The Retail Payment Costs and Savings in Albania JUNE 2018 Finance, Competitiveness & Innovation Global Practice Payment Systems Development Group © 2018 International Bank for Reconstruction © 2018 Bank of Albania and Development / The World Bank Sheshi “Skenderbej”, Nr.1 1818 H Street NW Tirana, Albania Washington DC 20433 Telephone: + 355 4 2419301/2/3 Telephone: 202-473-1000 Fax: + 355 4 2419408 Internet: www.worldbank.org Internet: www.bankofalbania.org E-mail: public@bankofalbania.org This volume is a product of the staff of the World Bank and Bank of Albania. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of the World Bank or the governments they represent. The World Bank and Bank of Albania do not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank and Bank of Albania concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this publication is subject to copyright. Because the World Bank and Bank of Albania encourage dissemination of their knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution is given. CONTENTS Foreword iii Acknowledgments iv Abbreviations v EXECUTIVE SUMMARY 1 I. METHODOLOGY 4 OWNERSHIP AND USAGE OF TRANSACTION ACCOUNTS AND II.  5 ELECTRONIC PAYMENT INSTRUMENTS ACCESS POINTS III. 7 COSTS OF RETAIL PAYMENTS BY ACTOR AND FOR THE TOTAL ECONOMY IV. 10 IV.1. Consumers 10 IV.2. Businesses 11 IV.3.  Government Agencies 13 IV.4.  Payment Service Providers and Payment Infrastructure Providers 14 IV.5.  Total Economy 20 V. ECONOMY SAVINGS 20 VI. CONCLUSIONS 21 ANNEX A: BREAKDOWN OF COSTS FOR THE SUPPLY SIDE BY PAYMENT INSTRUMENT 26 ANNEX B: INDICATIVE SURVEY QUESTIONNAIRES 29 REFERENCES 33 BOXES, TABLES, AND FIGURES Box 1: Survey Administration for the Different Stakeholders 6 Box 2: Sample Questions for the Consumer Survey 29 Box 3: Sample Questions for the Business Survey 30 Box 4: Sample Questions for the Government Agency Survey 31 Box 5: Sample Questions for the Payment Service Provider Survey 32 Table 1: Types of Retail Payments 4 Table 2: Utilization Ratios for Payment Cards across Time 7 Table 3: Annual Costs for Government Agencies by Payment Instrument (2014) 15   i ii   The Retail Payment Costs and Savings in Albania Table 4: Total Economy Costs for All Retail Payment Instruments Examined (2016) 19 Table 5: Substitution Scenarios and Savings at the Economy Level 21 Table 6: Substitution Scenarios and Savings for Indicative Use Cases across Different Actors 22 Figure 1: Indicative Transmission Methods 5 Figure 2: Access Points over Time 8 Figure 3: Active Payment Cards in Circulation over Time 9 Figure 4: Annual Costs for Consumers by Payment Instrument as Percentage of 2014 GDP 12 Figure 5: Per-Transaction Costs by Payment Instrument across Transmission Methods 12 Figure 6: Per-Transaction Costs by Transmission Method across Payment Instruments 13 Figure 7: Annual Costs for Businesses by Payment Instrument as Percentage of 2014 GDP 14 Figure 8: Volume of Payments/Transfers Processed by Paper-Based over Electronic Channels over Time 15 Figure 9: Volume of Payments/Transfers Processed by Payment Instrument over Time 16 Figure 10: Volume of Cash Deposits and Withdrawals over Time 17 Figure 11: Annual Costs for Payment Service and Infrastructure Providers by Payment Instrument as 17 Percentage of 2016 GDP Figure 12: Per-Transaction Processed Cost by Payment Instrument 18 Figure 13: Supply-Side Industry Cost Indicators for Albania (2016) 19 Figure 14: Total Economy Costs (Resource) by Payment Instrument as Percentage of 2016 GDP 19 FOREWORD Modernizing payment systems has become a prominent Moreover, it allows for the development of substitution component of the financial sector reform agenda in many scenarios (from more costly to less costly retail payment advanced and developing economies. For governments instruments) and the resulting cost savings. The guide and central banks, both strengthening the security and can be adapted to country-specific circumstances with- reliability of the national payments system and fostering out losing its comparative nature, across time and across the use of efficient payment instruments are important countries. It is built on four main principles: applicability, public policy goals. Greater use of electronic payments comparability, efficiency, and standardization. rather than cash and other paper-based instruments has important economic and social benefits, including lower- Promoting the smooth functioning of payment systems ing costs and thereby raising economic efficiency, and is at the core of the mandate of the Bank of Albania. broadening access to financial services for those currently The Bank of Albania undertakes the roles of operator excluded. In turn, the cost-efficiency benefits of electronic of payment systems and overseer, and a catalytic role payment instruments vis-à-vis paper-based payment of the national payment system development. Under instruments can be a strong incentive for reforms and the leadership of the Bank of Albania, a National Pay- actions by all stakeholders. However, national authorities ment System Committee was established to support and stakeholders may often face difficulties in underpin- and contribute to increasing the security, stability, and ning and monitoring retail payment reform efforts due the efficiency of the national payment system in Albania. In difficulty of ascertaining the costs incurred in the retail this context, the Bank of Albania and the World Bank payments market and the (potential) savings that could be Group signed a memorandum of understanding with the realized if certain reforms were to be undertaken. objective to implement the World Bank guide in order to measure the costs of retail payments in Albania. The The World Bank Group has developed A Practical Guide results of the study will enable the Bank of Albania to for Measuring Retail Payment Costs,1 which fills a gap in ground payment system policies in a comprehensive this space and allows the systematic identification, col- analysis of the retail payments market and to identify key lection, and measurement of data associated with the actions to accelerate a transition from cash/paper-based cost of retail payment instruments, for the demand and to electronic payments. The results will also enable the supply side of the payments chain. The guide comple- industry as well as consumers and businesses to identify ments other policy and guidance documents published potential actions that will lead to a more efficient use of as part of the World Bank’s Retail Payments Package. retail payment instruments. Ceyla Pazarbasioglu Elisabeta Gjoni Senior Director First Deputy Governor Finance, Competitiveness and The Bank of Albania Innovation The World Bank   The guide (World Bank 2016) is available at http://documents.worldbank.org/curated/en/255851482286959215/pdf/ 1.    iii 111216-WP-P155382-PUBLIC-ABSTRACT-SENT.pdf ACKNOWLEDGMENTS This report would not have been possible without the Team Leader of the Remittances and Payments Program, generous support of the Swiss State Secretariat for Eco- World Bank. Sharmista Appaya, Keler Gjika, Johanna Jae- nomic Affairs through the Remittances and Payments Pro- ger, and Gynedi Srinivas (all World Bank) provided useful gram in the Balkans, the Department for International comments as peer reviewers. The team is also grateful to Development through the Harnessing Innovation for Mahesh Uttamchandani (Practice Manager, Financial Financial Inclusion Program and the Bank of Albania. Inclusion, Infrastructure and Access, World Bank), Mario Guadamillas (Practice Manager, Europe and Central Asia, Lead authors of the report are Holti Banka (World Bank) Finance, Competitiveness and Innovation, World Bank), and Maria Teresa Chimienti (World Bank). Support was Massimo Cirasino (World Bank), Thomas Lammer (World provided by Anjeza Beja and Valentina Semi (both Bank of Bank), and Oya Pinar Ardic Alder (World Bank). Charles Albania). The overall guidance was provided by Harish Hagner provided editorial assistance and Naylor Design, Natarajan (Lead, World Bank). Ceu Pereira was the Task Inc. provided design and layout assistance. iv ACRONYMS AND ABBREVIATIONS ATM automated-teller machine GDP gross domestic product POS point of sale PIP payment infrastructure provider PSP payment service provider PSU payment service user All dollar amounts are U.S. dollars unless otherwise indicated.   v EXECUTIVE SUMMARY In order to accelerate access to bank accounts, one of this report’s recommendations is that the Bank of Albania, This report is based on a study of the costs of retail pay- banks, and other payment service providers (including ments in Albania using a methodology developed by the through the National Payment System Committee) could World Bank’s Payment Systems Development Group, part explore the viability of providing a basic transaction of the Financial Inclusion, Infrastructure, and Access at the account at little or no cost to all individuals and businesses Finance, Competitiveness, and Innovation Global Prac- that do not hold an account. Moreover, the current tice. By applying this methodology to Albania, the study account and payment product offerings might need to be aims to establish a sound economic baseline for the reviewed in light of improving overall design and afford- national retail payments system in terms of costs of differ- ability. Public and private sector strategies should align to ent payment instruments to better guide system develop- the needs, habits, and barriers to account ownership and ment and enable high-impact changes. Efficiency gains use of the unbanked and underserved (consumers and resulting from the migration to lower-cost retail payment businesses alike) and identify solutions with measurable instruments and a more efficient use of those instruments targets. Technological and business model innovations could have significant benefits for economic development that leverage ubiquity, are easy to use, and lower trust and reach to lower-income households. barriers should be encouraged. In parallel, initiatives on access and usage should be complemented by further ini- In 2014, 38.5 percent of Albanian adults had access to a tiatives to increase the financial literacy levels in Albania. bank account, while 24 percent and 3.5 percent owned a debit and credit card, respectively.2 Since the last demand- In terms of access points, during the period 2015–16, the side survey, supply-side data indicates that the number of net growth rate of bank branches and automated-teller bank accounts per 100,000 adults has grown modestly machines (ATMs) per 100,000 adults was negative (–4.2 (4.8 percent on an annual average basis). E-money, which percent and –3.2 percent, respectively), while the net was recently introduced to the Albanian market and there- growth rate of e-money agents and point-of-sale (POS) fore was not recorded in the 2014 demand-side survey, terminals was positive (+25.7 percent and +6.3 percent, reached 177,000 accounts in 2016. Debit and credit cards respectively). in circulation per 100,000 adults have also increased between 2014 and 2016, although growth slowed in 2016 Albanian consumers report making and receiving day- (from 10.4 percent to 5.5 percent for credit cards, and to-pay payments overwhelmingly in cash (96 percent of from 6.8 percent to 5.5 percent for debit cards). payments initiated as compared to 90 percent of pay- ments received). Across all business sizes surveyed, 99.2 percent of all payments in volume terms were received The latest Findex data (Demirgüç-Kunt et al. 2018) indicate that 2.  in cash, while 66 percent of initiated payments were rep- bank account access among adults in Albania has reached 40 percent. resented by cash. Businesses accepting electronic pay-   1 2   The Retail Payment Costs and Savings in Albania ments through a POS device are limited to 15 percent of via the Internet. Credit card payments follow closely, at the business sample. The planned implementation of the 1.2 million. As opposed to debit cards, credit cards were interbank direct debit is expected to contribute to more used more often for online payments (56 percent) than at cost-efficient bill payment services and less usage of cash the point of sale (44 percent). In 2016, 0.7 million direct among consumers and businesses. E-money products are debits (90 percent paper-based and 10 percent elec- also demonstrating potential in this respect. tronic) and 0.3 million e-money transactions (92 percent via the mobile phone network and 8 percent via the Inter- The demand-side surveys suggest that cash payments are net) were processed. concentrated at the point of interaction3 and for the pur- chase of groceries and necessity goods. However, other For their current use of retail payment instruments, the service channels and use cases also rely heavily on cash. demand-side actors (consumers, businesses, and govern- After groceries and necessity goods, Albanians make the ment agencies) face annual costs equivalent to 1.6 percent most payments for periodic bills and services; of these, 90 of Albania’s gross domestic product (GDP), the cost of cash percent are made in cash. Similarly, person-to-person alone being about 1 percent and consumers bearing the transfers are cash-based (95 percent). Although govern- highest share. On a per-transaction basis, though, paper- ment disbursements to a great extent are paid directly to based credit transfers are the costliest payment instrument the beneficiary’s bank account, old-age pensions are across users’ categories, followed only by cash. The annual channeled mostly through physical service channels and aggregated costs associated with electronic payment paid out in cash. instruments account for less than 0.35 percent of the GDP. Moreover, the analysis reveals that service channels more The annual cash-in/cash-out activities are also significant closely associated with the use of cash and other paper- for Albania, as indicated by the findings. In 2016, there based payment instruments (such as pay office, agent out- were about 22.5 million cash withdrawals in the country, let, and bank branch) are the costliest for users as a result of which 59 percent were made via debit card at an ATM of the time spent to arrive to the service location, wait in and 32 percent were made at bank branches. Another line, and undertake the transaction, compared with other 19 percent of cash withdrawals represent cash-out at the service channels such as the Internet. e-money agent. Moreover, there were approximately 10.5 million cash deposits, 85 percent of which were at bank The volumes of cash payments handled at the bank branches and 15 percent were at (e-money) agents. branch combined with the relative cost of the bank branch as a transmission method suggest that bank customers The report suggests that government and industry incen- may benefit from banks implementing agent models, thus tives (such as fiscal, monetary incentives, but also inno- cutting time-based costs, especially in rural areas, and vative business models underpinned by technology) may possibly also mitigating cultural and trust barriers. prove useful to boost acceptance of electronic payments at the point of interaction. In this context, efforts to com- The annual cost borne by the supply-side actors in provid- bat informality will also be significant in accelerating the ing retail payment services amounts to 1.31 percent of the shift to electronic payments at the point of sale, given GDP. Cash stands out again as the costliest payment that informality and cash have a symbiotic relationship. instrument on an aggregate basis, equivalent to 0.65 per- Expanding Internet/mobile banking and leveraging cent of the GDP. Debit card costs follow and are slightly e-money for bill payments are critical to phase out other higher than credit card costs, driven primarily by the costly transmission methods of cash and paper-based higher volume, then credit transfers, direct debits, and payment instruments (for example, pay office). e-money. As it is the case also for the demand side, on a per-transaction basis, paper-based credit transfer and Among all noncash payment instruments processed in paper-based direct debits become the most costly pay- 2016, credit transfers exceeded all other instruments ment instruments, followed by cash deposit/withdrawal based on volume. Specifically, the number of credit trans- costs. All electronic transactions are less costly than the fers reached about 7.7 million, of which 77 percent were equivalent paper-based ones, with electronic direct debit paper-based and 23 percent electronic. The second and electronic credit transfers being the least costly of all. most-used electronic payment instrument was debit Other important industry cost indicators derived in the cards, with 1.9 million transactions, 94 percent of which analysis include the average annual cost of maintaining a took place at the point of sale and 6 percent transpired physical POS terminal, 12,044 lek ($103.6); the average annual cost of maintaining an off-site ATM, 475,000 lek 3.  The point where the payer and the payee converge (typically the ($4,085); and the average annual cost of maintaining an point of sale). agent, 126,600 lek ($1,084). The Retail Payment Costs and Savings in Albania   3 Local card processing, under certain circumstances, could cent is linked to consumers, 14 percent is linked to busi- contribute to decreasing some of these costs, which is nesses, and 1 percent to government agencies. operationally strategic for banks (given the growth of pay- ment card transactions) and the payments system as a The study concludes that savings for up to 0.9 percent of whole (for example, as a result of local governance and the GDP or almost half of all costs of paper-based pay- decision-making). ment instruments can be achieved for the Albanian econ- omy by substituting paper-based with electronic payment At the economy level, where demand and supply side are instruments. When specific use cases are considered in combined and only resource costs are considered, cash the context of the Albanian payments system, a 50 per- remains the costliest payment instrument, at 1.7 percent of cent substitution of cash pensions with electronic credit the GDP. Payment cards follow at a distant 0.22 percent for transfers would generate annual savings of 385 million lek credit cards and 0.21 percent for debit cards. Costs associ- ($3.3 million) for the pensioners, 350 million lek ($3 mil- ated to paper-based credit transfers amount to 0.20 per- lion) for the government, and 500 million lek ($4.3 million) cent of the GDP, while jointly electronic credit transfers, for the payment service/infrastructure providers. Similarly, direct debits, e-money, and online money account for 0.17 if 50 percent of cash utility bill payments were substituted percent. Of all paper-based payment costs, 50 percent are with electronic direct debits, consumers would save about borne by consumers, 25 percent by businesses, 24 percent 750 million lek ($6.5 million), businesses would save about by the payment service/infrastructure providers, and 1 per- 1 billion lek ($8.2 million), while payment service/infra- cent by government agencies. With regard to electronic structure providers would save approximately 1.8 billion payment instruments, about 55 percent of the cost is lek ($15.5 million). linked to payment service/infrastructure providers, 30 per- 4   The Retail Payment Costs and Savings in Albania I. METHODOLOGY4 payments) authorization, switching, clearing, and netting and/or settlement services.6 Payment service users (PSUs) constitute the demand side of retail payment actors. PSUs can be broadly catego- Using and providing payment services generates costs for rized into consumers (the term households, individuals, the actors involved—that is, for PSUs and PSPs/PIPs. or persons could also be used), businesses (registered While costs can be classified several ways, here the enterprises, including retailers, sole proprietorships, and self- emphasis has been put on the practical application of the employed providers of goods and services), and govern- classification. The following two categories are consid- ment agencies (federal, regional, and local ones). All of ered: (i) nature (resource vs. transfer) and (ii) variability them typically have a dual role, being the payer for certain (fixed vs. variable). These types of costs are distinct within retail payment transactions and the payee for others. the category but not mutually exclusive between catego- ries, and no dedicated relationship exists between them. Payment service providers (PSPs) and payment infrastruc- That is, a cost element that is classified as a resource (or ture providers (PIPs) constitute the supply side of the pay- transfer) can be either fixed or variable. ments value chain. PSPs issue payment instruments (for example, cash, payment cards, e-money) to the payment Most of the cost data includes variable costs and can be service users (PSUs)/demand side and/or offer services that allocated up front to single transactions (for example, allow PSUs to initiate retail payments with a given payment transaction time). However, data for some cost elements is instrument, including supporting services (for example, obtained per period (for example, on a weekly, monthly, opening and maintaining transaction accounts). A central or annual basis), rather than on a per-transaction basis. bank is typically the issuer of banknotes and coins, while These are mainly fixed costs, such as fees associated with commercial banks, microfinance institutions, e-money the underlying transaction account and/or payment instru- institutions, and other nonbank PSPs are issuers of noncash ments. Upon data analysis, though, these cost elements (including electronic) payment instruments. In addition to are allocated to single payment transactions using trans- PSPs that issue payment instruments and (in the case of action volume as the main allocation key, in order to cre- noncash payments) typically maintain transaction accounts, ate a common basis. there are institutions that offer over-the-counter cash-in/ cash-out transaction services (for example, money transfer The methodology examines payment transactions in their operators) or additional functions (for example, currency overall context of their use by including important factors exchange bureaus, if foreign currency plays a considerable (that is, payment type, transmission method, and use role in a country’s retail payment system). On the other case) that are likely to affect retail payment costs. Payment hand, PIPs provide payment infrastructure services usually transmission method indicates how a payment is initiated to PSPs and not directly to PSUs. Typical services provided and/or received. The two main categories of transmission by the PIPs include cash logistic services or (for noncash methods are in-person payments (for example, point of interaction, ATM, bank branch, pay office, agent) and remote payments (the Internet, telephone/mobile phone TABLE 1: Types of Retail Payments network). Different use cases are also considered in the PAYER GOVERNMENT context of consumers (remittances, payments for goods/ PAYEE CONSUMER BUSINESS AGENCY services, payments for periodic bills, salaries, pensions, Consumer P2P P2B P2G social assistance, and so forth). Business B2P B2B B2G Government agency G2P G2B G2G5 The methodology provides insights on the potential sav- Note: B2B = business to business, B2G = business to government, B2P = business to ings that may arise when migrating from paper-based to person, G2B = government to business, G2G = government to government, G2P = electronic processing of payments. It is possible to con- government to person, P2B = person to business, P2G = person to government, and P2P = sider a number of different substitution scenarios (35 per- person to person. cent, 70 percent, and 100 percent conversion rates). For an in-depth read of the generic methodological framework, 4.  It is important to emphasize that the present methodology follows 6.  please see World Bank (2016). a functional approach rather than an institutional approach when Government-to-government payments will be considered only in 5.  identifying and measuring the costs associated with the PSP/PIP. the context of secondary activities, such as cash receipts and Depending on the market conditions, one institution can fall into checks, supply of change, and transfer of funds between own the PSP category as well as the PIP category. A typical example of accounts. this would be central banks, which in many countries not only issue currency (a PSP function) but also offer payment settlement and sometimes even payment switching, clearing, and netting services (PIP functions). The Retail Payment Costs and Savings in Albania   5 FIGURE 1: Indicative Transmission Methods Savings in each scenario are shown by using the overall OWNERSHIP AND USAGE OF II.  annual cost figures and considering certain reduction pro- TRANSACTION ACCOUNTS jections resulting from the migration of one payment AND ELECTRONIC PAYMENT instrument (for example, paper-based credit transfer) to INSTRUMENTS another (electronic credit transfer or e-money). At the time of the demand-side survey (2014), 38.5 per- Given that the surveys were conducted at different time cent of Albanian adults had access to a bank account,7 periods, the demand-side analysis portrays findings that with 33.5 percent having a personal account and 5 per- pertain to the 2014 calendar year, while the supply-side cent accessing an account through a family member. analysis portrays findings that pertain to the 2016 calen- Since e-money products were not yet available but were dar year. foreseen to become available in the near future, the sur- vey included access to online money (PayPal) as a proxy; Moreover, not all payment instruments are relevant for all as of 2014, 0.5 percent of the Albanian population had actors examined. Specifically, for consumers, the follow- access to online money. Bank account ownership falls to ing instruments are analyzed: cash, paper-based credit 23 percent of the adult population in rural areas, as com- transfer (interbank and intrabank), debit card, credit card, pared with 50 percent in urban areas. Across age catego- electronic credit transfer (inter- and intrabank), and online ries, adults who were 65 years old or older seem to display money. For businesses, the following instruments are the lowest ratio for bank account access (20 percent), as examined: cash, paper-based credit transfer (inter- and compared to adults who were 40 to 64 years old (45 per- intrabank), debit card, credit card, electronic credit trans- cent) and 18 to 39 years old (39 percent). Only 2 percent fer (inter- and intrabank), and electronic direct debit (intra- of the population reported using Internet banking at least bank). For government agencies, the following instruments once over the past six months. are analyzed: cash, paper-based credit transfer (inter- and intrabank), and electronic credit transfer (inter- and intra- Based on the survey findings, 24 percent of the adult pop- bank). For payment service and payment infrastructure ulation in Albania owns a debit card, and 3.5 percent of providers, the following instruments are examined: cash, the adult population owns a credit card. The highest debit paper-based credit transfer (inter- and intrabank), paper- card ownership ratio is observed among adults between based direct debit (intrabank), debit card, credit card, electronic credit transfer (inter- and intrabank), electronic Defined as an account held at private commercial banks that 7.  direct debit (intrabank), and e-money. enable users to make deposits and initiate/receive payments. 6   The Retail Payment Costs and Savings in Albania BOX 1 Survey Administration for the Different Stakeholders8 Consumers: Face-to-face consumer surveys were local government institutions, and public utility com- conducted by the Institute of Statistics in coordination panies. The study captures all main central govern- with the Bank of Albania and the World Bank Group in ment ministries and agencies. As far as local order to collect primary consumer data. The average government payments are concerned, this study is duration of the interviews was 30 minutes. The sam- focused mainly on the municipality of Tirana due to its pling frame was based on summary data for the enu- importance in terms of population and business activ- meration areas defined for the 2011 Albania Census ity concentration. Specifically, Tirana represents about of Population and Housing. The availability of a rela- 30 percent of the total population of Albania, while in tively recent sampling frame ensured that an effective terms of economic activity, businesses operating in and representative sample could be selected for the Tirana account for almost 45 percent of all businesses survey. Ultimately, 897 adults responded to the survey in the country. With regard to public utility payments, at the place of their residence. Overall, the final sam- as in the case of local government payments, this ple consisted of 50.2 percent of rural area respon- study also focuses on water bill payments in Tirana. dents and 49.8 percent of urban area respondents. PSPs/PIPs: For the supply side of the payments mar- There is an even distribution across most categories ket in Albania, no sampling technique was used, for the different sociodemographic characteristics given the small number of supply-side actors in the within the sample. market. Data was gathered from all 16 commercial Businesses: The National Business Register was used banks operating in Albania, the Bank of Albania (in its as a basis for the sampling frame of the business sur- PSP role as currency issuer and in its PIP role as the vey. The Institute of Statistics has maintained the regis- operator of the automated clearinghouse and the ter since the mid-1990s. A stratified sampling real-time gross settlement system), two e-money issu- technique was used in order to select the sample, with ers, one money transfer operator, and one bill pay- a final sample size of 716 businesses. Strata were ment initiator. A day-long workshop was held in Tirana based on six economic sectors (agriculture, industry, in January 2017 in partnership with the Bank of Alba- construction, trade, transport, and services) and four nia. All survey participants were present in order to be business sizes (micro, small, medium, and large). Con- familiarized with the objectives and the structure of sidering the importance of Tirana, the capital city, as the survey questionnaires. After the workshop, cus- the country’s main economic center, around 50 percent tomized questionnaires were sent electronically to of the companies in the sample were located there. Of each participating institution. Respondents were the 716 businesses selected to be interviewed, 104 given a month to fill out the questionnaires (given that refused to participate, while 150 businesses were shut data from the different departments had to be put down, idle, or no contact could be established. The together) and return them to the Bank of Albania and responding businesses came from all six economic the World Bank. Issues/questions that came up during sectors; the largest sectors were trade and manufactur- that time frame were addressed by the Bank of Alba- ing. In terms of geography, the responding businesses nia and World Bank experts. Data quality control fol- were allocated across 27 towns/cities, covering all 12 lowed, and in some cases, additional information and administrative districts of Albania. clarification was requested from the responding insti- tutions, particularly where inconsistencies and errors Government agencies: These can be divided into were detected. three main categories: central government agencies, It should be emphasized that even though some questions of this consumer survey are similar to Findex, in that they cover financial inclusion 8.  aspects, the overall structure and objective is different, in that the main purpose of this survey is to capture costs and savings associated with retail payments. The Retail Payment Costs and Savings in Albania   7 the ages of 30 and 39 (32 percent). Other age categories The vast majority of surveyed businesses in Albania (92 display similar debit card ownership ratios, except for percent) have access to a bank account, although account those of 65 years old or older, whose ownership ratio is ownership varies significantly based on the size of the only 7 percent. business, and merchants’ acceptance of electronic pay- ments appears to be limited. Forty percent of microbusi- Over three-quarters of debit card holders reported using nesses report not having a bank account, as compared their card for cash withdrawals only. This finding is sup- with 1 percent or less of small/medium/large businesses. ported in supply-side transaction data for 2014, accord- The majority of medium and large businesses (more than ing to which the average debit card holder made about 70 percent) report having three or more bank accounts. 14.9 ATM cash withdrawals, 1.5 POS debit card pay- Overall, slightly more than half of the total business sam- ments, and 0.04 Internet debit card payments. In 2016, ple use Internet banking, with peaks of 62 percent for the equivalent numbers were 15.2 for ATM cash with- large businesses.10 Only 2 percent of microbusinesses and drawals, 2.1 for POS debit card payments, and 0.1 for 14 percent of small businesses reported having a POS Internet debit card payments.9 terminal. E-commerce is offered rarely—by less than 10 percent across business sizes. The pattern for credit cards is different from debit cards, given that the utilization ratios across transmission meth- ods do not diverge as much. In 2014, as shown in supply- III.  ACCESS POINTS side data, the average credit card holder made 8.9 ATM cash withdrawals, 8.3 Internet credit card payments, and On the supply side, the analysis provides time series data about 4 POS credit card payments. In 2016, the numbers (2014–16) on certain payments infrastructure indicators for were 11.3 for cash withdrawals, 8.2 for Internet payments, Albania. Throughout this study, bank branches refer to and 6.3 for POS payments. commercial bank branches, while agents (outlets) refer to physical locations associated with e-money issuers/opera- Prepaid cards are incipient and also are used more fre- tors. POS terminals are deliberately specified as physical quently for withdrawals than payments, although this pat- terminals in order to distinguish them from virtual11 POS tern is reversing. In 2014, the average prepaid card holder terminals, which exist but are very limited in Albania (21 for made 0.8 cash withdrawals and 0.04 POS payments, while 2014, 30 for 2015, and 28 for 2016). It should be noted that in 2016 there were 0.9 cash withdrawals and 0.1 POS pay- e-money institutions were licensed by the Bank of Albania ments. Moreover, prepaid cards in circulation in 2016 for the first time in 2015. Therefore, no e-money statistics reached about 42,000, while e-money accounts in Albania reached about 177,000 in 2016. For the purposes of this study in Albania, micro- and small 10.  businesses have been defined as those having 1–4 and 5–19 9.  In order to derive such comparisons, the utilization/usage ratio is employees, respectively. used. The utilization ratio, generally, is defined as the number of 11. A virtual POS terminal is the web version of a physical debit/ payments initiated using a payment card at transmission channel credit card terminal machine. It is a software application that is A during period B/the number of payment cards in circulation hosted online, usually on the service provider’s servers, and can during period B. Utilization ratios are derived for each of the be accessed from any Internet-connected web browser. A virtual three payment cards for their use at the point of sale, over the terminal allows for the input of the customer’s card information Internet, and at ATMs, over time. Essentially, this indicator shows directly into a web-based payment form, which can then be used the number of payments/transactions a cardholder makes, on to process an electronic transaction. average, during a time period (typically a year). TABLE 2: Utilization Ratios for Payment Cards across Time DEBIT CARD CREDIT CARD PREPAID CARD YEAR ATM POS INTERNET ATM POS INTERNET ATM POS 2014 14.85 1.52 0.04 8.92 4.06 8.29 0.82 0.04 2015 14.89 1.73 0.07 9.86 4.72 7.46 1.05 0.08 2016 15.25 2.10 0.14 11.34 6.34 8.19 0.88 0.10 8   The Retail Payment Costs and Savings in Albania FIGURE 2: Access Points over Time Number of access points over time 8,000 7,111 7,000 6,540 6,689 6,000 5,000 4,000 3,000 2,000 811 826 800 891 1,000 699 448 485 474 0 Bank Branches ATM Agent Outlets Physical POS Terminals 2014 2015 2016 Number of access points (per 100,000 adults) over time 400 352 350 331 317 300 250 200 150 100 39.6 40.8 39.5 44 50 24 23 35 22 0 Bank Branches ATM Agent Outlets Physical POS Terminals 2014 2015 2016 Net growth rate of access points (per 100,000 adults) over time 30% 28.7% 25% 20% 15% 10% 9.1% 6.3% 4.4% 5% 3.0% –4.2% –3.2% 0% 5% Bank Branches ATM Agent Outlets Physical POS Terminals 2014–15 2015–16 The Retail Payment Costs and Savings in Albania   9 FIGURE 3: Active Payment Cards in Circulation over Time Number of active payment cards in circulation over time 900,000 871,611 826,280 800,000 777,195 700,000 600,000 500,000 400,000 300,000 200,000 100,000 73,804 81,125 85,937 27,200 32,873 41,765 0 Debit card Credit card Prepaid card 2014 2015 2016 Number of active payment cards in circulation (per 100,000 adults) over time 40,000 37,090 35,311 35,000 33,072 30,000 25,000 20,000 15,000 10,000 5,000 3,140 3,467 3,657 1,157 1,405 1,777 0 Debit card Credit card Prepaid card 2014 2015 2016 Number of active payment cards in circulation (per 100,000 adults) over time 30% 26.5% 25% 21.4% 20% 15% 10.4% 10% 6.8% 5.0% 5.5% 5% 0% Debit card Credit card Prepaid card 2014–15 2015–16 10   The Retail Payment Costs and Savings in Albania are available for the year 2014. It is also worth noting that COSTS OF RETAIL PAYMENTS BY IV.  in addition to the absolute number of access points cap- ACTOR AND FOR THE TOTAL tured across time, the number of access points per 100,000 ECONOMY adults is also derived, as well as growth rates per 100,000 adults, for more meaningful comparisons.12 Given that the costs are a function of the volume, among other variables, each relevant actor is analyzed by volume The declining net growth rate13 for the bank branches and of transactions, followed by the costs they bear.15 ATMs is a clear signal that commercial banks are moving away from physical access points and toward remote IV.1. CONSUMERS access points and, at the same time, as a result of the recent economic slowdown. While the number of bank Consumers in Albania initiate a total of almost 127.5 million branches (per 100,000 adults) increased by 9.1 percent payments annually (across different payment instruments, during 2014–15, it then decreased by 4.2 percent during use cases, and transmission methods), or 53 payments/ 2015–16. This is significant, especially when considering adult annually, of which 96 percent in cash. Paper-based the similar pattern of ATMs (per 100,000 adults), which direct credit transfers follow, with 2.2 percent of total pay- increased by 3 percent during 2014–15 and then decreased ments initiated. Debit and credit card payments account by 3.2 percent. for 0.6 percent each, while electronic credit transfers account for 0.4 percent and online money for 0.2 percent. Contrary to bank branches and ATMs, the number of However, the total number of payments initiated (particu- agent outlets and POS terminals (per 100,000 adults) has larly through cash) is likely underestimated.16 been increasing in Albania. The number of agent outlets has been growing by 27.5 percent during the last two When the data is analyzed by use case, it is not surprising years. The presence of agents is important for financial that the majority of payments are initiated for groceries inclusion purposes, given that they also tend to cover and necessity goods (64.5 percent, or 34.4 such payments remote areas. At the same time, the agent outlet is asso- annually per adult). Payments for periodic bills and ser- ciated with a large number of cash-in/cash-out processes, vices jointly account for 30 percent of all payments, while raising cost-efficiency concerns. With regard to POS ter- payments for durable goods, government payments, and minals (per 100,000 adults), their net positive growth rate remittances jointly constitute 5.5 percent of all annual ini- has been increasing, from 4.4 percent (2014–15) to 6.3 tiated payments. percent (2015–16). In terms of distribution across transmission methods, pay- In terms of active14 payment cards in circulation, while ments initiated at the point of interaction account for 78.5 there is a positive growth for all types of cards across time, percent, those at a PSP branch account for 11.5 percent, the growth rates have been decreasing, with the excep- tion of prepaid cards. For debit cards, the growth rate has The data are a reflection of the responses collected via the 15.  decreased from 6.8 percent to 5 percent, while for credit surveys and, as such, are as accurate as the responses. cards, it has decreased from 10.4 percent to 5.5 percent. 16. A number of factors contribute to this outcome. First, in 57 While ownership does not necessarily imply usage, the percent of the cases, the respondent is not the person responsible for the household’s finances. In only 23 percent of decrease in the growth rates of active debit and credit the sample, the responsible person is also the household head. cards still raises concerns. One factor that explains the Therefore, he or she might not be in charge of payments of a decrease is the low access to bank accounts, but other periodic nature (such as utility bills), resulting in a lower number of initiated payments being reported. The second reason has to reasons could be further explored by Bank of Albania. Pre- do with the nature of the questionnaire, which relies on paid cards show an uptrend in terms of growth rate, from respondents to self-report the number of payments. In other 21.4 percent to 26.5 percent. words, respondents use their perceptions when recalling the number of payments they make in a typical week, month, or year, and as such, they tend to report fewer payments than they actually made, particularly cash-related. The third reason relates Whenever there is reference to growth rates in the text in the con- 12.  to the sampling. Even though the sample was nationally text of access points, the term refers to growth rates of access representative, consumers who generally initiate a low number of points per 100,000 adults. The following data are used (from the payments might have been overdrawn in the sample, while, Institute of Statistics, Demographics): for 2016, total population conversely, consumers with higher levels of spending might have 2,876,591, adult population (15+ years old) 2,354,481; for 2015, been underdrawn, leading to an overall low number of payment total population 2,875,592, adult population 2,343,985; for 2014, initiated at the nationally extrapolated level. For instance, people total population 2,882,318, adult population 2,348,649. with university and post-university degrees who tend to earn Net growth rate takes into account newly established access 13.  more and therefore initiate more payments constitute jointly only points and access points that were removed. 15 percent of the sample. Education is a random variable, thus Active payment cards refer to debit and credit cards that have 14.  not controlled. been received and activated by their beneficiaries. The Retail Payment Costs and Savings in Albania   11 while the remaining 10 percent of payments are initiated nomic costs of 1 percent of the country’s GDP20 across all at an agent outlet, a paying office, or through the Inter- payment instruments. Cash alone accounts for 0.7 percent net. Internet payments constitute only 1.3 percent of all of the GDP.21 Paper-based credit transfers follow, at 0.12 payments. Further, when the transmission methods are percent of the GDP. It is worth pointing out that these over- matched to the payment instruments, it emerges that 99 all costs are driven by a high per-transaction cost, while oth- percent of all payments initiated at the point of interaction ers are driven by a high volume. For cash, both hold true. are through cash. The remaining one percent of payments are initiated through debit card (0.7 percent) and credit It is not surprising that the vast majority of costs associated card (0.3 percent). Among all bank branch payments, 81 with the use of paper-based instruments are time-based percent are initiated through cash, while the remaining opportunity costs, due to the fact that they require the 19 percent are initiated via a paper-based credit transfer. physical presence of and travel by the payer/payee in Among all Internet payments, 35 percent are initiated almost all instances. Similarly, it is expected that the major- through electronic credit transfer, 31 percent are initiated ity of costs for electronic payment instruments are direct through credit card, 20 percent are made by debit card, monetary costs stemming from the different fees charged. and 14 percent are online money transactions. Further, 10 Finally, with regard to paper-based credit transfers, even percent of all bill payments are initiated online. though a significant portion of the relevant costs are time- based, the fees associated with this instrument overshadow Consumers in Albania receive approximately 38 million the time-based costs. As a result, 90 percent of the paper- payments annually, or 16 payments/adult annually, of based credit transfer costs are direct monetary costs. which 90 percent are received in cash. Of the remaining 10 percent of payments that are not received in cash, 8 As the analysis reveals, Albanian consumers face the high- percent are received via direct credit transfer,17 and 2 per- est aggregate costs when travelling to a bank branch and cent are received via online money. then waiting in line and undertaking a transaction. How- ever, on a per-trip/per-transaction basis, the costliest pay- In terms of use case distribution, the majority of payments ment transmission method is the pay office, followed by are received for goods and services provided to individual the agent office. The pay office demonstrates the highest consumers by other individuals that are not formal busi- cost associated with a cash collection procedure. nesses and account for 28 percent. Salaries and pensions follow, with 23 percent and 20 percent, respectively, while By combining payment instruments with transmission remittances account for 15 percent and social assistance methods, this analysis reveals that paper-based payment benefits account for 14 percent. instruments, regardless of the transmission method, are more costly than electronic payment instruments. For ini- When analyzing the flow of received payments by trans- tiated payments, the costliest combination appears to be mission method, a significant portion is received at the a paper-based credit transfer at the bank branch, while point of interaction (29 percent), followed by PSP branches the least costly is a debit/credit card payment at the point and the ATM, with 21 percent. The Internet, as the only of interaction. It is worth noting that certain transmission exclusive noncash channel, accounts for 10 percent18 of methods do not enable payments by electronic means payments received, the agent outlet accounts for 7 per- (for example, pay office). cent, the paying office for 5 percent. If consumer use cases were matched to payment instruments instead,19 it IV.2. BUSINESSES would result that only 5 percent of all remittances (person to person) are received through the Internet. Albanian businesses initiate approximately 18.2 million payments annually, or 792 payments/business annually, of Overall, given their current use of payment instruments as which 66 percent are in cash. Cash usage varies by busi- described above, Albanian consumers incur annual eco- ness size. While microbusinesses initiate 93 percent of their annual volume of payments in cash, this percentage falls to 47 percent for large businesses. Direct credit trans- On the receiving end, direct credit transfers are always electronic. 17.  18.  It should be mentioned that all payments reported as being fers (both paper-based and electronic) and direct debit received through a credit transfer are linked to the Internet transfers are the second and third most-used instruments, transmission method. However, payments that are sent as direct respectively, regardless of business size. credit transfers from the payer but are obtained as cash from the payee (for example, cashing out a salary or pension that is deposited to the beneficiary’s account from a PSP branch or an ATM) are accounted for as cash and associated with transmission The 2014 GDP. 20.  methods other than the Internet. The 2014 GDP for Albania was estimated at 1.8 trillion lek or 21.  In fact, data does not allow for a detailed match between 19.  $14.5 billion. payment instruments and use cases. 12   The Retail Payment Costs and Savings in Albania FIGURE 4: Annual Costs for Consumers by Payment Instrument as Percentage of 2014 GDP Cash 0.700 Paper-Based Credit Transfer 0.120 Debit Card 0.100 Electronic Credit Transfer 0.060 Credit Card 0.020 Online Money 0.001 0.0% 0.1% 0.2% 0.3% 0.4% 0.5% 0.6% 0.7% 0.8% FIGURE 5: Per-Transaction Costs by Payment Instrument across Transmission Methods Paper-Based Credit Transfer 436 Lek ($3.75) 74 Lek ($0.63) 74.6 Lek ($0.64) Cash 62 Lek ($0.53) 49.5 Lek ($0.43) Electronic Credit Transfer 15.5 Lek ($0.13) Online Money 6.3 Lek ($0.05) Debit Card 5 Lek ($0.04) Credit Card 5 Lek ($0.04) 0 100 200 300 400 500 Initiated Received Deposit Withdrawal At the aggregate business level, the majority of initiated transfers account for half of such payments, followed by payments are for the purchase of retail and capital goods direct debits (27 percent) and then credit and debit card (37 percent of total annual volumes). Twenty percent of payments (at 13 percent and 9 percent, respectively). total payments initiated by businesses represent salaries, while periodic bills account for 14 percent, insurance 13 On the receiving side, 274 million payments, or 2,543 percent, services 10 percent, and other payments, such as payments/business, are received by businesses in Albania taxes and alike obligations, 6 percent. As a result, the point annually. The vast majority of all payments are received of interaction is the most popular transmission method for from microbusinesses.22 It is not surprising that 99 percent business-initiated payments (54 percent, of which the over- of all payments across businesses are received in cash, whelming majority is in cash), while the Internet and PSP given that 47 percent of the business sample is comprised branch payments are equally distributed (23 percent each). Moreover, slightly more than half of their payments are ini- It might seem a paradox that the vast majority of all payments 22.  tiated at the point of interaction. Of those, the overwhelm- are received from microbusinesses, but there are some legitimate ing majority is initiated in cash (almost 97 percent), while reasons why this is the case. First, due to their nature, microbusi- nesses accept many small-value payments on a daily basis. They the rest is initiated in debit and credit cards. Payments initi- are often retail “neighborhood” stores that provide goods and ated at a bank branch and through the Internet have an services, while some large businesses might be factories that do equal distribution of about 23 percent. Within a bank not necessarily transact as frequently. Second, microbusinesses constitute a large share of overall businesses in Albania. Combin- branch, cash accounts for slightly more than paper-based ed with the frequent number of transactions, this fact drives the credit transfers (57 percent to 43 percent). With regard to overall high number of payments that microbusinesses accept payments initiated through the Internet, electronic credit throughout the country. The Retail Payment Costs and Savings in Albania   13 FIGURE 6: Per-Transaction Costs by Transmission Method across Payment Instruments Time spent for a single trip/transaction by transmission method Monetary cost for a single trip/transaction by transmission method (in minutes) Pay Office 42 Pay Office 127 Lek ($1.09) Agent Outlet 33 Agent Outlet 106 Lek ($0.91) Bank Branch 29 Bank Branch 67 Lek ($0.57) ATM 16 ATM 37 Lek ($0.32) Point of Point of 1 2.3 Lek ($0.02) Interaction Interaction 0 10 20 30 40 50 0 30 60 90 120 150 of micro- and small businesses that (i) show the biggest transfers at 870 million lek ($7 million), direct debits at 757 gap in terms of adoption of POS terminals and, at the million lek ($6 million), debit cards at 541 million lek ($4.3 same time, (ii) account for over 90 percent of all payments million), and credit cards at 517 million lek ($4.1 million). received by all businesses in Albania. When breaking Table 23 summarizes the costs for all businesses by pay- down volumes received by business size, it emerges that ment instrument and by fixed and variable cost compo- debit and credit card payments are received by all types nent. Interestingly, in terms of cost of cash relative to total of businesses but occupy a larger share among large busi- costs, microbusinesses bear the highest percentage. nesses. As expected, electronic credit transfers and direct (Cash accounts for 63 percent of the overall costs borne debit transfers are not relevant among micro- and small by microbusinesses.) Medium businesses are ranked sec- businesses, while they are used more heavily by medium ond, given that the cost of cash accounts for 60 percent of and large businesses. Data also indicates that the largest overall costs, followed by small businesses (cost of cash share of all electronic payments received by businesses is accounts for 54 percent) and large businesses (cost of cash represented by electronic credit transfers, not card pay- accounts for 37 percent). ments, which follow in the second place. In total, businesses in Albania incur annual costs equal to The only relevant use case considered on the receiving 0.6 percent of the GDP in the context of receiving and side of business payments is the purchase of goods and initiating payments across all payment instruments, with services. A distinction is made based on transmission cash alone representing 0.3 percent of the GDP or half of methods, point of interaction and Internet. Consequently, total costs. In comparison, the cost of paper-based credit the overwhelming majority of payments across different transfers is estimated at 0.134 percent of the GDP, elec- types of businesses are received at the point of interaction tronic credit transfers is estimated at 0.048 percent, direct (99.6 percent) and in cash (99.5 percent). This was debits is estimated at 0.042 percent, debit cards is esti- expected, given that e-commerce is quite limited in Alba- mated at 0.03 percent, and credit cards is estimated at nia. Medium-size businesses seem to have a larger share 0.028 percent. of payments received through the Internet (7 percent of total payments received), compared with small businesses IV.3. GOVERNMENT AGENCIES (only 0.2 percent) and large businesses (2.5 percent). Of the 0.4 percent of payments received through the Inter- The limited data does not allow for a detailed analysis of net, 79 percent are due to electronic credit transfers, 11 payment flows pertaining to government agencies. How- percent are due to direct debit, 6 percent are due to credit ever, it can be inferred that on the initiation side, govern- cards, and 5 percent are due to debit cards. ment agencies in Albania make approximately 17.5 million payments to consumers annually. (On the other hand, When aggregating the costs (by payment instrument) irre- government-to-business data is not available.) About 50 spective of the size of business, the most costly payment percent of all payments initiated by the government are in instrument is cash. The annual cost of cash for all busi- cash, while paper-based credit transfers account for about nesses in Albania amounts to approximately 5 billion lek 20 percent of all payments initiated, and electronic credit ($40 million), followed by the cost of paper-based credit transfers account for 30 percent. It is estimated that gov- transfers at 2.4 billion lek ($19.5 million), electronic credit ernment agencies receive a total of 1.8 million payments, 14   The Retail Payment Costs and Savings in Albania FIGURE 7: Annual Costs for Businesses by Payment Instrument as Percentage of 2014 GDP Cash 0.300 Paper-Based 0.134 Credit Transfer Electronic 0.048 Credit Transfer 0.042 Direct Debit 0.030 Debit Card Credit Card 0.028 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 mostly from consumers at an annual volume of approxi- GDP. The second largest cost results from electronic mately 1.3 million payments, and about half a million pay- credit transfers, at 0.0035 percent of GDP, while the third ments from businesses. largest cost derives from the use of paper-based credit transfers, at 0.003 percent of GDP. The reason why elec- The government payments analysis follows a different tronic credit transfers are slightly costlier, on aggregate, structure as compared with the consumer and business than paper-based credit transfers is the larger volume of sections, given the difference in nature of these payments. electronic credit transfers being used at the government For example, for many services offered by commercial level in Albania. banks and postal offices pertaining to government pay- ments, the government agencies are not imposed any PAYMENT SERVICE PROVIDERS AND IV.4.  explicit fees. Three types of costs are considered for gov- PAYMENT INFRASTRUCTURE PROVIDERS ernment agencies: time-based costs (the time needed for the execution of the transactions can be considered the In 2016, banks and e-money institutions in Albania pro- time needed to process the payment through the pay- cessed 12 million payments/transfers,23 up from approxi- ments systems of the Bank of Albania), communication mately 10.2 million in 2014 and 11 million in 2015. Of costs (the main communication costs relate to Internet those, 44.7 percent were electronic payments in 2016, as providers and SWIFT communication), and infrastructure compared with 33.4 percent in 2014 and 37.5 percent in costs (the Albanian Treasury System has been financed by 2015. It is imperative that this transition momentum to a World Bank Group loan and maintained by the state electronic payments (at least among noncash payment budget). Cost approximations with regard to cash, paper- instruments) is sustained in the country. based credit transfers, and electronic credit transfers are made, but the total costs are likely to be substantially In terms of volume, credit transfers account for the vast underestimated. In fact, it was pointed out that local gov- majority of payments processed by the PSP/PIP in Albania. ernment payments and bill payments are focused on the However, their share has been decreasing over time (due to municipality of Tirana. Therefore, the share of cash pay- the slow growth rate compared to other instruments) and ments is expected to be higher outside Tirana due to antagonized by payment cards and e-money payments, scarce penetration of banking infrastructure, thus leading which have been on a steep rise. Debit cards in particular to even greater costs (and cost-saving opportunities). In account for the second largest share of all payments. Direct addition, social security payments, which are entirely dis- debits’ share, on the other hand, had a slight increase bursed in cash, could not be surveyed. Thus, it is plausible during 2014–15 and a static pattern during 2015–16. that they present the highest gains from switching to elec- tronic instruments. In terms of growth rate, e-money payments/transfers pro- cessed by the supply side displayed the highest growth The annual cost associated with cash is higher, by a large rate in 2016, about 62 percent, followed by prepaid card difference, than that associated with the other two pay- ment instruments used, paper-based credit transfer and As defined here, payments/transfers included only transactions 23.  electronic credit transfer. Table 3 summarizes the overall made for the purchase of goods/services, government 2014 costs for the government, by payment instrument. obligations, and transfers among parties affected using electronic The annual cost of cash accounts for 0.016 percent of the payment instruments—that is, excluding cash and checks. The Retail Payment Costs and Savings in Albania   15 TABLE 3: Annual Costs for Government Agencies by Payment Instrument (2014) CASH PAPER-BASED CREDIT TRANSFER ELECTRONIC CREDIT TRANSFER Central government — 52 million lek ($416,000) 66 million lek ($528,000) Local government 200 million lek ($1.6 million) — — Public utilities 85 million lek ($680,000) — — Total 285 million lek ($2.5 million) 52 million lek ($416,000) 66 million lek ($528,000) Total as % of GDP 0.0160 0.0030 0.0035 FIGURE 8: Volume of Payments/Transfers Processed by Paper-Based over Electronic Channels over Time Volume of payments/transfers processed by paper-based vs. electronic channels over time 8,000,000 6,819,709 6,866,993 6,639,133 7,000,000 6,000,000 5,361,445 5,000,000 4,127,906 4,000,000 3,414,296 3,000,000 2,000,000 1,000,000 0 2014 2015 2016 Volume of payments/transfers processed by paper-based vs. electronic channels over time (%) 80% 70% 66.6% 62.5% 60% 55.3% 50% 44.7% 40% 37.5% 33.4% 30% 20% 10% 0% 2014 2015 2016 Electronic/Mobile Non-Cash/Non-Cheque Paper-Based payments, at 56 percent. In the same period, debit and over time, but their growth rate has dropped significantly, credit card payments have been growing by 31 percent from 27 percent in 2014–15 to 7 percent in 2015–16. Sim- and 26 percent, respectively. The analysis also shows an ilarly, cash deposits displayed a significant hike between ongoing transition from paper-based/branch-assisted to 2014 and 2015, reaching a growth rate of 94 percent. electronic credit transfers. While in 2014 electronic credit However, that growth rate turned negative between 2015 transfers accounted for 17 percent of all credit transfers and 2016, as cash deposits declined by 14 percent. processed, in 2016 they accounted for 23 percent. Overall, it is estimated that the supply side bears an In 2016, the retail payments supply side in Albania pro- annual cost equivalent to almost 9 billion lek ($154 mil- cessed about 22.5 million cash withdrawals24 and 10.5 mil- lion), or 1.31 percent of Albania’s 2016 GDP, associated lion deposits.25 Cash withdrawals have been on the rise with all examined payment instruments. (For a detailed breakdown of costs for each payment instrument, please 24.  Debit, credit, and prepaid card withdrawals are included here see the annex at the end of the report.)26 Cash stands out since such transactions are treated separately from card payments and are linked to cash transactions and cash costs. Withdrawals also include e-money cash-out at agents of e-money institutions. Even though prepaid cards are examined in terms of payment 26.  Deposits include e-money cash-in at agents of e-money 25.  volume flows and issuance over time, they are not examined in institutions. terms of cost because no such data was provided. 16   The Retail Payment Costs and Savings in Albania FIGURE 9: Volume of Payments/Transfers Processed by Payment Instrument over Time Distribution of payments/transfers volume processed by instrument over time (%) 80% 73.4% 69.5% 70% 64.8% 60% 50% 40% 30% 20% 16.3% 11.8% 13.6% 10.4% 8.9% 9.0% 10% 5.6% 6.1% 6.1% 0.2% 0.0% 0.3% 1.4% 0.3% 2.1% 0% 2014 2015 2016 Prepaid Card E-money Direct Debit Credit Card Debit Card Credit Transfer Growth rate of payments/transfers volume by instrument over time 150% 121.2% 120% 90% 61.6% 60% 56.3% 26.0% 30.9% 30% 22.9% 17.8% 8.7% 8.8% 1.7% 1.7% 0% 2014–2015 2015–2016 Credit Transfer Credit Card Direct Debit Debit Card Prepaid Card E-money as the costliest payment instrument on an aggregate action processed, based on the payment instrument and basis, equivalent to 0.65 percent of the GDP (8.9 billion the transmission method used. The transaction costs are lek or $77 million) and responsible for almost half of all derived by using the variable29 cost components associ- costs. Cash is associated with almost triple the cost of the ated with the different payment instruments and transmis- second most costly payment instrument, the debit card, at sion methods. 0.25 percent of the GDP (3.3 billion lek or $29 million).27 • Processing a paper-based credit transfer is the costliest Credit cards rank third, with 0.21 percent of the GDP (3 of all transactions (550 lek or $4.73). It costs almost nine billion lek or $26 million),28 followed by credit transfers, at times as much as processing the equivalent transfer via 0.1 percent of the GDP (1.3 billion lek or $11.5 million); the Internet (66 lek or $0.57). Processing a paper-based direct debit, at 0.06 percent of the GDP (0.8 billion lek or direct debit, which comes second in terms of per-trans- $7 million); and e-money, which accounts for 0.04 percent action costs (460 lek or $3.96), is also approximately of the GDP (0.5 billion lek or $4.5 million). nine times more expensive than processing an equiva- lent electronic direct debit (50 lek or $0.43). It is of significance to the supply-side industry in Albania to have an understanding of the average cost per trans- • Cash deposit and cash withdrawal processes are asso- ciated with almost half the per-transaction cost of pro- cessing paper-based credit transfers and direct debits,  he fact that debit cards on aggregate are more costly than 27. T credit cards is due primarily to the larger volume of debit ranging from 200 lek ($1.72) to 232 lek ($2.00). Specif- card transactions processed by the supply side, as supposed ically, cash deposits/withdrawals at bank branches are to credit cards. slightly costlier than withdrawals at ATMs, followed by  he reason why debit cards show higher cost than credit 28. T cards on an aggregate basis is the following: Even though cash-in/cash-out processes at agents. credit card transactions are associated with higher per-transaction fees than debit cards, the higher volume of the debit card transactions overshadows the transaction fee 29. A variable cost component refers to that category of cost that difference, thus driving the cost of debit cards higher than changes when the volume and/or value of the transaction that of credit cards. changes. The Retail Payment Costs and Savings in Albania   17 FIGURE 10: Volume of Cash Deposits and Withdrawals over Time Volume of cash deposits and withdrawals over time 25,000,000 22,439,910 20,914,754 20,000,000 16,456,402 15,000,000 12,222,592 10,557,906 10,000,000 6,301,008 5,000,000 0 2014 2015 2016 Growth rate of cash deposits and withdrawals over time 100% 94% 80% 60% 40% 27% 20% 7% –14%% 0% –20% 2014–15 2015–16 Cash deposits Cash withdrawals FIGURE 11: Annual Costs for Payment Service and Infrastructure Providers by Payment Instrument as Percentage of 2016 GDP Cash 0.048 0.65 Debit Card 0.048 0.25 Credit Card 0.048 0.21 Credit Transfer 0.10 Direct Debit 0.06 E-money 0.04 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 • Credit and debit card payments stand in the range of the lower processing cost corresponds to e-money 170 to 188 lek ($1.46 to $1.62) in terms of per-transac- transactions channeled through the mobile phone net- tion processing costs. It is not surprising that credit work and the higher processing cost to the transac- card processing costs are higher than the equivalent tions channeled through the Internet. debit card processing costs, primarily driven by higher interchange fees. The same argument also explains In addition to transaction costs, the analysis also presents the higher processing costs for the card-not-present cost indicators related to various important activities and (that is, Internet) payments. infrastructures pertaining to the supply side of retail pay- ments in Albania. Specifically, it is estimated that the aver- • Per-transaction e-money processing costs are slightly age cost of opening a bank account in the country, from below payment card processing costs in Albania, rang- the industry’s perspective, amounts to 322 lek ($2.8), while ing from 150 to 157 lek ($1.35 to $1.46). Specifically, the equivalent average cost for opening an e-money 18   The Retail Payment Costs and Savings in Albania FIGURE 12: Per-Transaction Processed Cost by Payment Instrument Average cost per transaction processed (in lek) Paper-based credit transfer 550 Paper-based direct debit 460 Cash deposit at bank branch 232 Cash withdrawal at bank branch 220 ATM cash withdrawal 208 Cash-in at agent 203 Cash-out at agent 200 Credit card via internet 188 Credit card at POS 180 Debit card via internet 175 Debit card at POS 170 E-money via internet 157 E-money via mobile phone network 150 Electronic credit transfer 66 Electronic direct debit 50 0 50 100 150 200 250 300 350 400 450 500 550 600 Average cost per transaction processed (in US$) Paper-based credit transfer 4.73 Paper-based direct debit 3.96 Cash deposit at bank branch 2.00 Cash withdrawal at bank branch 1.89 ATM cash withdrawal 1.79 Cash-in at agent 1.75 Cash-out at agent 1.72 Credit card via internet 1.62 Credit card at POS 1.55 Debit card via internet 1.51 Debit card at POS 1.46 E-money via internet 1.35 E-money via mobile phone network 1.29 Electronic credit transfer 0.57 Electronic direct debit 0.43 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 The Retail Payment Costs and Savings in Albania   19 FIGURE 13: Supply-Side Industry Cost Indicators for Albania (2016)30 Average annual cost Average annual cost Average cost of Average cost of of maintaining a of maintaining a opening a bank account: producing a debit card: bank account: POS terminal: 322 lek ($2.8) 900 lek ($7.7) 1,900 lek ($16.3) 12,044 lek ($103.6) Average annual cost Average annual cost Average cost of opening Average cost of of maintaining an of maintaining an an e-money account: producing a credit card: e-money account: off-site ATM: 168 lek ($1.4) 1,000 lek ($8.6) 1,500 lek ($12.9) 475,000 lek ($4,085) Average annual cost of maintaining an e-money agent: 126,000 lek ($1,084) TABLE 4: Total Economy Costs for All Retail Payment Instruments Examined (2016) TYPE OF COSTS TOTAL COSTS PER STAKEHOLDERS RESOURCE COSTS TRANSFER COSTS STAKEHOLDER Demand Consumers 12.2 billion lek 6.4 billion lek 18.6 billion lek side $105 million $55 million $160 million Businesses 7.8 billion lek 2.8 billion lek 10.6 billion lek $67 million $24 million $91 million Government 328.3 million lek 87.8 million lek 416.1 million lek agencies $2.8 million $0.8 million $3.6 million Supply PSPs/PIPs 13.4 billion lek 4.6 billion lek 18 billion lek side $115.2 million $39.6 million $154.8 million 33.7 billion lek Total economy costs $290 million 2.5% of GDP FIGURE 14: Total Economy Costs (Resource) by Payment Instrument as Percentage of 2016 GDP Cash 1.70 Credit Card 0.22 Debit Card 0.21 Paper-Based Credit Transfer 0.20 Electronic Credit Transfer 0.06 Paper-Based Direct Debit 0.04 Electronic Direct Debit 0.03 E-Money 0.03 Online Money 0.001 0.0 0.5 1.0 1.5 2.0 30. This methodology does not gather data that would allow for the calculation of costs associated with the maintenance of bank branches. 20   The Retail Payment Costs and Savings in Albania account amounts to 168 lek ($1.4). Such costs are derived almost half of the joint cost of all paper-based payment by accounting for the over-the-counter time spent to instruments.31 It is also worth noting that a hypothetical open accounts, cash-handling time costs, other overhead complete phase-out of paper-based funds transfers (that costs (for example, marketing, compliance), and the num- is, a 100 percent conversion rate) would lead to total sav- ber of new accounts opened during 2016. ings of 9.5 billion lek or $81.7 million. Specifically, replac- ing paper-based credit transfers with electronic credit transfers could save 8.2 billion lek or $70.5 million. Simi- IV.5. TOTAL ECONOMY larly, in the case of paper-based direct debits, in a hypo- Overall, the Albanian economy bears costs equivalent to thetical 100 percent replacement scenario, savings would 2.5 percent of the country’s GDP associated with the use, amount to 1.3 billion lek or $11.2 million. issuance, and processing of all retail payment instruments. This translates roughly into 33.7 billion lek or $290 million. Hypothetical savings scenarios have been calculated and It is worth noting that, in order to avoid overestimations at are reported in table 5. Three conversion rates are used the economy level, the transfer costs among stakeholders (35 percent, 70 percent, and 100 percent) in order to have been removed from the calculations, since they con- project detailed savings from paper-based to electronic stitute revenue for some stakeholders. As such, only payments, constructing various scenarios. Only variable resource costs for each stakeholder are taken into account resource costs from both the demand and supply side are for the estimation of total economy costs for the different taken into account in order to derive the potential sav- payment instruments. Moreover, given that the demand- ings. It is worth noting that the substitution scenarios are and supply-side costs pertain to different years (2014 and constructed in a way that allows for feasibility. For 2016, respectively), in order to create a common-year instance, in the case of substituting electronic credit basis, demand-side cost figures are projected to 2016, transfer for cash, only branch payments and payments using inflation rates. Thus, all costs that refer to the total made for utility bills and remittances are considered, economy are associated with the calendar year of 2016. given that it is not likely in the current retail payments system that cash payments made at the point of sale Cash accounts for 1.7 percent of the GDP, followed by would be substituted with electronic credit transfers. credit cards, at 0.22 percent; debit cards, at 0.21 percent; Also, since the supply side does not deal with actual pay- and paper-based credit transfers, at 0.2 percent, while all ments in the case of cash, the equivalent of a payment is other payment instruments jointly (electronic credit trans- a cash deposit/withdrawal at a bank branch, which can be fers, direct debits, e-money, online money) account for replaced with an electronic credit transfer; a cash deposit/ 0.17 percent of the GDP in terms of cost. withdrawal at an ATM, which can be replaced with three debit/credit card purchases at the point of sale; or a cash In terms of payment instruments, at the economy level, deposit/withdrawal at an agent, which can be replaced paper-based payment instruments are overwhelmingly with an e-money transfer or purchase.32 the costliest instruments, accounting for about 1.95 per- cent of the GDP annually. About 50 percent of the paper- The figure is derived by using the 70 percent conversion rate for 31.  based cost is borne by consumers, 25 percent is paid by the following set of substitutions: cash-to-electronic credit businesses, 24 percent is borne by PSP/PIP, and 1 percent transfer + paper-based credit transfer-to-electronic credit transfer is paid by government agencies. + paper-based direct debit-to-electronic direct debit. In the calculations, it is assumed that one cash withdrawal from a 32.  bank branch can be substituted with an electronic credit transfer or With regard to electronic payment instruments, which an electronic direct debit, given that that consumers typically cost about 0.55 percent of the GDP, about 55 percent is make cash withdrawals of larger-than-average value at bank branches, and electronic credit transfers are also of larger-than- borne by the PSP/PIP, 30 percent is borne by consumers, average value. As such, it is safe to assume that the entire amount 14 percent is borne by businesses, and 1 percent is borne of money withdrawn would be equivalent to the value required for by government agencies. the electronic credit transfer, thus leading to the one-to-one substitution. In the case of cash-to-debit card and cash-to-credit card, on the supply side, a 1:3 ratio is used. Using Bank of Albania data on the aggregate volume and value of ATM cash withdrawals V.  ECONOMY SAVINGS and debit/credit card payments at the point of sale, the per- transaction values can be derived, leading to the conclusion that the average ATM cash withdrawal value is three times that of an Overall, when substituting cash and other paper-based average debit/credit card POS purchase value. For cash-to-money, payments with electronic payments at the economy level, on the supply side, a 1:1 ratio is used.International experience shows that consumers typically withdraw the entire amount of cash it is estimated that, at hypothetical conversion rates of 35 from the agent that another party sent them through e-money percent and 70 percent, Albania can generate annual sav- transfer. As such, a one-to-one substitution makes sense in this ings ranging from 0.5 percent to 0.9 percent of its GDP, or case as well. For more details, see Banka (2013). The Retail Payment Costs and Savings in Albania   21 TABLE 5: Substitution Scenarios and Savings at the Economy Level SUBSTITUTION SCENARIO/ SAVINGS: 35% CONVERSION RATE SAVINGS: 70% CONVERSION RATE SAVINGS: 100% CONVERSION RATE SAVINGS DOMAIN % OF % OF % OF LEK US$ 2016 GDP LEK US$ 2016 GDP LEK US$ 2016 GDP Cash k electronic credit transfer 3.5 billion 30.1 million 0.252 7.3 billion 62.8 million 0.526 11.8 billion 101.5 million 0.850 Cash k debit card 1.9 billion 16.3 million 0.136 3.9 billion 33.5 million 0.280 7.9 billion 68 million 0.570 Cash k electronic direct debit 1.8 billion 15.5 million 0.130 4.2 billion 36.1 million 0.302 7.7 billion 66.2 million 0.554 Cash k credit card 1.7 billion 14.6 million 0.122 3.5 billion 30.1 million 0.252 7.6 billion 65.4 million 0.548 Cash k online money 1 billion 8.6 million 0.072 2.1 billion 18 million 0.150 3 billion 25.8 million 0.216 Cash k e-money 160 million 1.4 million 0.012 362 million 3.1 million 0.026 681.6 million 5.9 million 0.049 Paper-based credit transfer 2.2 billion 18.9 million 0.160 4.3 billion 37 million 0.310 8.2 billion 70.5 million 0.600 k electronic credit transfer Paper-based credit transfer 760 million 6.5 million 0.055 1.5 billion 12.9 million 0.108 2.7 billion 23.2 million 0.195 k online money Paper-based credit transfer 312 million 2.7 million 0.022 760 million 6.5 million 0.055 1.1 billion 9.5 million 0.080 k e-money Paper-based direct debit 460 million 4 million 0.033 700 million 6 million 0.050 1.3 billion 11.2 million 0.094 k electronic direct debit Paper-based direct debit 296 million 2.5 million 0.020 637 million 5.5 million 0.046 1.2 billion 10.3 million 0.086 k e-money The following hypothetical substitution scenarios are VI. CONCLUSIONS worth highlighting. They assume a 50 percent conversion rate and focus on a subset of instruments, actors, and/or The simulations based on the demand- and supply-side use cases that were considered to have the most relevant analysis suggest substantial improvements in efficiency implications for the Albanian payments system and the can be realized in Albania if cash and other paper-based economy as a whole: payments, or parts thereof, were migrated to lower-cost • Indicatively, in the case of utility bill payments, if 50 electronic retail payment instruments. Since the current percent of current payments were instead conducted cost of paper-based payments at the economy level through electronic direct debits, consumers would amounts to about 1.9 percent of the country’s GDP, sav- save about 750 million lek ($6.5 million), businesses ings from a migration of 70 percent of paper-based to would save about 1 billion lek ($8.2 million), and pay- electronic payments could reach 0.9 percent of the GDP. ment service/infrastructure providers would save While the savings would be absorbed by a number of approximately 1.8 billion lek ($15.5 million). actors, the largest benefits would be associated with con- sumers and businesses, since they also bear the largest • Similarly, in the case of pensions, if half of the current share of costs, followed by the payment service/infrastruc- cash pensions were distributed through electronic ture providers.33 credit transfers, the annual savings would be in the range of 385 million lek ($3.3 million) for consumers, 350 million lek ($3 million) for the government, and While a number of countries are in the process of implementing 33.  this guide, Guyana has already completed the process. The results 500 million lek ($4.3 million) for the supply side. show that the cost associated with cash and other paper-based payment instruments (that is, checks, paper vouchers, and • If half of current sent and received cash remittances paper-based credit transfers) for consumers and government were channeled through electronic credit transfers, agencies, on the initiating and receiving side of retail payments, consumers would save about 300 lek ($2.6 million) amounts to about 2.6 percent of the country’s GDP (approximately $80 million), with 95 percent of this cost attributed to cash (roughly annually, while payment service providers would save $76 million). The findings also indicate that the country could about 800 million lek ($6.7 million). potentially save up to 1 percent of its GDP (approximately $31 million) if about 75 percent the current paper-based consumer and • On the business side, if 50 percent of business-to- government transactions were performed via electronic channels. business cash payments were replaced with electronic In terms of specific use cases, under the assumption of a 50 percent substitution rate, the country could save up to $6 million credit transfers, businesses would save about 750 mil- if utility bill payments were paid via electronic credit transfers lion lek ($6.5 million), while payment service providers instead of cash, and about $2 million if salaries were received via would save about 1.6 billion lek ($13.8 million). electronic credit transfer instead of cash. 22   The Retail Payment Costs and Savings in Albania TABLE 6: Substitution Scenarios and Savings for Indicative Use Cases across Different Actors USE CASE: Payments for grocery and necessity goods (P2B) Substitution scenario: Cash k debit card / Conversion: 50 percent Savings for consumers Savings for businesses Savings for PSP/PIP 120 million lek 600 million leks 400 million lek $1 million $5.2 million $3.5 million USE CASE: Payments for utility bills (P2B) Substitution scenario: Cash k electronic direct debit / Conversion: 50 percent Savings for consumers Savings for businesses Savings for PSP/PIP 750 million lek 1 billion lek 1.8 billion lek $6.5 million $8.2 million $15.5 million USE CASE: Pensions (G2P) Substitution scenario: Cash k electronic credit transfer / Conversion: 50 percent Savings for consumers Savings for government agencies Savings for PSP/PIP 385 million lek 350 million lek 500 million lek $3.3 million $3 million $4.3 million USE CASE: Remittances (P2P) Substitution scenario: Cash k electronic credit transfer / Conversion: 50 percent Savings for consumers Savings for PSP/PIP 300 million lek 800 million lek $2.6 million $6.7 million USE CASE: Remittances (P2P) Substitution scenario: Cash k e-money / Conversion: 50 percent Savings for consumers Savings for PSP/PIP 150 million lek 450 million lek $1.3 million $3.9 million USE CASE: Inventory payments (B2B) Substitution scenario: Paper-based credit transfer k electronic credit transfer / Conversion: 50 percent Savings for businesses Savings for PSP/PIP 750 million lek 1.6 billion lek $6.5 million $13.8 million Note: B2B = business to business, G2P = government to person, P2B = person to business, and P2P = person to person. As the transaction costs of exchanging goods and ser- substantially. However, the volume of these transactions vices are reduced, a more efficient use of electronic pay- is still low compared with the total volume of cash and ments could entail significant benefits to the country’s other paper-based transactions, and the growth rates are economic development and growth. In the case of Alba- still too low to achieve the identified substitution scenar- nia, lower costs might have the potential to boost the ios on their own—that is, without a deliberate interven- reach of noncash and electronic payment services to tion. E-money products are also demonstrating potential. lower-income households and small retailers, thereby The planned implementation of the interbank direct debit improving the efficiency of the national payment system would also contribute to more cost-efficient bill payment as access to and usage of modern payment instruments services. Still, the prevalence of cash may cripple authori- is broadened. Overall, about half of the current cost of ties’ and the market’s commitment and efforts in the direc- using payment instruments could be saved annually in tion of increasing electronic payments. Albania, with such estimates being “net” of other posi- tive externalities, including reduced informality and Since less than 40 percent of the adult population has higher tax revenue. access to a bank account (and less than 200,000 e-money accounts), addressing access to transaction accounts is a The question arises as to what could be done to acceler- first-order priority. The Bank of Albania, banks, and PSPs ate Albania’s progress toward reducing cash usage and, (including through the National Payment System Com- in parallel, increasing electronic payments. It is import- mittee) could explore the viability of providing a basic ant to note that some early developments are pointing transaction account at little or no cost to all individuals in the right direction. The overall use of electronic credit and businesses that do not hold an account. Where basic transfers, direct debits, and payment cards have improved accounts exist, these typically result from market-led ini- The Retail Payment Costs and Savings in Albania   23 tiatives, and to a lesser extent from legal/regulatory Changes in the transmission methods associated with requirements. It is equally important that all transaction paper-based payment instruments appear to be urgent, accounts have a minimum functionality that makes it as they result in the most substantial costs from the per- possible to send and receive payments electronically at spective of the users. For instance, it takes the average little or no cost (and to store value safely). The current Albanian about 42 minutes to get to a pay office, wait in account and payment product offerings might need to line, and make a transaction and about 29 minutes to be reviewed in the light of improving overall design and travel to, wait in line in, and transact at a bank branch. The affordability. Public and private sector strategies should volumes of cash payments handled at the bank branch align to the needs, habits, and barriers to account own- combined with the relative cost of the bank branch as a ership and use of the unbanked and the underserved transmission method suggest that bank customers may (consumers and businesses alike) and identify solutions benefit from banks implementing agent models, thus cut- with measurable targets. Technological and business ting time-based costs, especially in rural areas, and possi- model innovations that leverage ubiquity, are easy to bly also mitigating cultural and trust barriers. Expanding use, and lower trust barriers should be encouraged. In Internet/mobile banking and leveraging e-money for bill parallel, initiatives on access and usage should be com- payments are critical to phase out other costly transmis- plemented by further initiatives to increase the financial sion methods of cash and paper-based payment instru- literacy levels in Albania. ments (for example, pay office). Without universal electronic payments acceptance, The analysis also highlights that banks bear significant Albanians have no or little incentive to own and/or use a costs from payment cards, to a large extent driven by vol- transaction account. On the other hand, since payments ume (especially for debit cards) but also as a result of card are a two-sided market, merchants are less likely to processing fees. Specifically, of all debit card–related accept electronic payments, as low customer demand costs, about 31 percent are dedicated to various variable further drives down interest on the merchant side. This and fixed fees (that is, data processing, license fees, inter- report shows that in Albania, the volumes of transactions change fees), and of all credit card–related costs, about generated by electronic payment instruments as a result the same percentage, 31 percent, also go to variable and of a shift from cash to electronic at the point of interac- fixed fees. Local card processing, under certain circum- tion would be the most significant (consumer-initiated stances, could contribute to decreasing some of these payments in cash at the point of interaction amount to costs, which is operationally strategic for banks (given the 99 million) and could save businesses up to 600 million growth of payment card transactions) and the payments lek ($5.2 million) every year. Government and industry system as a whole (for example, as a result of local gover- incentives (for example, fiscal, monetary incentives, but nance and decision making). also innovative business models underpinned by tech- nology) may prove useful to boost acceptance of elec- Overall, even though cash still dominates and is quite tronic payments at the point of interaction. In this costly for Albania, there are promising signs. However, context, efforts to combat informality will also be signifi- trends alone are not sufficient for changes, and as such, cant in accelerating the shift to electronic payments at deliberate actions are needed by the industry, govern- the point of sale given that informality and cash have a ment authorities, and the Bank of Albania to materialize symbiotic relationship. the necessary changes on the ground that will accelerate the shift to electronic payments. ANNEXES ANNEX A BREAKDOWN OF COSTS FOR THE SUPPLY SIDE BY PAYMENT INSTRUMENT CASH DEBIT CARDS Time-based costs associated with over-the-counter Commercial banks in Albania bear an annual cost of deposits/withdrawals as well as back office cash handling approximately 3.3 billion lek (about $29 million), or 0.25 account for about 35 percent of all annual cash costs. percent of the country’s GDP, associated with debit cards. Overhead costs occupy about 18 percent of all cash costs, Putting aside the overhead costs, which account for 26 and they include marketing/advertising as well as compli- percent of all debit card costs, all types of fees jointly ance.34 Safety costs account for approximately 16 percent account for the largest share, about 38 percent, of all and include vault maintenance, insurance, fraud preven- debit card costs. tion, and losses/theft. ATM maintenance cost is close to Excluding overhead, data processing fees—which are that of safety, accounting for 15.5 percent of all cash cost, per-transaction fees paid to the card network for the facil- followed by currency transportation, which accounts for itation of debit card transactions—account for the single about 14 percent. Currency printing, which is an activity largest share, about 15 percent. Interchange fees (paid to performed at the Bank of Albania, is estimated at about issuing banks) account for 9 percent of costs. Settlement 1.5 percent of all cash costs and includes the printing of fees, paid for both local currency and foreign currency cash and the minting of coins. It should not come as a transactions, account for about 7.3 percent of costs. surprise that this cost component is small compared with License fees, which are fixed fees also paid to the card the other cost components, since the printing and minting network on a monthly/annual basis for card branding, of new currency does not happen with the same intensity account for 6.6 percent of all debit card costs. every year. Overall, about 23 percent of all costs borne by banks are The breakdown also enables the allocation, to some fixed and variable fees paid annually to the card compa- extent, of costs to the different access points. Specifically, nies (or 58 percent of all fees paid), about 9 percent of the ATM maintenance cost of 1.4 billion lek ($12 million), costs are fees (interchange) paid to issuing banks (or 23 accounting for about 15.5 percent of all cash costs, is of percent of all fees paid), and 7 percent of costs are settle- particular interest to the industry, as it stands out as a sig- ment fees for local and foreign currency transactions (or nificant cost component directly linked to ATMs. Similarly, 19 percent of all fees paid). time-based costs (over-the-counter and back office) as well as safety costs and currency transportation costs can In terms of other cost components, IT system mainte- be linked to branch operations. Specifically, these cost nance, which covers software and hardware development components account jointly for about 65 percent of all for interfaces with networks and other centralized infra- cash costs or 5.6 billion lek ($48 million). structures allowing for debit card transactions, accounts for about 8.5 percent of overall cost. Card production costs are in the same range (about 8.2 percent of costs) and 34.  Given that overhead costs pertain to multiple payment include enrollment with the card company, plastic and chip instruments, they are linked to each instrument separately by purchase, and personalization and registration with the using volume of payments processed as a weight to allocate domestic bureau (PayLink). POS management accounts for the respective shares. a slightly smaller share, about 7 percent of the costs, fol- lowed by safety costs (fraud prevention, insurance, losses/ 26 The Retail Payment Costs and Savings in Albania   27 theft) at 6 percent of costs, and loyalty programs provided Other cost components include the IT system mainte- to customers, also at about 6 percent of costs. nance, accounting for about 7 percent of costs; card pro- duction, at 7 percent of all costs; POS maintenance, at The breakdown of the cost components allows for some about 6 percent of costs; loyalty programs, at 6 percent; disaggregation of access point costs. Specifically, given and safety, which is also close to 6 percent of the costs. that interchange fees and data processing fees are slightly higher for card-not-present transactions versus physical As in the case of debit cards, a cost allocation by access POS debit card transactions, and also using the share of point can also be derived for credit cards, using a similar Internet and POS transactions, it could be inferred that of approach. Internet-processed credit card processes all debit card costs, about 5 percent is associated with account for about 24 percent of overall credit card costs, Internet transactions, 55 percent is associated with physi- while physical POS-processed credit card activities (includ- cal POS transactions, and the other 40 percent is over- ing POS maintenance) account for about 26 percent of all head, card production, and loyalty program costs, costs, with the remaining 50 percent being overhead, card associated with overall operations. production, loyalty programs, and safety. CREDIT CARDS CREDIT TRANSFERS Commercial banks in Albania bear an annual cost of Commercial banks bear an annual cost of 1.3 billion lek about 3 billion lek ($26 million), or 0.21 percent of the ($11.5 million), or 0.1 percent of the country’s GDP, asso- country’s annual GDP associated with credit cards. Over- ciated with credit transfers. Aside from the overhead head accounts for about 30 percent of all costs, while cost, which accounts for 33 percent of all costs, the next fees jointly account for roughly 38 percent of all credit largest item is that of processing fees, at 16 percent (for card costs. The remaining costs are split between card interbank credit transfers), followed by IT maintenance production, POS maintenance, IT systems, loyalty pro- cost, at 14 percent; over-the-counter transaction time grams, and safety. All cost items (but not the actual costs) cost, at 12 percent; archiving of paper forms and other for credit cards are identical to those presented under back office paper-related processes, at about 11 per- debit cards.35 cent; issuance (that is, production and transportation to the branches) of the credit transfer paper forms, at 8 per- Overhead cost aside, data processing costs show as the cent; and safety (that is, fraud prevention and losses/ highest ranked cost component, accounting for 13 per- theft), at 7 percent. cent of all costs, followed by license fees and interchange fees, which account for about 11 percent and 8 percent of What is interesting to notice is that, jointly, all activities all costs, respectively. Settlement fees for local currency associated with paper-based credit transfers (over-the- and foreign currency transactions are ranked further down, counter time cost, archiving, and issuance of paper credit after system maintenance and card production, at about transfer forms) account for 30 percent of all costs. In 6.5 percent of all costs. other words, all processes related to paper-based credit transfers cost banks about 404 million lek ($3.6 million) The fees paid to the card company account for about 24 annually. percent of all costs (or 63 percent of all fees paid), while those paid to issuing banks account for about 8 percent of The aforementioned observation regarding paper-based all costs (or 20 percent of all fees paid). Settlement fees credit transfer costs, along with the processing fee data account for about 6 percent of all costs (or 17 percent of and share of paper-based and electronic credit transfers, all fees paid). allows for an attribution of costs to access channels (that is, branch-assisted/paper-based versus electronic/Inter- 35.  Given that certain cost components were reported jointly for both net). Indeed, the branch channel/access point is associ- debit and credit cards, the cost disaggregation for the variable ated with 44 percent of all credit transfer costs (590 cost components (including per-transaction fees, such as data million lek or $5 million), 16 percent with the Internet processing fees, interchange fees, and settlement fees) was achieved by using the volume shares of debit and credit cards. channel (213 million lek or $1.8 million), and the remain- However, for fixed-cost components, such as license fees, ing 60 percent (800 million lek or $7 million) are overhead secondary source information from the industry was used. and safety costs. Specifically, 60 percent of the license fees were allocated to credit cards and 40 percent to debit cards. Overall, in terms of fees, while they account for the same percentage approximately of debit and credit card costs, the breakdowns are different. It should DIRECT DEBITS be further noted that while debit card transactions processed are higher in number compared to credit cards, the per-transaction Commercial banks bear an annual cost of about 800 mil- fees paid for credit cards (that is, data processing, interchange, lion lek ($7 million), or 0.06 percent of the country’s GDP, settlement) are higher for credit cards than debit cards. associated with direct debits. In addition to the 48 per- 28   The Retail Payment Costs and Savings in Albania cent of the overhead cost, archiving the paper form of GDP. Excluding the 43 percent of the overhead cost share, direct debits shows as the single largest cost compo- the remaining 57 percent is split as follows: agent-related nent, accounting for 16 percent of all costs. The IT sys- costs (commission fees and training) accounts for 19.5 tem maintenance accounts for about 10 percent of costs, percent of overall costs, followed by mobile phone net- followed by over-the-counter transaction time for paper- work access and maintenance, with 13 percent of all costs; based direct debit forms, with about 9 percent; issuance online system access and maintenance, with 8 percent of of paper forms, with around 9 percent; and safety, with all costs; back office customer service (addressing inqui- close to 8 percent of all costs. All cost components pre- ries related to e-money), with about 7 percent; safety sented under direct debits are identical to those pre- (fraud prevention, losses/theft, insurance), with 5 percent; sented under credit transfers.36 The only difference is and fees other than agent commission (regulatory fees, that for direct debits there are no processing costs, given audit fee, membership in various associations), with about that only intrabank direct debits are processed as of now 4.5 percent. in Albania. Overall, agents cost e-money institutions about 102 Jointly, all activities associated with paper-based direct million lek or $0.9 million annually, accounting for 19.5 debits (over-the-counter time cost, archiving, and issu- percent of all e-money costs.37 In terms of the other ance of paper direct debit forms) account for 34 percent transmission methods, activities associated with mobile of all costs. In other words, all processes related to phone network processing amount annually to about paper-based direct debits cost banks about 287 million 68 million lek or $0.6 million (13 percent of all costs), lek ($2 million) annually. The costs associated with the while those associated with online processing amount electronic channel (that is, the Internet) account for 10 to about 43 million lek or $0.4 million (approximately percent of all costs (80 million lek or $0.7 million), with 8 percent of all costs). It should be mentioned that the the remaining 56 percent accounting for overhead and higher per-transaction processing cost for online money safety costs. transactions as compared with mobile phone network e-money transactions (derived in one of the previous sections) is driven primarily by the low scale of online E-MONEY money transactions. E-money institutions in Albania bear an annual cost of 525 million lek ($4.5 million), or 0.04 percent of the country’s Not all e-money agent-related costs pertain to e-money, given 37.  that agents are involved in a large number of cash-in/cash-out processes. Indeed, this methodology has allocated part of the 36.  Given that certain cost components were reported jointly for both agent costs under cash (based on the volume of cash-in/cash-out credit transfers and direct debits, the cost disaggregation was processes) and the remaining (based on e-money-related achieved by using the volume shares of credit transfers and direct inquiries) under e-money. debits. ANNEX B INDICATIVE SURVEY QUESTIONNAIRES BOX 2 Sample Questions for the Consumer Survey Section 1: Volume and value of payments In a typical month, do you RECEIVE at least one transfer of funds from relatives and/or friends without 1.  exchanging goods/services? ❑ Yes   ❑ No If yes, proceed with the following questions. Otherwise, continue with the next use case. 2. How often do you receive these payments in a typical month? _____________ times 3. What is the payment instrument(s) via which you receive transfer? ❑ Cash  ❑ Credit transfer  ❑ Mobile money 4. How many of those payments do you receive via the following service channels? BRANCH AGENT INTERNET PHONE _____________  _____________  _____________  _____________ Answers in:  ❑ Number  ❑ Percentage Section 2: Cost elements of payments When you visit the following service points, how long does it typically take you to get there? Please answer 1.  in minutes. BRANCH ATM AGENT PAY OFFICE _____________  _____________  _____________  _____________ How much do you typically pay in fees when you cash out from the following service points? Please answer 2.  in local currency. BRANCH ATM AGENT _____________  _____________  _____________   How long (in minutes) does it take, on average, to initiate a payment via one of the following payment instru- 3.  ments—from the moment you start inputting the data until the payment has been completed? DEBIT CARD CREDIT CARD DIRECT CREDIT TRANSFER MOBILE MONEY _____________  _____________  ___________________  _____________ Section 3: Sociodemographic data Do you have access to an account through which you can make and receive payments? ❑ Yes   ❑ No 1.  2. Do you personally use Internet banking? ❑ Yes   ❑ No 29 30   The Retail Payment Costs and Savings in Albania BOX 3 Sample Questions for the Business Survey Section 1: Volumes, values, and distribution of payment transactions initiated and received How many transactions have you RECEIVED with each payment instrument over the past fiscal year for [insert 1.   use case]? What was the average value per transaction by payment instrument? NUMBER OF MINIMUM MAXIMUM PAYMENT INSTRUMENT TRANS-ACTIONS TRANS-ACTION VALUE TRANSACTION VALUE Cash Debit card Credit card Direct debit transfer Direct credit transfer Mobile money Total Section 2: Cost elements of payments How long does it take to accept a payment at the point of interaction from the time you announce the 1.  amount until the transaction has been completed? Please estimate in minutes/seconds per transaction. DIRECT DEBIT DIRECT CREDIT MOBILE CASH DEBIT CARD CREDIT CARD TRANSFER TRANSFER MONEY _____________  _____________  _____________  ____________  ____________  ___________ How much does your business pay in charges/fees to banks for depositing cash per month? 2.  _____________(in local currency)   On average, how much do you pay monthly for communication charges associated with the acceptance of 3.  the following payment instruments at the point of interaction? DEBIT CARDS CREDIT CARDS MOBILE MONEY _____________  _____________  _____________   Section 3: Business information 1. Company name (optional): ______________________________________________________________________ 2. Contact person (optional): _______________________________________________________________________ 3. Number of full-time employees: ___________________ Number of part-time employees: ________________ 4. Average hourly salary paid over the past fiscal year per employee ______________ (in local currency). The Retail Payment Costs and Savings in Albania   31 BOX 4 Sample Questions for the Government Agency Survey Section 1: Volumes, values, and distribution of payments initiated and received What is the percentage distribution of payments INITIATED with each payment instrument via each service 1.   during the past fiscal year? PAYMENT INSTRUMENT DIRECT DIRECT DEBIT CREDIT DEBIT CREDIT MOBILE TRANSMISSION METHOD CASH CARD CARD TRANSFERS TRANSFERS MONEY Point of interaction Branch ATM Agent outlet Pay office Internet/designated lines Mobile phone network Total 100% 100% 100% 100% 100% 100% Section 2: Cost elements of payments 1. How much per month do you pay to payment service providers (for example, agent outlet, bank branch, and so forth) for collecting payments on your behalf ? __________ (in local currency) Please report how much time you or your employees spend in an average week examining bank account 2.  statements. __________ hours __________ minutes Section 3: Government agency information 1. Agency name (optional): ______________________________________________________________________ 2. Contact person (optional): _______________________________________________________________________ 3. Number of outlets: __________________ 4. Number of payment centers: __________________ 5. Number of bank accounts: __________________ 32   The Retail Payment Costs and Savings in Albania BOX 5 Sample Questions for the Payment Service Provider Survey Section 1: Volumes, values, and distribution of payments processed For debit cards issued by your institution, how many transactions were initiated through each service channel 1.   over the past year, and what was the average value per transaction? AVERAGE TRANSACTION AMOUNT TRANSMISSION METHOD NUMBER OF TRANSACTIONS (IN LOCAL CURRENCY) Point of interaction Branch ATM Agent outlet Pay office Internet/designated lines Telephone/mobile phone network Section 2: Cost elements of payments What are the typical production costs you incur per debit card issued? 1.  COST PER DEBIT CARD AVERAGE NUMBER OF DEBIT CARDS (IN LOCAL CURRENCY) ISSUED PER YEAR Manufacturing Personalization Distribution to customer Personalization was done:  ❑ In-house   ❑ By a PSP/PIP   ❑ By a third party Section 3: PSP information 1. Company name (optional): ______________________________________________________________________ 2. Contact person (optional): _______________________________________________________________________ 3. Number of branches: ___________________ 4. Number of full-time employees: ___________________ Number of part-time employees: ________________ 5. Average hourly salary paid over the past fiscal year per employee ______________ (in local currency). REFERENCES Banka, Holti. 2013. “M-PESA at the Point of Sale: Expanding Financial Inclusion and Reducing Demand for Physical Cash.” Journal of Payments Strategy and Systems 7, no. 4 (winter): 359–69. Demirgüç-Kunt, Asli, Leora Klapper, Dorothe Singer, Saniya Ansar, and Jake Hess. 2018. The Global Findex Database 2017: Measuring Financial Inclusion and the Fintech Revolution. Washington, DC: International Bank for Reconstruction and Development/The World Bank. https://globalfindex.worldbank.org/ INSTAT (Institute of Statistics Albania). n.d. http://www.instat.gov.al/en/ World Bank. 2016. Retail Payments: A Practical Guide for Measuring Retail Payment Costs. Washington, DC: International Bank for Reconstruction and Development/The World Bank. http://documents.worldbank.org/curated/en/255851482286959215/pdf/111216-WP- P155382-PUBLIC-ABSTRACT-SENT.pdf 33