Managers at a mobile-money business in Ghana underestimated the incentive contract that performed best for the company, according to a field experiment described in an October research revision. The authors’ risk-adjusted assessment favored other contracts for agents, showing why company returns need to be examined alongside how pay and risk are shared.
Francis Annan and Collin B. Raymond worked with MTN MobileMoney. Their paper, Beliefs Over Contracts, revised on October 24, compares managerial predictions with contracts tested from May to July 2025. It is a working paper, not a peer-reviewed finding.
Administrative records covered about 2,300 participating agents in 473 communities. A franchising-style contract ranked fourth of five in managers’ aggregate predictions but led the researchers’ company-profit comparison. Agents paid a fee in exchange for a higher commission on transactions. The estimated revenue gain was modest, and the division of returns was central to the results.
The paper’s profit measure concerns transaction revenue after agent payments, rather than MTN’s consolidated earnings. Its preferred accounting also models uncapped payment terms and recovery of agents’ debt at exit. Agent returns were lower under that approach; this should not be read as a measured reduction in cash disbursements.
What an incentive contract changes
An incentive contract determines how compensation responds to performance. A commission pays a share of transactions or sales. A threshold bonus begins after a target is reached. A tournament rewards performance relative to other participants.
Those arrangements can create different decisions even when a business starts with the same expected compensation budget. Equal expected spending does not guarantee equal payments after people change their behavior or leave.
Consider a hypothetical sales team. A target-based bonus may encourage extra effort near the cutoff, while leaving someone who considers the target unreachable with little additional reason to pursue it. A ranking-based reward depends partly on colleagues’ performance. A fixed payment offers greater certainty but does not reward another transaction directly.
These examples explain the design choices. They are not estimates of how every sales team would respond.
Why the agent network matters
Mobile-money agents connect cash and digital balances, allowing customers to deposit or withdraw money. Their availability therefore matters even when the underlying payment system operates electronically.
The GSMA’s 2026 industry report puts worldwide mobile-money transaction value at $2.1 trillion in 2025, with 2.3 billion registered accounts and 593 million accounts active within 30 days. Accounts are not unique people: one person can hold more than one.
Those global figures do not describe the experiment’s effects. They show the scale of a sector in which companies depend on local service networks as well as software.
Assess the division of returns
For a business evaluating compensation, additional revenue and additional profit are different outcomes. A scheme can increase transactions while paying out enough extra compensation to reduce the amount retained by the company.
The reverse is also possible: a company may retain more largely because the commercial arrangement shifts payments or risk toward the people delivering its service. That requires a separate assessment of agent earnings, participation and the ability to withstand a poor trading week.
MBN’s reporting on how work organization affects employee training examines another case in which management arrangements influence the value of an investment. Compensation similarly needs an operating explanation, rather than an assumption that a more elaborate scheme will work better.
The Ghana experiment does not identify a contract suitable for every employer, franchise or platform. Its useful lesson for managers is narrower: record the expected effects before a trial, then compare revenue, compensation and participation on the same basis. A plan can beat expectations for one side of a business relationship while leaving the other side worse off.
