Transcript 投影片 1

•
•
•
•
Chapter 15 Market Demand
Two goods
Individual demand for good 1: xi(p1,p2,mi)
Market demand x(p1,p2,m1,m2,…, mn)=
x1(p1,p2,m1)+ x2(p1,p2,m2)+…+ xn(p1,p2,mn)
• Market demand depends on the
distribution of income.
• Earlier we talked about inverse demand:
given a quantity, the demand gives us the
MRS for the consumer to demand such.
• For market demand, we can interpret the
inverse demand the same way. Since at
optimum, everyone chooses so that MRS
equals the relative price ratio and the
price ratio is common. Hence, given a
quantity (quantity demanded by all), the
demand gives us the MRS for all.
• One way to characterize the demand is to
describe how responsive it is. For
instance, how responsive demand is to
price changes, income changes or price
changes of other goods.
• Price elasticity: x(p1,p2,m), ∆x/∆p1 is not a
good measure because it is not unit free.
We use (∆x/x)/(∆p1/p1) instead.
• It is negative for most demands, but
sometimes we ignore the negative signs.
• ||>1: elastic, ||=1: unit elastic, ||<1:
inelastic
• Linear demand x=a-bp, a,b>0, then ||=|bp/(a-bp)|
• At the midpoint, ||=bp/(a-bp) where
p=a/2b so ||=1. When p is higher than
the midpoint, ||>1. When p is lower than
the midpoint, ||<1.
• Can calculate the relationship between
revenue and elasticity. Revenue R=px, so
marginal revenue (the change of revenue
when the quantity is increased)
MR=dR/dx=p+(dp/dx)x=p(1+(dp/p)/(dx/
x))=p(1+1/), when negatively sloping,
this equals p(1-1/||).
• MR=p+(dp/dx)x, when you increase the
quantity, for the marginal unit, you get an
additional revenue p, but since there is
only one price, all the units sold before
have to be sold at the same price. Thus
the revenue is changed by the amount as
well (dp/dx)x.
• Notice that MR at x=0 equals p (no units
sold before) and MR at x>0 typically is
less than p when downward sloping.
• MR=p(1-1/||). So when ||=1, MR=0.
When ||>1, MR>0. When ||<1, MR<0.
The intuition is, when demand is elastic, a
price decrease makes quantity increase a
lot. So it is worthwhile to decrease the
price or increase the quantity. Hence
MR>0. On the other hand, when the
demand is inelastic, a price decrease can
hardly change quantity. Thus not
worthwhile to decrease the price or
increase the quantity, so MR<0.
• For linear demand p=(a-x)/b,
MR=p+(dp/dx)x=p-x/b=(a-x)/b-x/b=(a2x)/b. So the slope is twice negative.
• Simple observation, a firm will not
choose a price at which the demand is
inelastic. By decreasing the quantity
(increasing price), since MR<0,
decreasing the quantity increases R.
• Decreasing the quantity decreases cost
typically. R is increased, C is decreased,
so total profit is increased.
• Convenient to regress ln(x) on ln(p)
because if ln(x)=c+fln(p), then
dln(x)/dln(p)=(dx/x)/(dp/p)=f.
• Unit elastic demand has MR=0 or
R=constant. So any rectangular under the
demand curve has the same area.
• Can also work conveniently in this way.
To generate a demand of constant
elasticity, we want ln(x)=k+fln(p) where
k,f are constants. Let ek=c. Hence
ln(x)=ln(cpf) or xp-f=c. For instance, if
unit elastic, then xp-(-1)=c or xp=c. Work
out the examples of f=-2, f=-1/2 yourself.
• Income elasticity can be similarly defined
as (∆x/x)/(∆m/m). Income elasticity>0:
normal, <0: inferior, >1: luxury.
• Income elasticities tend to cluster around
1 because budget constraint is binding.
For instance:
p1x1 + p2 x2=m and
p1x1’+ p2 x2’=m’. Let ∆ denote the
difference. Then p1∆ x1+p2∆x2=∆m or
p1∆ x1/∆m+ p2∆x2/∆m=1 or
[(p1 x1)/m](∆ x1/x1)/(∆m/m)+ [(p2 x2)/m]
(∆x2 /x2)/(∆m/m)=1. The weighted
average (expenditure share) of the
income elasticities is 1.