from understanding the competition to policy making* Biliana - - PowerPoint PPT Presentation

from understanding the competition to policy making* Biliana Alexandrova Kabadjova Banco de México

First t workshop

• p on Agent-Based

d Modeling g for Bankin ing g and Finance, , Bank of Italy, , February ary 10th 2009

* The views expressed in this presentation are those of the authors and do not involve the responsibility of Banco de Mexico

 Motivation  Model of competition among payment card

providers (developed with Professor Edward Tsang and Dr. Andreas Krause)

 Model of intranetwork competition

(developed for the Mexican Payment Card Market)

 Conclusions

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

 Growing importance of the payment cards in the context of

modern economies

 Public interest in the pricing and rules governing this market

• The determination of the Interchange Fees (IF) has come under scrutiny by

competition authorities as market participants, specially large retailers associations and the payment card industry, have different views with respect to the level on which the IF should be set; and

• The payment service that the banks provided through the payment cards

are intimately related to other bank services (e.g. credit lines, deposit accounts, etc.):  It is difficult to determine and calculate the cost incurred by issuers and acquirers, which correspond exclusively to the cards’ payment services;

 The analytical models of the market assume a full acceptance of

the payment card;

 The complex underlying structure of the payment cards market

allows us to represent it as Multi-agent model

• There are network externalities on both sides of the market; and
• Usage externalities with relation to other payment methods as paper-based

instruments

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

 The first model was created by William Baxter, 1983

(Department of Justice, EU).

 More recently, an important body of literature has been created,

among which are the analytical models:

• Jean-Charles Rochet and Jean Tirole, 2002
• Richard Schmalensee, 2002
• Joshua Gans and Stephen King, 2003
• Julian Wright, 2003
• Sujit Chakravorti (Bob) and Roberto Roson, 2006
• Graene Guthrie and Julian Wright, 2007

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

Consumer Merchant Sells good at price p Marginal net benefit bS Marginal net benefit bB Issuer Pays p+f (f – consumer fee) Acquirer net cost cI Pays p – a (a – interchange fee) net cost cA Pays p – m (m – merchant discount)

Costumer Merchant Interactions at the Point Of Sale

Payment Card provider

Costumer’s fees and benefits Merchant’s fees and benefits

Custome r Benefits Merchant Fixed Fee

Customer’s decision to use the card Customer’s decision to hold a card Merchant’s decision to hold a card

Number of Customers using the card Number of Merchants accepting the card

Banks ks’ Profits

Number of Customers having the card Number of Merchants using the card Publicity Cost Custome r Fixed Fee Merchant Benefits

Banks’ Marke ket t Share re

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009 Consumers Fixed Fees Number of Consumers per Card

Profit per Card Issures

Costumer Merchant Interactions at the Point Of Sale

Payment Card provider

Strategy Profit and Market Share

Number of Issuers Consumers Fixed Fee Merchants Fixed Fee Consumers Benefits (Variable Fee) Merchants Variable Fee Marketing Effort Profit per Card Issuer 9 5.74 0.00 (0.65) 1.00 10.62 5,179,486.23 5 1.69 0.09 0.61 1.00 8.49 2,073,281.64 2 0.11 0.00 0.14 0.90 6.97 1,100,374.47

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

 The Model of Competition among card providers could be used to

advance our understanding of the real payment card market;

 The model is a good representation of the behavior of consumers

and merchants;

 We found that in a market with 9 competitors, the fixed fees

charged to consumers are higher then in a market with 5 or 2 card providers, which imply a higher profit for the competitors:

• Given the price structure established for those kind of markets, with more
• ptions in a market consumers find difficult to choose the lower price card.

 Machine learning techniques could be used as normative

framework to study the optimal price structure and level of the electronic payment instruments under specific scenarios

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

Consumers Merchants Interactions at the Point of Sale (POS)

Issuers

Fixed price and Variable benefits Fixed price and Merchant discount

Acquirers’

Interchange Fees per business line

Issuers’ Income Acquirers’ Income

Benefits from accepting the card Benefits from using the card

 Initial Conditions

• Consumers and merchants are assigned to a fixed locations
• The level of Interchange Fee is assigned
• The level of income is determined for each consumer
• The marginal profit is determined for each merchant
• Issuers and Acquirers decide their own prices of the payment

card

• The initial number of cards and electronic payment methods are

randomly distributed to consumers and merchants respectively

 Loop of Interaction at the point of sale

• Consumers’ decisions at each interaction
• Periodical decisions of consumers and merchants (after h

interactions)

 According to the level of Interchange Fee, the rate of

growth of the network is observed

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009



Selects a business line



Selects a merchant m

Mc,bl

• We assume that the each consumer knows a limited number of merchants for

each business line, located near him at a distance dc,m

• Given that is the number of common payment methods between

consumer and the merchant and is the number of consumers’ payments, we define the probability to select a merchant as following:



Decides how much to spend with the merchant



Decides which payment method to use

• In the case the merchant has applied price discrimination

 The consumer pays with card if the transaction’s benefits are grater then the price increase  Otherwise pays with cash

bl c

M m p c p m c m c p c p m c m c m c

N N d N N d

,

' ' , ' , , , ,

1 1 m1 m2 m3 c1

m c

d ,

m c

d ,

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

m c

d ,

p m c

N ,

p c

N



The consumer c

C

• Decides to join the network with probability
• In the affirmative case, chooses an issuer
• If already has a card, he could leave the network with probability



The merchant m

M

• Decides to accept cards with probability
• In the affirmative case, chooses a acquirer
• If already accepts cards, he could leave the network with probability

c c c

b x b exp exp

m m

b y exp 1

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

m m m

b y b exp exp

c c

b x exp 1

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

Modeling the network externalities’ impact on the network’s growth

Scaling at the minimum level of

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

Modeling the network externalities’ impact on the network’s growth

Scaling at the minimum level of

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

 This is the first model, in which the relationship

among IF and network growth is studied

 The model reproduces multiple scenarios of

consumers’ and merchants’ awareness of network externalities, which is a significant factor for the demand of electronic payment instruments

 From our observations, the level of Interchange

Fee has an important impact on the payment card adoption rate:

• This is relevant as the use of more efficient payment

instruments could be translated to important social savings

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

 The payment card market is characterized by:

• complex price structures,
• network externalities among card holders and

merchants accepting cards

• usage externalities among cards and other payment

methods

 These factors make difficult to asses

competition in the market;

 In order to go further in our understanding of

the payment card industry a computational models and algorithms might be useful.

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009



After h number of interactions, which is determinate by individual Poisson distribution, consumers and merchants evaluate the possibility to join or to leave the payment card network



These decisions are determined by the following factors:

• On the consumers’ side:
• is the number of merchants that the consumer knows that accept cards
• is the number of merchants with which the consumer has made his shopping
• and represent the degree of consumer’s appreciation of the positive

network externalities

• and are the consumer’s inertia to join or to leave the network
• is the consumer’s appreciation of a payment card over a cash
• On the merchants’ side:
• is the number of consumers that want to use a card with the merchant
• is the number of consumers that shop with the merchant
• is the number of consumers that have used their cards with the merchant
• and represent the degree of merchant’s appreciation of the positive

network externalities

• and are the merchant’s inertia to join or to leave the network
• is the merchant’s appreciation of a payment card over a cash

First workshop on Agent-Based Modeling for Banking and Finance, Bank of Italy, February 10th 2009

c

b

x x

m

b

y y