Micro 091021

Document Sample
Micro 091021 Powered By Docstoc
					     Micro 1124

Pricing with market power
                   Pricing with market power
• In monopoly chapter we assumed that each monopolist
  seller only sold one homogenous good and had to sell all
  units at the same price
• In this chapter text considers sellers with market power
  for which these assumptions are not true
   – market power: the seller faces a demand curve for its own
     products that is not perfectly elastic
• First few pages deal with price discrimination: charging
  different prices for different units sold
   – different units can be sold to the same buyer, or to different
   – note a key assumption for price discrimination: there must be
     something that prevents buyers from reselling the units that they
     have bought, to other buyers
                First-degree price discrimination
• This is the reference case in which every unit sold is sold
  at the highest price that any buyer is willing to pay for it
   – first-degree price discrimination is sometimes also referred to as
     perfect price discrimination
• As figure 11.2 shows, this implies that there will then be
  no consumer surplus left in the market for this good
• Note also that with first-degree price discrimination, the
  profit-maximizing quantity for the monopolist is where the
  MC intersects the demand curve (why?)
   – is this solution still consistent with the general rule: produce and
     sell until the marginal revenue equals marginal cost? (Ans: yes)
• What happened here to the efficiency loss from
   – Ans: the equilibrium is efficient (why?)
               First-degree price discrimination
• Although it may not be possible in practice, one can
  interpret it this way: Suppose the seller knows each
  individual’s demand curve, and hence can compute the
  individual’s consumer surplus at each price
• The seller can then offer each individual the following: if
  you pay me an amount that is one cent less than your
  consumer surplus, I will allow you to buy as many units
  as you like at a price that is equal to my MC
• If you do not accept this offer, I will not sell you any units
   – remember, we have ruled out the possibility that the consumer
     could by any units from someone else who had bought it from
     the monopolist
• A self-interested consumer would accept this offer. If all
  consumers do, the number of units sold will be where the
  market demand curve intersects the MC, and there is no
  consumer surplus left
                   Price discrimination, cont’d
• Text recognizes that perfect price discrimination is not
  possible in practice but cites the examples of doctors,
  lawyers, accountants, as sellers that try to use it in
  practice, at least to some extent
• Second-degree price discrimination: when the price per
  unit depends on how many units the consumer buys
   – note again that in order for this to work, it must be impossible (or
     at least costly) for consumers to resell units that they have
     bought, to others
• An example of second-degree price discrimination is
  block pricing: for some X, the first X units bought per
  period are priced at Px per unit; if more than X units are
  bought, the price per unit falls to Py<Px per unit
   – this pricing scheme is applied to each individual buyer
                   Price discrimination, cont’d
• An example of block pricing is the pricing of electricity to
  households in some communities
   – note that there can be more than two prices
• Block pricing may be particularly useful in cases of
  natural monopoly
   – remember, natural monopoly is when economies of scale are so
     large that it is only efficient for a single firm to serve the whole
   – it can be used together with monopoly regulation to attain an
     efficient level of production, although it generally is not privately
     profitable to produce at that level if only two prices are allowed
• Third-degree price discrimination is the most common
  form. This refers to cases where various methods are
  used to charge different buyers different prices
               Third-degree price discrimination
• As before, it is critical that resale of the seller’s goods or
  services are not possible
   – clearest cases of this practice would be student or seniors’
     discounts for certain goods or services
   – other cases could be charging different prices for the same good
     in different outlets
• The key element in third-degree price discrimination is
  that there must exist buyer groups with different price
  elasticities of demand
   – if such groups exist and the seller can charge different prices to
     members of each group, it will generally be profitable to do so
• Suppose there are two buyer groups indexed by 1 and 2
  that can be charged different prices. The monopolist’s
  profits will then be
  Third-degree price discrimination with two buyer groups
• Using calculus (on the board) and simplifying, one can
  derive the following rule for optimal pricing:
      MC (Q1  Q2 )  P (1  11 )  P2 (1   2 1 )

• Clearly, prices in the two markets will only by different if
  the demand elasticities η(1) and η(2) are different
• For an illustration, see Fig 11.5 in the text (board)
• The text now gives several examples of third-degree price
  discrimination in which the seller tries to separate out
  buyer groups with different elasticities by making the
  product slightly different, or at least make it seem as if they
  are slightly different
        Examples of third-degree price discrimination
• One amusing example is the one referring to the
  producer of vodka who sells the same product at
  different prices, only in different bottles
   – note that the names cited are clearly designed to appeal to
     different buyer groups
   – I sort of suspect this method may be used by sellers of mou tai
• Other interesting examples are airline travel (business
  class vs economy), and the practice of giving discounts
  to people who present coupons from the newspapers, as
  well as “mail-in rebates”
   – note the empirical evidence on different price elasticities for
     coupon users and non-users in Table 11.1
   – Note also that the different price elasticities for airline travel are
     likely to be due in part to the fact that in many cases, business
     travel is paid by someone else, not the traveller
              Intertemporal price discrimination
• The example of hard-back and paper-back editions of
  books is interesting (Fig 11.7)
   – Note that in this case, the seller is a monopolist because of
     copy-right legislation
   – Separation of markets in this case is accomplished by releasing
     the hardback version of the book many months before the paper-
   – People who are eager to buy the book right away probably have
     a less elastic demand, so they will buy the hardback edition even
     if it is expensive; if the paperback edition were available right
     away, many of those who buy the hardback edition would
     probably buy the cheaper paperback edition instead
• Another example is peak-load pricing
   – it is practiced in situations where the demand for some good or
     service fluctuates over time
                        Peak-load pricing
• As the text notes (around Fig 11.8), peak-load pricing is
  both more profitable for the seller than charging the
  same price at all times, and more efficient for society
  (because the marginal cost is lower at off-peak)
• Price discrimination through peak-load pricing is
  therefore quite different from regular third-degree price
  discrimination, because in the regular case the MC is the
  same (or essentially the same) for each buyer group
   – consider, for example, the book example, or the air travel
• Peak-load pricing, in contrast, refers to cases where the
  MC depends on the number of people consuming the
  good or service (it usually occurs for services) at a
  particular time
              Peak-load pricing, two-part tariffs
• Examples of peak-load pricing are ski-lifts, amusement
  parks, movie theaters, etc., where prices usually are
  lower on weekdays when demand is low
• Other examples of peak-load pricing can be electricity in
  some countries, or public transportation
   – note that, contrary to regular goods and services, in cases where
     demand fluctuates over time in a way that affects MC, it is not
     efficient for all units to be sold at the same price
   – so peak-load pricing is sometimes practiced by regulated
     monopolies as well, in order to attain an efficient pattern of
     production and use
• Section 11.4 in the text talks about pricing through a two-
  part tariff. It is commonly used by sellers such as
  amusement parks, golf clubs, gymnasiums, book clubs,
                          Two-part tariffs
• A two-part tariff requires consumers to pay some type of
  membership fee to the seller, before they are allowed to
  buy any units. Once they have paid the one-time fee,
  they can buy as many units as they like at a fixed price
• Notice again that it can only be used in situations where
  it is possible to prevent resale
• As the text notes (Fig 11.9), if all consumers are
  identical, a two-part tariff can be used to extract all
  consumer surplus
   – for a seller who can do this, the profit-maximizing price per unit
     for members is the marginal cost
   – note again that with a two-part tariff, the amount produced of the
     good/service is efficient from society’s point of view
• However, if consumers are not identical, the solution is
  more complicated (Figs 11.10, 11.11)
                       Two-part tariff
• If consumers are not identical, the seller may still have to
  set the same entrance fee for all buyers, and charge the
  same price per unit
• Note that the case when the seller can charge different
  entrance fees to all buyers, and knows what the demand
  curves are, we are back in the case of first-degree price
  discrimination (explain why)
• Also note that the two-part tariff is related to the idea of
  block pricing (why?)
• Pay particular attention to the case of a constant MC and
  two linear demand curves in Fig 11.10, and do exercise
  10 at the end of the chapter (next slide)
• In this example, the seller must charge the same price
  per unit of use, and the same membership fee, to all
                 Two-part tariff exercise
• The demand curves for the two types of buyers are
  Q(1)=10 – P, Q(2)= 4 – 0.25 P. It is assumed that MC=0,
  but that the seller has fixed costs of 10,000 per week.
  There are 1000 buyers of each type, and the demand
  functions refer to units bought per week
• You are first asked what your maximum profit would be if
  you only cater to type 1 buyers and set your membership
  fee and unit price at their optimum levels.
• Second, you are asked to show that you can make more
  money by setting a common two-part tariff that is
  attractive enough so that both groups will join
• Third, you are asked how your solution would change if
  there were 2000 additional potential buyers of type 1
              Two-part tariff, concluded; bundling
•   Look also at table 11.3 with pricing for mobile phone
    plans, which represent a menu of two-part tariffs
    – can you explain why the phone companies think they will make
      more money by offering all these different plans, rather than a
      single plan?
• Note, though, that this case may not, strictly speaking,
  be a good example of sellers with considerable market
  power (why not?)
• The concept of bundling refers to the idea of selling
  combinations of things at a package price
• The discussion in the text explains how bundling is a
  potential way of the seller extracting more of the
  consumer surplus when goods are bundled in such a
  way that consumer’s valuation of the bundle components
  are negatively correlated
• In the diagram on next slide, if prices of the two goods
  are sold separately at P0 and P1, buyers in quadrant II
  will buy both goods (since their reservation prices are
  above the prices charged); those in quadrant IV will buy
  neither good, while those in quadrants I and III will buy
  one good but not the other
• Suppose P0=P1, and now suppose the valuations are
  negatively correlated to the extent that there is no buyer
  in quadrant II, but many buyers in quadrants I and III,
  and some in quadrant IV (disregard the B-line for now)
• With this assumption, the seller earns P0=P1 from each
  buyer in quadrants I and III
• Now suppose the seller bundles these two goods so that
  if you buy one, you must also buy the other
                   The axes measure the maximum amount that a
                   consumer is willing to pay for goods 1 and 2. P1 and
     r2            P2 are hypothetical prices charged for the goods

               I                            II




          IV         B

                               P1                                r1
• Just to make the point dramatically, suppose the price of
  the bundle is PB=P0=P1
   – that is, two for the price of 1!
• But this means that any buyer whose valuations of the
  two goods is at a point above and to the right of the
  dotted line B will buy both goods
• In this case, this also means that bundling pays: the
  seller will earn the same revenue from all buyers in
  quadrants I and II as before, and will now earn additional
  revenue from those above the dotted line B in quadrant
   – note: if there were buyers in quadrant I who bought both goods
     before, there would now be much less revenue from them
• So in this extreme case of negative correlation bundling
  will certainly raise total revenue
   – probably total revenue will be higher if the bundle price is
     somewhat higher than P0=P1
• The text also distinguishes between “pure bundling” and
  “mixed bundling”, where the former means that the
  consumers’ only option is to buy the goods together,
  while the latter means that the consumer has a choice
  between either separate prices or a bundle (where the
  bundle price is lower than the sum of the separate
• When consumers’ valuations are not perfectly
  (negatively) correlated, mixed bundling may be the
  preferred strategy
                         Mixed bundling
• In Fig 11.12, there are three buyers of two goods, with
  reservation prices given by (good 1 first): A {3.25, 6},
   B {8.25, 3.25}, C {10,10}
• In exercise 11 you are asked which strategy is best,
  separate pricing, pure bundling, or mixed bundling. (The
  assumption here continues to be that MC of either good
  is zero)
   – let’s show on the board that mixed bundling is the best option
• Also look at the discussion of “bundling in practice”
• One set of examples is cars with various luxury
  “options”, which are often offered as packages
   – some luxury cars use pure bundling
• Another interesting example is restaurant menus which
  offer things “a la carte” or “complete dinners”
               Bundling, concluded; Advertising
• The logic of this last example is similar to that in the
  earlier diagram
   – a Western meal often has an “appetizer” (small dish before the
     main course), the main course, and a “dessert”
   – a bundle captures both those who really like an appetizer but
     don’t care for dessert, and those whose tastes to the other way,
     as well as those who like both appetizer and dessert a little bit
   – it is in this sense that the bundling can capture more of the
     consumer surplus than pricing and selling each component
• The last main topic is advertising: spending money
  reminding consumers about your excellent product
   – Note that sellers in perfectly competitive markets do not
     advertise; why not?
• Empirical evidence shows that advertising does increase
  sales for certain kinds of products. So the question is:
  how should a seller jointly determine production, pricing,
  and advertising?
• The answer must depend on the extent to which
  advertising expenditure increases consumer demand for
  the product
• The text frames the problem this way: when the firm
  spends A CNY on advertising, the demand function for
  its product is Q(P,A), with the partial derivative of Q with
  respect to A being positive
• The firm’s profit function will then be written:
   π = P*Q(P,A) – TC(Q) - A
To find the optimum expenditure on advertising and
the optimum price and production, one finds the two
first-order conditions:

            Q              Q
       QP        TC ' (Q )    0
    P        P              P
         Q              Q
       P       TC ' (Q )     1  0
    A    AP              A

These equations can be rearranged and combined
to yield the “rule of thumb” for advertising:
                 A   A
                PQ    P
                     Advertising, concluded
• This rule gives the optimum advertising budget A as a
  fraction of total sales revenue PQ; the ratio should equal
  the ratio of the elasticity of produce demand with respect
  to advertising ( ηA) to the (absolute value of) the
  product’s price elasticity of demand (- ηP)
   – to interpret this rule of thumb, remember that the ratio (P-MC)/P
     equals (1/(- ηP)). That is, for a given ηA, more money should be
     spent on advertising when the markup of price over MC is large
   – this makes sense intuitively
• If there is time we will discuss the Appendix on transfer
  pricing as well

Shared By: