Professional Documents
Culture Documents
Mathematic of Growth Rates
Mathematic of Growth Rates
Michael Bar
July 9, 2018
Contents
1 Measuring Rates of change. 1
1.1 Discrete time variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Continuous time variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Logarithmic scale 4
3 Important Approximations 5
3.1 Rate of change of a product and ratio . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Rule of 70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.3 Rates of Change of sum and di¤erence . . . . . . . . . . . . . . . . . . . . . 7
De…nition 1 Time series - the values of a variable recorded at di¤erent points in time
constitutes a time series.
Time series is collected by a number of di¤erent agencies in the economy. For example
the Bureau of Economic Analysis (BEA) collects data about National Income and Product.
Federal Reserve System collects data on Monetary aggregates and Interest rates. The Bureau
of Labor Statistics collects data on Employment and wages. Data are also measured in
di¤erent time intervals, so we have annual data , which is recorded once a year, quarterly
data recorded four times a year. We also have data recorded every minute such as stock
prices on the NYSE.
1
The GDP is continuous by nature since every instant something is being produced. However,
we are not able to measure the GDP in the economy at every instant. The BEA can estimate
the output only after it was sold to the buyers. Therefore, we are going to treat economic
variables such as GDP as discrete variables.
Example 1 Suppose the price of a good was $76 at 2000 and $87 at 2001. What is the rate
of change in the price from 2000 to 2001?
Solution:
pt+1 pt 87 76
= = 0:14474 = 14:474%
pt 76
Consider the special case when the rate of growth is a constant over time, say g. That is
yt+1 yt
yt
= g for all values of t. This implies that yt+1 = (1 + g)yt . Hence, we can express the
value of y at time t in the following way
yt = (1 + g)t y0 (2)
Example 2 Suppose that you invest $1000 in a trust fund that promises 5% annual interest
rate. How much money will you have in the fund after 10 years?
Solution:
Example 3 US GDP per capita grows at constant rate of 2% per year. After how many
years will it double?
Solution:
2y0 = (1 + 0:02)t y0
2 = (1:02)t
ln(2) = t ln(1:02)
ln(2)
t = t 35
ln(1:02)
2
Example 4 Korea grows at 4%. How long would it take for Korea to catch up with the US
if the US GDP grows at a constant rate of 2% per annum and the Korean GDP is just half
the size of US GDP?
Solution:
d ln(y(t))
(3)
dt
To show why this gives the rate of change, use the chain rule to get
d ln(y(t)) 1 dy(t)
=
dt y(t) dt
3
2 Logarithmic scale
Suppose that a variable y grows at constant rate g and has initial value of y0 . Then the
value of y at time t is given by
yt = y0 (1 + g)t
Now, if we take the natural logarithm of yt , we get that ln (yt ) is a linear function of time:
This is a linear function of time, with slope of ln(1 + g) and intercept ln(y0 ). Now we can
show that the slope of this function is approximately equal to the growth rate, for small g,
i.e. ln(1 + g) t g
Proof. We need to prove that
ln (1 + g)
lim =1
g!0 g
Notice that when g ! 0, both the numerator and the denominator in the limit go to zero.
In other words, we have a limit of the form of 00 . Using L’Hopital’s rule we get
ln (1 + g) 1= (1 + g)
lim = lim =1
g!0 g g!0 1
Example. Suppose that you deposit $1000 in a savings account, with interest rate
of 3%. The amount of money you have in the in the savings account at any time t is
st = 1000 (1 + 0:03)t , and is shown in the next graph.
1900
1800
1700
1600
S_t
1500
1400
1300
1200
1100
1000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
t
Now the ln of savings is ln (st ) = ln (1000) + t ln (1 + 0:03) and shown in the next graph.
4
ln(savings)
7.6
7.5
7.4
7.3
ln(S_t)
7.2
7.1
6.9
6.8
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
t
Notice that ln of savings is a linear function of time. Also observe that the slope of ln (st ) is
approximately equal to
7:5 6:9 0:6
= = 3%
20 20
The properties of logarithmic scale are very useful every time we look at data that is
growing over time. By looking at the original data we cannot tell whether it is growing
at constant rate or not. But when we plot the ln of the variable, we can see right away if
the variable is growing at constant rate or not. That is, if the ln of the variable looks like
linear function of time, we conclude that the original variable is growing at constant rate.
Moreover, we can immediately compute the approximate growth rate of the original variable
from the slope of the ln, as shown in the previous example.
3 Important Approximations
In this section we prove important approximation rule, for small growth rates.1
2. The growth rate of the ratio is approximately the di¤erence of the growth rates
d
x
x^ y^
y
1
The approximations are in general good for growth rates g 10%.
5
Proof. 1. From the de…nition of growth rate, we have
xt+1 yt+1
cy =
1+x
xt yt
Taking ln of both sides
cy = x^ + y^
x
d
x
= x^ y^
y
cy
x x^ + y^
2. Omitted.
3.2 Rule of 70
Proposition 1 Suppose that a variable is growing at constant rate g per period. Then, for
small g, the approximate doubling time is
70
t
100 g
6
Proof. Let the initial value be y0 . Then after t periods we have:
y0 (1 + g)t = 2y0
(1 + g)t = 2
Notice that y0 cancels out, so the doubling time is independent of the initial value. Taking
logs and solving for t:
t ln (1 + g) = ln 2
ln 2 0:69315 70
t =
ln (1 + g) g 100 g
xt+1 yt+1 xt + yt
x[y =
xt yt
xt+1 xt xt yt+1 yt yt
=
xt yt xt xt yt yt
xt+1 xt xt yt+1 yt yt
=
xt xt yt yt xt yt
xt yt
= x^ y^
xt yt xt yt
Example 5 Suppose that during the last year, the price of a product increased by 2% and
the quantity sold increases by 1:5%. What is the approximate growth rate of the revenue?
Solution:
P[ Q P^ + Q^ = 2% + 1:5% = 3:5%
Remark: If we wanted the exact growth rate of the revenue, then letting the initial values of
price and quantity be P0 and Q0 we have
new price new quantity
z }| { z }| {
(1 + 0:02) P (1 + 0:015) Q
1 + P[Q= = (1 + 0:02) (1 + 0:015) = 1:0353
P Q
Thus, the exact growth rate is 3:53%, and the approximate growth rate is 3:5%.
7
Example 6 Suppose that during the last year, the real GDP is grew at 2:5% and population
grew at 1%. What is the approximate growth rate of GDP per capita?
Solution:
d
Y
Y^ N ^ = 2:5% 1% = 1:5%
N
Remark: if we wanted the exact growth rate, then letting the initial levels of GDP and
population be Y0 and N0 , we have
Y Y (1 + 0:025) Y
= = 1:014851485
N new N (1 + 0:01) N
Thus, the exact growth rate is 1:485%, and the approximate growth rate is 1:5%.