Academic Jobs - Home of Higher Ed Logo

Alternative delivery channels: 10 Research Papers exploring ADCs

8 min read0 vues

At a glance

Ten studies on alternative delivery channels, from Tanzania and Kenya to agent networks in Indonesia and Nigeria, and what they show about mobile money, service quality and financial inclusion.

a man wearing a hard hat
Photo by Fotos on Unsplash

Alternative delivery channels don't reliably lift customer satisfaction, and on some service dimensions they weaken it. That is the finding of the most recent paper in a decade-long run of research on ADCs, published in the Journal of King Abdulaziz University: Islamic Economics.

The study surveyed 680 customers of Islamic and conventional banks in Tanzania. At conventional banks, ADCs significantly weakened the link between satisfaction and both tangibility and assurance. At Islamic banks, they weakened reliability and empathy.

For bank managers, that complicates a decade of channel spending. For regulators in markets where agents now outnumber branches, it raises a sharper question: what happens to service quality when the branch stops being where banking happens?

What counts as an alternative delivery channel

The label covers a wide range of infrastructure, and the papers don't always agree on where it stops. The International Finance Corporation's 2014 handbook, written for financial services providers, defines ADCs as channels that expand the reach of services beyond the traditional bank branch. The handbook is a project guide rather than a study: channel strategy, platform components, vendor selection, implementation. It was written when an estimated 2.5 billion adults had no account.

The Mastercard Foundation and BFA sharpened the definition in their 2016 review of African banks and microfinance institutions. An ADC, they wrote, is any channel that is not a full-service brick-and-mortar branch staffed by the provider's own employees. That takes in ATMs, agents, mobile banking, online banking, call centres, limited-service branches and roving staff.

Grouping a cash machine with a shopkeeper holding float for a telco is analytically awkward. It's part of why the findings below don't line up neatly.

Tanzania: channels don't fix service quality on their own

Mustafa Omar Mohammed, Mohamed Cherif El-Amri and Yussuf Charles Yussuf collected five-point Likert responses from 680 customers and ran structural equation modelling in AMOS. They tested six service-quality dimensions drawn from the CARTER model, which adds Shari'ah compliance to the usual service measures.

Nearly all six dimensions were positive and statistically significant. Responsiveness was the exception at conventional banks. The interesting result sits in the interaction terms. ADCs moderated the service-quality relationship in both bank types, and the moderation ran negative where it was significant.

The Shari'ah compliance path was flat: a standardised estimate of 0.001, with a p-value of 0.965. The authors call that counterintuitive against international evidence, and point to earlier work showing many Tanzanian Muslim customers keep using conventional banks for reasons of awareness and product availability rather than conviction.

Their policy recommendation is blunt. Tanzania needs infrastructure investment and a comprehensive legal and regulatory framework for ADCs. The country's central bank counted more than 66 million ATM transactions and close to 60 million mobile banking transactions in 2020, which is the scale at which a thin rulebook becomes a consumer problem.

The paper is published in volume 39 of the journal.

Kenya: a channel-by-channel ranking from the banks themselves

Maurice Malenge's study at Kisii University took a census rather than a sample: 86 IT and strategy directors at all 43 commercial banks operating in Kenya, surveyed at head offices in Nairobi. Each director rated how far a channel contributed to financial performance.

Agency banking came top, internet banking last, and all four channels cleared the 5% significance threshold.

ChannelStandardised coefficientt-value
Agency banking0.3755.092
ATMs0.3455.109
Mobile banking0.2353.806
Internet banking0.1963.460

The recommendations are unusually practical for this literature: put more transactions on the ATM, extend what agents can do, tighten security because customers avoid internet banking over fraud fears, and make mobile banking move money rather than just display balances.

One caveat applies to most of the survey work here. These coefficients measure what senior bankers believe, not audited margins by channel, and the thesis shares that limit with the wider field.

a man holding a stack of boxes

Photo by Fotos on Unsplash

Nigeria: agents and POS move inclusion, internet banking doesn't

Quadri, Malik-Abdulmajeed, Akinwumi and Omotosho took a different route. Rather than asking bankers, they ran ordinary least squares on secondary data from 2013 to 2021, drawn from the World Bank's Global Findex and the Central Bank of Nigeria's e-payment statistics.

Point-of-sale and mobile money operations had a positive and significant effect on financial inclusion at the 5% level. Web and internet banking played a limited role. The authors conclude that more off-site ATMs and licensed agents are the practical route to rural residents, and publish the work in The International Journal of Applied Business.

Read beside the Kenyan study, the pair makes a point about what is being measured. Kenya's dependent variable was bank performance, where internet banking still registers. Nigeria's was inclusion, where a browser-based service reaches almost nobody without a laptop, a data plan and a formal address.

The practitioner literature: three documents worth reading alongside the journals

MIX surveyed 217 financial services providers in early 2017 and followed up with 26 interviews across Ghana, Tanzania and Uganda. Agents were the most commonly deployed channel, yet 55% of the experiments and future plans reported involved mobile phones or the internet. Providers named five goals most often: cutting costs, improving convenience, reaching new clients, mobilising deposits and diversifying revenue.

The value of that work is its metrics framework. MIX tested each indicator against three conditions: that it supports a decision, that providers can report it without new systems, and that it can be benchmarked against other deployments. Plenty of channel dashboards fail all three.

The Mastercard Foundation review, produced with BFA, is the clearest statement of what customers actually get. Convenience, lower fees and the comfort of not carrying cash to a branch. Its list of barriers is equally concrete: network downtime, patchy connectivity and agents running out of cash.

Where the agent model breaks

CGAP and MSC put numbers to the viability problem in Indonesia. Of roughly 64,000 agents that would be needed to serve excluded populations in rural frontier areas, about 90% were judged likely to face economic or operational trouble. Three-quarters of agents earned less than $20 a month; 21% failed to break even.

Liquidity is the recurring failure. Sixty-three percent of agents reported barriers to managing cash, and around 30% of active agents faced working capital shortfalls of IDR 5 million to 15 million, or roughly $300 to $1,000. Per-agent setup and support cost banks about IDR 4.5 million a month, against IDR 300,000 for e-money providers running lighter systems.

India shows the other side of the ledger. Business correspondent outlets grew from 34,621 in 2010 to 988,299 in 2019, helped by a mandate for a banking access point within five kilometres of every village and by Aadhaar-enabled e-KYC, which logged 799 million successful identity checks and an estimated direct saving of $1.5 billion. The same policy push produced a spike in dormant accounts, because access was supplied faster than demand.

The cost arithmetic behind the channel shift

CGAP, working with the Inter-American Development Bank and the consultancy Akya, priced the shift for banks running agents, branches and remotely managed ATMs at the same time. Agent transactions cost roughly $0.27 to $0.58 each, about half the cost of a branch or ATM transaction at low volumes.

At one major Mexican bank, moving 20% of transaction volume onto the agent network was estimated to be worth 30 basis points of return on equity. CGAP has also put the capital cost of an agent channel at about one-thirtieth of building a branch network, which explains why the model spread before the evidence on service quality caught up.

a man holding a box

Photo by Fotos on Unsplash

The adoption numbers the papers are chasing

The environment has changed since most of this research was written. The World Bank's Global Findex, based on surveys of about 145,000 adults across 141 economies, puts account ownership at 79% of adults worldwide, and 75% in low- and middle-income economies. In Sub-Saharan Africa, 40% of adults held a mobile money account in 2024, up from 27% in 2021.

GSMA counted 2.1 billion registered mobile money accounts in 2024, with 514 million active on a 30-day basis, and 28 million registered agents, of which 10 million were active monthly. There were 755 registered agents per 100,000 adults, double the ratio in 2021. Transaction volumes hit 108 billion, worth $1.68 trillion.

Against that, 1.3 billion adults still have no account, and the World Bank notes that nearly 900 million unbanked adults own a phone. The Findex analysis treats digital payments as the entry point rather than the end state.

The gap the next round of papers has to close

Divide active accounts by registered ones in the GSMA data and roughly three-quarters of accounts sit idle in any given month. That ratio has held even as the absolute numbers climbed, which suggests registration campaigns are outrunning use.

No paper in this set links agent-level economics to customer-level service quality. The Tanzania study measures satisfaction without costing the channel. The CGAP and MSC diagnostics cost the channel without measuring satisfaction. The Kenyan thesis asks directors, not customers.

Both facts stand on their own. More people are being reached than at any point in the history of retail finance, and many of the counters doing the reaching are marginal businesses run by people earning under $20 a month. The research has counted the first. It hasn't yet explained how the second survives.

  • alternative delivery channels
  • ADC research
  • mobile money
  • agent banking
  • financial inclusion
Portrait de Sarah West
A propos de l'auteur

Sarah WestVoir auteur

Academic Jobs In House Author

Sarah West is a Customer Relations & Content Specialist at AcademicJobs, fostering excellence in research and teaching as a specialist in academic trends and professional development. Holding a doctorate in Higher Educa…

Discussions

Sort par :

Soyez le premier à commenter cet article !

vous

Vous serez invité à vous connecter avant de publier votre commentaire.

Browse par faculté

Browse par sujet

Frequently Asked Questions

🏦What is an alternative delivery channel (ADC)?

An alternative delivery channel is any route to a financial service that isn't a full-service branch staffed by the provider's own employees. The definition used by the Mastercard Foundation and BFA covers ATMs, agents, mobile banking, online banking, call centres, limited-service branches and roving staff. The International Finance Corporation describes ADCs as channels that extend services beyond the traditional bank branch.

📄Which research papers on alternative delivery channels should I read first?

Start with the Tanzanian study in the Journal of King Abdulaziz University: Islamic Economics, which tests whether ADCs change the relationship between service quality and satisfaction across 680 bank customers. Follow it with Maurice Malenge's Kisii University thesis on 43 Kenyan commercial banks, which ranks channels by their contribution to performance, and the MIX survey of 217 providers, which is the best guide to measurement.

📉Do alternative delivery channels improve customer satisfaction?

Not automatically, and not evenly. The Tanzanian study found that ADCs significantly weakened the effect of tangibility and assurance on satisfaction at conventional banks, and weakened reliability and empathy at Islamic banks. Both bank types did use ADCs to strengthen responsiveness. The authors conclude that channels alone don't substitute for service quality.

📊Do ADCs improve bank profitability?

The evidence points that way, with caveats. Directors at all 43 Kenyan commercial banks rated agency banking highest for contribution to performance, followed by ATMs, mobile banking and internet banking. CGAP's cost work found agent transactions run at roughly $0.27 to $0.58 each, about half the cost of branch or ATM transactions at low volumes. Those Kenyan figures are perceptions, not audited margins.

🛒What is agent banking and how does it work?

Agent banking uses third parties, usually shopkeepers or small business owners, to handle deposits, withdrawals and bill payments on behalf of a licensed institution. Agents hold float, verify identity and are paid commission. The model costs far less to build than a branch network, which is why it spread quickly across Sub-Saharan Africa and South Asia.

💵Why do so many agent networks struggle to stay viable?

Transaction volumes are thin and cash management is expensive. In Indonesia, CGAP and MSC found three-quarters of agents earned under $20 a month and 63% reported barriers to managing liquidity. Around 30% of active agents faced working capital shortfalls of roughly $300 to $1,000. Providers that visit agents less than weekly lose visibility on both problems.

📱How does mobile money relate to alternative delivery channels?

Mobile money is usually treated as one ADC among several, but it now drives most of the growth in account ownership in low- and middle-income economies. The World Bank's Global Findex puts mobile money ownership at 40% of adults in Sub-Saharan Africa in 2024, up from 27% in 2021. GSMA counted 2.1 billion registered mobile money accounts globally that year.

⚖️What do regulators need to do about ADCs?

The Tanzanian authors call for infrastructure investment and a comprehensive legal and regulatory framework covering ADCs, arguing that weak rules and poor network coverage constrain what channels can deliver. India's experience shows what a regulatory push can do: business correspondent outlets grew from 34,621 in 2010 to 988,299 in 2019, helped by a rule placing a banking access point within five kilometres of every village.

💻Is internet banking as useful as mobile money for financial inclusion?

In Nigeria, it isn't. A 2024 study using 2013 to 2021 data from the Global Findex and Central Bank of Nigeria statistics found point-of-sale and mobile money operations had significant positive effects on inclusion, while web and internet banking played a limited role. The authors recommend more off-site ATMs and licensed agents to reach rural residents.

📐How should a bank measure whether its channels are working?

MIX set three tests for any indicator: it has to support a decision, be reportable without new systems, and be benchmarkable against other deployments. In practice that means tracking transaction volume and value by channel, deposits held on accounts opened through agents, and channel-level direct costs, rather than counting registered customers alone.

🔍What is missing from the research on alternative delivery channels?

Two things. No paper in the current set links the economics of an individual agent to the service quality a customer receives. And the activity data is thin: in the GSMA figures, roughly three-quarters of registered mobile money accounts are not active in a given month, which means registration counts overstate real use.