UCLA School of Public Health Community Health Sciences DepartmentCommunity Health Sciences Department
 

CHS 247. Population Change and Public Policy

WHAT EVERYONE SHOULD KNOW ABOUT e AND LOGARITHMS

 by M.A. Strong (University of Pennsylvania)

 Where does “e” come from?

Say that $1 is put into a bank that gives 100% interest.  If interest is compounded once per year, the value (V) of our investment becomes: 

V(1) = initial principal x (1 + interest rate)

                                        = $1 x (1  + 1) = $2

(where the number in parentheses following V indicates the frequency of compounding and “x” means multiplied by).

If interest is compounded twice a year, an interest rate of  50% (i.e., 100/2) or .50
per $1  will be used for each 6-month period.  Thus at the end of the first 6 months we would have:

$1 x (1+ 0.50) = $1.50

During the next 6 months is $1.50 would grow to:

V(2)     =          $1.50  x  (1+ 0.50)       =          $2.25

           =          $1  x  (1 + 0.50)  (1 + 0.50) 

           =          $1  x  (1+0.50) 2

Similarly, if interest is compounded quarterly, the quarterly interest rate would be 25% (or .25 per dollar).

End of 1st quarter: $1 x (1 + 0.25) = $1.25
2nd quarter:   $1.25  x  (1+0.25) =  [$1 x (1+0.25)] (1+0.25) =$1.5
3rd quarter: $1.56 x   (1+0.25) = {[$1 x (1+0.25)] (1+0.25)} = $1.95
4th quarter:  V(4) = $1.95 x  (1+0.25)  =  $1 x (1+0.25) (1+0.25)  (1+0.25)(1+0.25)
         = $1 x (1+0.25) 4  = $2.44

 

Thus:   V(1) = $1 x (1 + (1.0 /1) 1 = $2.00
  V(2) = $1 x (1 + (1.0 /2) 2 = $2.25
  V(4) = $1 x (1 + (1.0 /4) 4 = $2.44              

    

In general:  V(m) = $1 x (1 + (1.0 /m)m
  where m represents the frequency of compounding in 1 year.

The limiting case is where interest is compounded continuously, i.e., where m becomes infinite. 

This gives us our definition of e:

limit  V (m)   =   limit   [$1  X  (1 + (1/m)m]   =   e (dollars)
m + ¥                m + ¥

As you can see in the following diagram, the value of our $1, as it is compounded more and more frequently (i.e., as m gets larger) does not get infinitely large, but approaches a limit.

Generalizations:

1)  More years:  if $1 becomes $e after 1 year of continuous compounding, and we leave this $e in the bank a second year, each dollar in $e will grow to e and the value of our savings account will be:
 

$e x e = $e2

which would become in three years:  $e3

or, more generally, after t years, $et

2)  Different principal (i.e., initial investment amount):  since each $1 grows to $et  in t years, if we invested $A to start with, our final value will be A times as large or  $Aet.

3)  Different interest rate:  in the initial formula, replace 100% (or 1.00 per dollar) with r:
 
V(m) = $1 x (1 + (r/m))m = [(1 + (r/m) (m/r)] r    
                                   = [(1 + (1/z)z]r   if we let m/r = z and r/m = 1/z
   
V(m) = [(1 + (1/z))z]r  Since m ® ¥  made (1 + (1/m))m = e, 
  and if  m ® ¥ , z = (m/r) ® ¥ too,
       = [e]r     \ (1 + (1/z))z = e.

Applying all three generalizations:

V(m) = A (1 + (r/m))mt  $A rather than $1, is the principal.
  r/m means that in each of m periods only
    1/mth of the interest rate is applied.
  mt = compoundings per year x the number
  of years, i.e., the total no. of compounding
   
       = A [(1 + (r/m))(m/r)] rt     if we let m/r = w, then r/m = 1/w as above
       = A [(1 + (1/w))w]rt  
   
       = Aert    thus m ® ¥ implies m/r=w ® ¥,
  and (1 + (1/w))w = e

This analysis can be applied to any growth process where growth takes place continuously at a constant rate r, including population growth.  Note that r, previously the interest rate, can no be reinterpreted as the instantaneous rate of growth of the function Aert.


LOGARITHMS

A logarithm is the power to which a base must be raised to attain a particular number:

f the base is 4, then since 42  = 16

                                    log4 16 = 2

                                       ­    ­     ­

                                  base    no.    base to this power = no.

“common” logs use 10 for the base

log10 1,000  = 3   (since 103 = 1000)

log10    100  = 2   (since 102 = 100)

log10      10  = 1   (since 101 = 10)

log10        1  = 0   (since 100 = 0)

log10        0.1  = -1   (since 10-1 =  (1/10)1 = (1/10) = 0.1

log10        0.01  = -2   (since 10-2 =  (1/10)2 = (1/100) = 0.01

   etc.

 In general,  y = bt  « t = log b  y

In analytic work, natural logs, using e as the base, are more convenient to use.  Note that loge is frequently written ln.

Since    y = bt + t = log b y

            y = e3 + 3 + ln y = ln e3

            y = en + n + ln y = ln en

Example:  What will the rate of growth have to be for a population P to double in 20 years (i.e., what growth rate is required if population 20 years from now is two time the current population –2*P)?

2P = P ert

 2  = e 20r

            ln 2 = ln e20r

            ln 2 = 20r

       0.6931 = 20r

r = (0.6931/20) = 0.035  or a growth rate of (0.035 * 100) = 3.5 percent

To calculate doubling time, if you know the growth rate:

  t= 69.3/r   where t is the doubling time and r is the growth rate express as a %

OR

t= 0.693/r  where r is the growth rate expressed per person (NOT per cent)