Costs and Prices in a Modern Telecommunications Market

Download Report

Transcript Costs and Prices in a Modern Telecommunications Market

Costs and Prices in a Modern
Telecommunications Market
D. Mark Kennet, Ph.D.
for
Public Services Regulatory Commission of Armenia
Network Industries
• Distinguish between industries that contain network
infrastructure and economic networks
• All public service industries have network
infrastructure
–
–
–
–
Electricity
Water
Gas
Telecom
• But of these, only telecom is a true network industry
What is a network industry?
• An economic network is an industry in which the
addition of a client brings both private (to the client)
and public (to the other clients) benefits
• Thus, demand for the service provided by the
network depends on the number of other clients as
well as the intrinsic value of the service
• A telephone network of one client is useless; the
value of the network increases to existing clients
whenever a new client joins
The presence of these network effects
influences public policy
• Two major policy implications:
– Universal service
– Interconnection requirement
• Universal service is justified by the public benefit of
increasing the client base
• Interconnection is essentially the same argument –
interconnecting networks provide their respective
clients a much larger network than they had without
interconnecting
Regulatory policy must take network
effects into account
• Traditional regulation attempts to assign costs to
services and then set tariffs to costs, which include a
“reasonable” return
• This approach may or may not be appropriate in the
presence of network effects together with
competition, potential or real
• The structure of prices may be as important as the
level; that is, a wider variety of pricing plans may
prove necessary than a simple two-part tariff (rental
plus usage)
Relevant characteristics of telecom
network costs
• High proportion of fixed costs
– In fact, the network is essentially 100% fixed costs relative
to usage
• Assignment of costs to subscriber lines and usage is
essentially arbitrary and depends on accounting rules
• An economically ‘efficient’ tariff would be a simple
flat monthly fee
• Demand considerations generally make the efficient
tariff unsuitable, at least as a single-price option
Cost concepts
•
•
•
•
•
•
•
Total cost
Fixed cost
Variable cost
Average cost
Marginal cost
Incremental cost
Standalone cost
Total cost, fixed cost, variable cost
• Total cost is simply the total cost of the
telephone network capable of providing the
services of interest
• Fixed cost is that cost that does not vary with
the amount of service
• Variable cost is that part of cost that varies
with output
• Mathematically: TC = FC + VC
Average cost, marginal cost,
incremental cost
• Average cost (AC) is just total cost divided by output
– Not particularly helpful in the case of telecom, since telecoms produce
multiple products
– May serve as a reference point for some analyses
• Marginal cost (MC) is the cost of producing an additional unit
of output (e.g., the cost of an additional minute of use, line,
etc.)
– Many telecom situations where MC = 0
• Incremental cost (IC) is very similar to MC, except that it is
often used to describe a change in the number of services
offered, as in Total Service Long Run Incremental Cost
(TSLRIC):
TSLRIC(i) = (TC(I,i) – TC(I,0)/i
Standalone cost
• Standalone cost is that cost which would be incurred
if only the product (or group of products) of interest
were being produced (without the others)
• This concept is mostly used as a reference point: If a
price charged for a service is greater than its
standalone cost, then entry into that market is likely
• Example: international long distance services, which
is why most countries have liberalized this market
Evaluating cost concepts in
telecommunications
• Fixed costs: A very high percentage of total
costs are fixed in telecom
• Variable costs: There is no variable cost for
usage (outside the peak hour), and only a
small variable cost for a new access line
• Marginal costs: Similar to variable costs
• Incremental costs: Significant, but still very
small relative to total costs
Aside on traffic-sensitive (TS) and nontraffic-sensitive (NTS) costs
• An approximation to measuring long-run incremental cost of
traffic is to separate costs into TS and NTS
• TS costs could be correctly defined as those costs that change
as peak busy-hour volume changes (capacity costs)
• NTS costs could be correctly defined as anything else
• Unfortunately, standard practice does not accord with this
common-sense approach:
– NTS costs are defined as those costs that can be directly attributed to
a subscriber line
– TS costs are anything else
– This approach represents an effort to load all costs to usage, which
may be politically popular but does not accord with economic
efficiency
Comparing telecom costs to those of
other infrastructure industries
• All infrastructure (“network”) industries have
high fixed costs
• But only telecom has zero variable costs for
many of its outputs
• Other infrastructure industries have significant
costs that vary with usage – e.g., electricity,
gas, water, railroads
Tariffs and prices
• Economics teaches us that in a competitive market, price of a
good or service will be driven to marginal (or incremental)
cost
• The opposite of pure competition is pure monopoly. In this
case, price is determined by the demand for the good or
service, and output is determined at the profit maximizing
point for the monopolist
• Monopoly prices are always higher, and outputs always lower,
than the competitive outcome
• This result, for many people, justifies the regulation of
monopoly: Make prices lower, and output higher
Problems with this approach
• Regulation is costly
• The efficient price is when price is set to MC,
but in telecom, MC is ALWAYS less than AC, so
firm will not make money
• There is no theory that tells us the best way to
make up the difference between MC and AC
• In telecom, the notion that there is even a
monopoly is doubtful
Is telecom a monopoly?
• If we consider only fixed voice services, the answer is
yes
• However, studies in some countries suggest that in
some populations, wireless service is substituting for
fixed service, which eliminates the notion of fixed as
a monopoly
• The monopoly power that does exist is limited and is
a result of interconnection policies and control of
telephone numbers
Interconnection prices and monopoly
power
• Every telephone company is a monopoly for
terminating traffic
• Each company can use interconnection prices
to partially control the market by making it
more costly for rivals to complete calls
• Thus, interconnection policy is key in any
regulation scheme
Number portability
• Once a client – especially a commercial client
– has chosen a number, he does not want to
change it
• That gives the company that controls that
number some monopoly power over that user
• Number portability is a regulation that
attempts to reduce that power
Improving telecom performance
without regulating prices
• Because regulation is costly, it may be
desirable to at least consider how to mitigate
monopolistic tendencies without it
• There are several schemes that reduce the
amount of price regulation
• All of them depend on the use of price
discrimination
Uniform vs. nonuniform prices
• Models we have considered till now have all involved
one price
• However, firms can also engage in nonuniform
pricing
• Nonuniform prices can arise in either competitive or
noncompetitive market
– Competitive example: quantity discounts – it costs less to
sell large quantity
– Noncompetitive example: two-part tariffs for wireless
operators
Discriminatory prices
• Vary with customer and/or with quantity
purchased
• Not cost-based
• Firm must have some market power, or prices
will be driven to cost
• Is market power sufficient? I.e., can any
monopoly price discriminate?
Market power and price
discrimination
• Market power is not sufficient
• Suppose monopoly tries to price discriminate. Then
– It will try to sell to each individual customer on demand
curve, but
– Customers on low end of demand curve will attempt to
resell to customers on high end, reducing monopoly profit
• Firm must be able to eliminate resale in order to
price discriminate
Resale
• Resale won’t occur when the costs exceed the
benefits
• Some goods can’t be resold – e.g., a filling in my
tooth
• In other cases transaction costs may be too high –
e.g., perishable goods, extra taxes, extra shipping
costs
• A firm wishing to price discriminate will try to find
ways to raise the cost of resale
Ways of raising resale costs
• Bundling
• Adulteration
• Vertical integration
Bundling
• Basic idea: goods that can be resold easily are
bundled with goods that cannot
• Examples
– Warranties/product support bundled with product
– Restaurants bundle atmosphere with food service
Adulteration
• Basic idea: Render commodity unfit for resale
• Examples
– Additive to rubbing alcohol makes it undrinkable
– No.2 heating oil is less filtered than diesel fuel, so it can’t
be used as transport fuel
– Telco uses same facility to deliver both high-value data
services and low-value emergency services, rendering
high-value services not resellable for other purposes
Vertical integration
• Basic idea: Control downstream suppliers as
well as current market
• Example
– Telco has local monopoly on wireline access
– It wants to sell at low price to LD and analog
wireless supplier, but at high price to digital
wireless
– Solution: Buy all three downstream firms
How price discrimination raises
profit
• Price discrimination enables producer to earn
some or all of consumer surplus
• Firm is able to charge high prices to highdemand customers, low prices to low-demand
customers
Price discrimination, graphically
Price
Profit with perfect
price discrimination
Profit under
uniform price
Demand
MC
MR
Quantity
Types of price discrimination
• First degree: Perfect price discrimination; firm
captures all consumer surplus
• Second degree: Average price paid varies with
quantity purchased
• Third degree: Price varies by customer type
First degree price discrimination
• Example: goods sold at auction
• Another example: housing market
• An interesting note: Markets under perfect
price discrimination have the same welfare
properties as pure competition and the
allocation of resources is Pareto optimal
Second degree price discrimination
• Example: quantity discounts, declining block
rates
• Virtually all markets have second degree price
discrimination to some extent
Third degree price discrimination
• Example: airline tickets
• Another example: wireless plans
• A firm must find a way to allow customers to sort
themselves out into sub-markets in order for this to
work
– Airlines accomplish this by requiring advance purchase for
cheap tickets
– Wireless customers self-select based on expected usage
Nonlinear pricing
• Basically, this is a form of second degree pd
• Generally takes the form of a two-part tariff
• Other approaches include multi-part tariff,
declining block rate
Nonlinear pricing, graphically
expenditure
Total expenditure
(slope = marginal
Expenditure)
Average expenditure
quantity
Contrasting 2nd and 3rd degree pd
• For successful 3rd degree pd, the producer must
– Be able to identify different demands
– Have information on those demands
– Prevent resale
• For successful 2nd degree pd, the producer must
– Prevent resale
– Keep down the number of small purchasers
– But doesn’t have to identify different groups
Tie-ins as pd
• A form of bundling where two separable
commodities sold together
– E.g., radio with batteries
– Or shoes with shoelaces
• Minimum price for goods sold separately
constrained by minimum demand, but by
summed minimum demand when sold
together
Tie-in example
Consumer 1
Consumer 2
Value of A
$8
$9
Value of B
$3
$2
Tie-in example, continued
• If firm sold products separately, and wanted to
max profits, it would charge $8 for A and $2
for B, for a total profit of $20
• If firm bundles the products, it can charge $11
for the A-B bundle, and still sell to both
customers for total profit of $22
Caveats
• Note that if both customers valued B at $3, there is
no difference between the tie-in profit max and the
separated sale profit max
• In general, customers must be heterogeneous in
taste over all products for tie-ins to work
• In general, there must be monopoly power in both
markets for tie-ins to work
Simplest approach
• Accept that while competition in the industry
may not be perfect, it does exist between
wireless and wireline operators
• Regulation is thus unnecessary; operators will
offer a variety of pricing plans in order to
attract clients
• These plans will lead to a desirable outcome
because of price discrimination
Caveats for simple approach
• Approach may not work if one operator has
already taken both fixed and wireless
operations
• In general, very strong competition policy at
the very least would be required to have any
hope of making this work
• This approach would still require a careful
regulation of interconnection
A less simple – but still light-handed –
approach
• In this approach, the regulated firm is free to
set prices, but subject to a cap placed on the
pricing of a bundle of services
• The cap is adjusted downward according to a
formula that depends on productivity
• The approach is designed to elicit “optimal”
price discrimination since the firm can adjust
prices to reflect varying demand elasticities
Conclusion
• Telecom regulation is complex
• Telecom shares some characteristics of other
infrastructure industries, but there are
important differences
• Regulators should strive to interfere as little as
possible but work toward improving industry
performance