Transcript Price elasticity of demand

Pricing in Public Transport
Niclas Krüger, PhD in Economics
SAMOT, Karlstad University
Swedish National Road and Transport Research Institute
Centre for Transport Studies, Stockholm
TRENoP, Royal Institute of Technology
Why Revenue Management in PT?
• Varying but important source of financing PT
in most countries and cities
Revenue Management-1
• In Sweden approx 50% of costs of PT is covered by
fare revenues
Revenue= Price*Quantity
• Question:
What quantity should we sell to whom at which price?
Revenue Management-2
• Revenue management is “the process of understanding,
anticipating and influencing consumer behavior in order
to maximize revenue or profits from a fixed, perishable
resource (such as airline seats or hotel room
reservations).” (Source: Wikipedia)
• Has been a success in service industry for 25 years!
• Focus here:
• How can pricing achieve business and social goals?
• How can demand and revenues be managed by
pricing?
Price evolution of PT in a Swedish region
- The Baumol effect (wage increase with constant labor productivity)
- The Mohring effect (demand=>supply=>demand…)
Monthly pass price (blue line)
Gas price (read line)
Consumer Price Index (green line)
How do people make choices?
• People compare the benefits and costs of their travel
choices
• Benefits are measured as Willingness to Pay for trip
• What determines WTP for PT?
– Supply (space, OD availability, average waiting
time)
– Travel time (quality adjusted) & reliability
• In Sweden travel time is approx valued to 6
EUR/hour for leisure travels and 15 EUR/hour for
work travels
• Work commuters have probably high WTP for fast
public transport; this makes it possible to
implement BRT (Bus Rapid Transfer) and charge
extra for this
Measuring price sensitivity
• Price elasticity of demand is defined as the
responsiveness of the quantity demanded of a good
or service to a change in its price:
%

C
h
a
n
g
e
i
n
q
u
a
n
t
i
t
y
d
e
m
a
n
d
e
d

Q
Q
E


%

C
h
a
n
g
e
i
n
p
r
i
c
e

P
P
• If E<1, a price increase will increase revenues!
• If E>1, a price increase will decrease revenues!
• Vice versa for price decreases
Price elasticity
• The price elasticity differs depending on customer
• For example single ticket:
– A: visits Stockholm, does not care if 3 EUR or 4 EUR
– B: lives in Stockholm and close to work, does not have
montly pass, but is thinking about walking or taking tram=>
4 EUR is too much
– Hence, empirical question whether A or B dominates!
• Availability of alternative transport modes affects
willingness to pay and price elasticity (car owners
have low variable cost per trip)
Empirical evidence on price elasticity
• Empirically, the effect of a 10% increase of price, gas
price, supply and car ownership on PT-demand is
(Johan Holmgren):
• Demand decreases with 4% wrt to price increase
(=revenue increase!)
• Demand increases with 3,4% wrt gas price
• Demand increases 5.5% wrt to supply (=kilometers)
• Demand decreases 13.7% wrt car ownership
• Even if average price elasticity is low, it can be high
for certain times, customers and types of travel; we
have to identify these…
Optimal pricing from company perspective
• Difference between company’s optimal pricing and social
optimal pricing!
• The PT-company is a local monopoly and can artificially cut
capacity in order to get more from the remaining customers
• Monopoly sets:
Price = Variable Costs + Markup
• Markup is dependent on price elasticity:
=> The lower price sensitivity, the higher the markup
Social optimal pricing
• Social optimal price (=maximum social welfare):
Price = Variable Costs
• Intuition: As long as the passenger covers the costs she
causes (her WTP>P=VC), it is beneficial for society
• The variable costs (=extra cost associated with one extra
passenger) is often but not always very low
• If there are positive external effects (environment, better
supply), P<Variable Costs might be optimal
• Fixed costs do not matter for optimal pricing and can lead
to a vicious cycle; fixed costs affect however the amount
available for other public goods like education and health
care
Comparison
• Profit maximization vs social optimum
Prices in the market economy
• Prices have 2 basic functions:
• show relative scarcity
• What should we consume and produce?
• determines allocation of the good
• Who should consume and produce and how much?
• Example: Apartments in the city
• It is better to let prices show relative scarcity in order to
ensure effectiveness in transport systems. To enable different
groups to use PT it is better to use lump-sum transfer
payments!
What objectives do we want to achieve with
price structure?
Demand curve is downward-sloping:
 One price cannot achieve all objectives!
 Where is PBreak-even?
Solution: Price differentiation
• Neither monopoly-pricing nor social-optimum-pricing
feasible
• Instead: Price differentiation with budget-restriction
• Prices determined both by costs and demand
• Marginal cost plus markup, where markup dependent
on price-elasticity
• Rule1: Take a low markup from the customers who
have high price elasticity and vice versa
• Rule2: The sum of markups should be so that:
Fare revenues=Total cost-Subsidies
=> This minimizes the impact of the markup on
demand! Hence, close to social optimum and we breakeven too…
Example: Price differentiation with 2 groups
• One price is seldom optimal! Imagine:
•
•
•
•
100 customers WTP 8 EUR
100 customers WTP 5 EUR
If one price: 5 EUR=> revenue=1000 (200*5>100*8)
If two prices: => revenue=1300 (100*8+100*5)
• In general:
• Divide the market into segments:
• Heterogeneous groups with homogeneous customers which
are easy to identify and target
• Increase price for groups who are insensitive
• Decrease price for groups who are sensitive
=> Discounts increase revenues for price-sensitive groups!
Price differentiation- but how?
• Estimate price elasticity by experimenting:
• change price and see what happens. Compute E!
• Use the following proxies:
• Time/season
• Daytime, morning rush-hour, evening, weekend, summer season
• Payment mode
• Cash, SMS, credit-card, prepaid card
• Types of customers
• Age, occupation, income, …
• Product differentiation
• We do not know who has high WTP and who has low WTP; but
offering a set of choices the customer reveals his true
preferences…
Quality differentiation
• Nothing new:
• Different classes (business, economy)
• Extras (newspaper, fruit, internet)
• ”It is not because of the few thousands francs
which would have to be spent to put a roof over
the third-class carriage… What the company is
trying to do is prevent the passengers who can
pay the second-class fare from traveling third
class.” Dupuit (1804-1866)
Quantity differentiation: Trips
• Price depends on the number of travels
• We have single tickets, prepaid cards and monthly passes;
customer selects ticket according to expected number of trips
• Usually, the per trip price is lower for monthly passes than for
single tickets. Why?
• General for quantity discount: downward-sloping demand-curves=>
each additional unit is valued less
• Lower price sensitivity for single tickets
• Transaction costs
• Competition with car (we come back to that…)
• E-tickets: Make it possible to implement more complex pricestructures: no monthly passes, discount based on number of
actual trips within a period; contactless Visa/Mastercards with
daily caps, see London)
Quantity differentiation: Distance
• 3 different basic schemes for fare structure
• 1) Flat-fare
• One price independent of distance; simple, cheap
• 2) Zone-based
• One price in a zone, different prices between zones,
threshold effects!
• 3) Station-to-station
• Constant price per km (”justice”)=> No!
• Based on demand => Yes! (higher price on commute routes)
• A route-based scheme makes it possible to let fast routes with
high fares compete with slow and cheap services
Peak load pricing-1
• Transport capacity is costly (busses, trains)
• Traffic demand is unevenly distributed
– E g: rush hour morning and afternoon car traffic
Peak load pricing-2
• It would be very costly to satisfy whole peak demand
since capacity would be mostly unused other times
• MC is high during peak times (since we need extra
capacity for one extra passenger)
• Hence, taking a higher price at peak times will reduce
peak demand and increase off-peak demand
• Less capacity is needed…
• Another aspect: During rush hours if 50 persons want
to board buss it can take at least 10 minutes: Each
passenger is willing to pay ca 1 EUR for avoiding this!
Peak load pricing-3
• Under peak load pricing we need to take into
account the variabel capacity cost:
• P=VOC+VCC
– VOC: Variabel operating cost (gas, cleaning)
– VCC: Variabel capacity cost (interest payment and capital
depreciation per cycle with shifting demand)
– Peak prices should be distance based and higher in central
parts
• Variabel cost means cost for additional seat, but:
– VOC=> Cost for actual service level
– VCC=> Cost for potential service level
Peak load pricing-4
• PT is part of the whole transport system
• Under certain conditions it is suggested that
peak price in PT should be lower!
– Reasons: Cars external effects on environment and
congestion/costly road capacity should give
incentive to switch mode
• However, it is not always feasible
– Capacity used at maximum in PT
– A low price increases travels mostly by PT-users
and not so much by mode switchers
Dynamic pricing-1
• Dynamic demand model:
• qt=f(qt-1,x) with δqt/δqt-1>0
• Demand this year is dependent on demand last
year
• Reasons:
• Habit formation
• Threshold to switch mode
(car purchase/driving license)
Dynamic pricing-2
• Model solution: P*SR<P*static<P*LR
• Interpretation:
• Optimal dynamic price is lower than static optimal price
• Hence, a low price this year can be regarded as an
investment to buy future market shares!
Other pricing methods:
Advance booking
• In the context of long distance PT: Advance
booking
• Waiting to book means trade-off between
better information about travel needs and the
risk of getting rationed
– If capacity is high, everybody will wait to last
minute to book a ticket
– If capacity is low, some will wait and some will
book early; this (rational) behavioral differences
can be exploited for price discrimination
Other pricing methods:
Tying
• Selling the transport service and other services/products as a
package:
Movie
Transport
Consumer A
6 EUR
10 EUR
Consumer B
10 EUR
6 EUR
• Separate: Sell transport for 6 & movie for 6=>Revenue=24
• Package can be sold for 16 =>Revenue=32
• Example:
Internet on board; concert-tickets including bus-ticket
Other pricing methods:
Competitive pricing
• Sometimes we have one or more competitors
• Your price decision depends on what you think your
competitor will do and vice versa
• Underbidding each other, the company with the
lowest cost will win and sell the service
• Example:
Swebus (national buss company) is since summer 2009
allowed to pick up travelers within regions on routes
normally served by PT. They set a lower price than the PTcompany. Can PT compete with Swebus? Hardly, since
each traveler is an additional revenue for Swebus, they can
always underbid PT
Pricing as part of the marketing-mix-1
• 4P in marketing (marketing instruments):
-
Price
Product (number of departures/service quality)
Place (availability/distribution)
Promotion (information/advertisement)
=> These 4 marketing instruments are interdependent!
• Example:
- Price affect service quality perception
- Service quality affects price we can charge
Pricing as part of the marketing-mix-2
• In general: the marketing budget should be spent so that an
additional dollar has the same additional (marginal) effect for
all 4 instruments
Example: In optimum, a dollar spent on advertising increases revenue
with 10 dollars, whereas a dollar spent on service quality
improvement yields 20 dollars in revenues=> we should re-allocate
money from ad-budget to service improvement until they are equal
• The Dorfman-Steiner relation:
Share of advertising budget of revenue
= Advertising elasticity/Price elasticity
=> in PT low price elasticity but even lower advertising elasticity
=> low ad-budget!
Behavioral pricing-1
• Behavioral pricing is about bias in behavior (deviations from
rational behavior) affecting the response to pricing decisions
• i) Extremeness aversion: People do not choose extremes
• Experiment with 2 groups:
• Results:
• 1. More choose the Standard model!
• 2. Some choose the expensive Deluxe model!
Conclusion: Always provide at least 3 versions!
Behavioral pricing-2
• ii) Self-control issues:
– We know what is good and bad for us but still we do not
act according to our knowledge (e g smoking)
– We perceive future benefits as too low
– We perceive the initial costs as too high
• Reasons:
– High ”discount rate”
– Reconciliation of cognitive dissonances
• Implications: low price now, high price later
Behavioral pricing-3
• iii) Response to payment mode and time
• Payment mode:
People are less price sensitive wrt card payments
compared to cash
• Difference between time of purchase and time of
payment:
People are less price sensitive if payment and purchase are
separated (buy now, pay later)
PT-example: Prepaid-cards, SMS-tickets, credit cards
Behavioral pricing-4
• iv) Framing and loss aversion
• Prospect theory:
• Reference point is important
• Bigger impact of price increases
than of decreases
• Diminishing impact of both
price increases and decreases
• Mental accounting:
• We have different reference prices for different expenses
(Food, travel, housing…)
Behavioral pricing-5
• Implications for pricing:
• Marketing of price changes:
• Better to raise prices infrequent but considerably
• A moderate price increase (=loss) might outweigh a big
service improvement (=gain)
• If many tickets: many small price decreases can
outweigh one big price increase
• A big price decrease can outweigh a small price
increase
• Price and quality often observed to be sticky
• It is important to set the right reference price
(framing)
Behavioral pricing-6
• Example for how to change reference price:
• Newspaper seller called:
• 1. Weekend only: 100 SEK per month
• 2. Whole week: 150 SEK per month
• What does the seller want to sell?
• Option 2!
• Option 1 is merely providing us with a reference price
• Option 2 is perceived as relatively cheap since we get
the double amount of newspaper for only 50% more
money
Want more?
• Literature:
Oz Shy: How to price
(especially Chapter 1 for an overview and the sections
labeled “Regulated public utility”)
Nils Fearnley: Inventive Pricing of Urban Public Transport
Stefano DellaVigna: Economy and Psychology: Evidence
from the Field