From Cash to Digital Payments: Evidence from a Two-Sided Market Intervention in Tanzania

How to cite

Roessler, Philip; Qi, Shuo; Toth, Russel; Tsai, Tiffany. 2026. "From Cash to Digital Payments: Evidence from a Two-Sided Market Intervention in Tanzania." AEA RCT Registry. AEARCTR-0017662. https://doi.org/10.1257/rct.17662-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Philip Roessler* , Shuo Qi , Russel Toth , Tiffany Tsai

* Primary investigator

Registry ID
AEARCTR-0017662
DOI
https://doi.org/10.1257/rct.17662-1.0
First registered
Last updated
Status
On going
Trial dates
2023-03-11 – 2026-12-31
Countries
Private
JEL codes
O33; D83: O16; L14; O17; L81
Registry record
http://www.socialscienceregistry.org/trials/17662

Abstract

In many emerging economies, cash remains the primary medium of exchange despite widespread access to mobile phones and mobile money. Cash dependence is costly, leaving households and firms vulnerable to theft and loss, and limiting access to formal financial services and credit. One fundamental barrier to the transition from cash to digital payments is a coordination problem, in which merchants’ and consumers’ willingness to switch depends on expectations about adoption by the other side. In a field experiment across 79 markets in Tanzania, we study whether locally targeted, two-sided incentives and cross-sided information shift expectations about market-wide adoption and lead to sustained increases in digital payment use. Leveraging detailed merchant network data, we estimate within-market spillover effects and decompose the channels through which digital payment use spreads. We further explore the role of information and network structure in shaping diffusion dynamics.

Intervention

Note that all interventions are implemented at the market level.

In each selected market, we target existing Lipa kwa Simu (LKS) merchants and consumers to incentivize more frequent use, increase the visibility of digital payments, and shift beliefs about their prevalence and acceptability, helping to address the coordination problem that sustains cash reliance in retail markets.

C (Control): no activation program during the study period.
T1 (two-sided Incentives): recruit and incentivize 4 LKS merchants and 6 LKS consumers per market to increase usage of digital payments.
T2 (two-sided Incentives + Cross-Sided Information / Ambassadors): same as T1 (recruited merchants and consumers receive incentives), but participants are also onboarded as Ambassadors to share relevant information and materials about LKS with other merchants and consumers in the market, aiming to encourage broader adoption and use of digital payments through increased awareness, public signaling, and correcting of misperceptions about the prevalence and acceptability of LKS use.

Experimental design

The study is a market-level randomized controlled trial conducted across 79 retail markets in Tanzania. Markets were randomly assigned to one of three arms:
(i) Control;
(ii) Two-Sided Incentives (T1); and
(iii) Two-Sided Incentives plus Cross-Sided Information (T2).

In treatment markets, the intervention targeted existing users of Lipa kwa Simu (LKS) among both merchants and consumers. The interventions were designed to increase the frequency and visibility of digital payment use within markets.

In T1 markets, selected merchants and consumers received small monetary incentives tied to their actual use of LKS. In T2 markets, the same incentive structure was implemented, and selected participants additionally served as Ambassadors who promoted LKS use through informational materials and peer outreach.

The primary analysis compares markets assigned to either treatment arm (T1 or T2) to Control. We also compare T1 and T2 separately to Control assess the incremental effect of the Ambassador activities.

Randomization occurred at the market level with markets blocked by baseline market size and baseline LKS adoption. Interventions were implemented over approximately eight weeks per market, with staggered start dates. Primary outcomes measure digital payment use at the extensive and intensive margins and beliefs about market-wide adoption.

Design details

Randomization method
Randomization was conducted in office using a computer-generated random assignment in Stata.
Randomization unit
The primary unit of randomization is the market. Each market represents a geographically defined trading area (approximately 100-meter radius) in which merchants and consumers interact regularly. Treatment assignment was conducted at the market level to capture coordination and spillover effects across merchants and consumers within the same retail environment. Within treated markets, subsets of merchants and consumers were randomly selected for participation in incentives and Ambassador activities for implementation and exposure variation.
Clusters
79 markets
Observations
790 merchants and consumers (5 merchants and 5 consumers surveyed per market at endline)
Arms
35 markets in Control; 22 markets in T1; and 22 markets in T2
Primary outcomes
Digital payments use (extensive margin): Whether respondents (merchants and consumers) have recently used LKS. Digital payments use (intensive margin): Number of LKS payments recently made. Coordination beliefs (perceived adoption): Respondents’ beliefs about the prevalence of LKS acceptance and use among market participants (same-side and cross-side).
Secondary outcomes
(A.) Digital payments use during the program period (A1.) Extensive margin: whether respondents used LKS in the past two months (A2.) Intensive margin: number of LKS transactions in the past two months (B.) Beliefs and expectations about future use (B1.) Forward-looking expectations about market-wide LKS acceptance and use over the next year (C.) Program exposure and diffusion indicators (C1.) Exposure to Ambassadors or Rewards participants (C2.) Receipt of LKS-related information or materials (C3.) Peer influence from other merchants on LKS use (D.) Awareness- and proximity-based exposure measures (D1.) Ability to name LKS-accepting merchants (D2.) Geographic proximity to Ambassadors or adopters (D3.) Local density of LKS adoption

Cited by

No citation data available for this registration yet.

Keywords