Banking correspondent agents
Local shops, post offices and kiosks act as licensed agents of a bank, so people in towns with no branch can pay bills, deposit, withdraw and open accounts nearby.
Start small
Partner one regulated bank or credit union with 3 to 5 trusted shopkeepers in an unbanked area for a 6 month pilot, using a point-of-sale device or phone app at low equipment cost. Start with bill payments and benefit cash-out, track agent income and dormancy monthly, then add savings accounts.
The problem
Peripheral and rural communities often have no bank branch, so people travel hours or rely on cash and informal lenders. Agents bring basic financial services into existing local businesses at a fraction of branch cost.
Where it fits
Thrived where the regulator allowed banks to contract non-bank retailers as agents, with the bank keeping the risk, and where government benefit payments and bill payments created steady transaction volume. It needs reliable connectivity, agents with enough cash liquidity, and fair commissions.
The principle underneath
Piggybacking on shops people already visit spreads the fixed cost of a financial access point across an existing business, so services reach places too small for a branch. Regulation that keeps liability with the bank lets trust flow to the agent.
Who to call
CGAP (World Bank-hosted financial inclusion centre); BFA Global (Bankable Frontier Associates); MSC (MicroSave Consulting); Central Bank of Brazil financial inclusion team; Reserve Bank of India business correspondent guidance.
Watch out for
Low commissions and low volumes make agents quit, stranding customers. Agents often stay limited to payments rather than savings and credit. Weak consumer protection and branches that refuse to honour agent-opened accounts erode trust, and some agents substitute for branches instead of complementing them.
Where it comes from
Brazil formalised correspondentes bancários from 1999, with key rules in 2003, 2008 and 2011, reaching every municipality by 2002. India adopted the business correspondent model from 2006 and scaled it through Pradhan Mantri Jan Dhan Yojana from 2014. Regulatory frameworks are public.
Where it's worked
Brazil (all municipalities) · 1999 to 2011
In 1999, 1,679 of 5,626 municipalities had no bank presence; by 2002 every municipality had some financial service, and correspondents grew from 63,509 in 2000 to 165,228 in 2010, becoming the main payment channel.
DocumentedSource
Brazil (national household survey) · 2013
About 400,000 agents; 68% of households paid at least one bill at a correspondent, with the biggest impact in towns under 75,000 people, but only 4% opened an account and 6% accessed credit through one.
DocumentedSource
India (national PMJDY assessment) · 2015 to 2016
10% of 1,627 surveyed agents were dormant; affected customers saved less, travelled 6 to 8 km to branches and about 65% returned to cash or informal savings.
DocumentedSource
Rural India · 2019
Peer-reviewed financial modelling found customer service point kiosks took over seven years to break even due to high costs and low volumes, and proposed a model reaching break-even in about three years.
Peer reviewedSource
Where it failed
In India, MicroSave found 10% of business correspondent agents dormant in 2015, leaving customers with unusable accounts, long trips to branches and a shift back to cash savings. A 2019 study found rural Indian kiosks took over seven years to break even. In Brazil, agents were used mostly for bill payments, with only 4% of people opening an account through one.
Related Fruit
- TreeMember-owned savings and credit groupsNeighbours pool regular small savings into a fund they own and govern, and lend it back to each other, giving people financial services where banks don't reach.
- TreeWomen's self-help group federationsSmall neighbourhood groups of women save together, then federate upward so their members gain a collective voice in local government.