical Mechanics, Macmillan, New York, 1972.

You might also like