Can diversification reduce small business vulnerability?

How to cite

Nardi, Leandro. 2019. "Can diversification reduce small business vulnerability?." AEA RCT Registry. AEARCTR-0005181. https://doi.org/10.1257/rct.5181-3.0

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

Leandro Nardi*

* Primary investigator

Registry ID
AEARCTR-0005181
DOI
https://doi.org/10.1257/rct.5181-3.0
First registered
Last updated
Status
Completed
Trial dates
2019-10-28 – 2021-03-01
Countries
Brazil ()
Registry record
http://www.socialscienceregistry.org/trials/5181

Abstract

This study investigates whether and how diversification in small businesses reduces the vulnerabilities to which poor entrepreneurs are exposed. We indirectly (and randomly) change the level of diversification of a group of small businesses in Brazil by increasing entrepreneurs' ability to sell through a platform. Treatment thus consists of online training sessions aimed at increasing entrepreneurs' platform-related sales and marketing capabilities. Main analyses focus on how treatment impacts platform sales and core-business sales (to grasp potential diversification effects). Other analyses consider how treatment impacts the diffusion of platform sellers within geographic regions.

Intervention

All businesses in the sample are, to some extent, diversified since they sell both core business products/services and platform services, such as bill payment, phone recharges. Businesses randomly assigned to treatment receive platform-related sales and marketing training to increase the proportion of platform sales (i.e., revenues from the diversified source) vis-a-vis total business revenues.

Experimental design

About half of businesses were assigned to treatment. Treatment was stratified in two levels: industry/sector and preexisting platform sales experience (i.e., a group with higher number of platform transactions in the nine months preceding the beginning of the trial versus a group with lower number of platform transactions in the same period). To avoid contamination, most of the businesses in the study are located in separated geographic areas. This design also allows for a ZIP-code-level analysis of treatment effects on platform diffusion.

Design details

Randomization method
Randomization done in office by a computer.
Randomization unit
Individual business.
Clusters
315 businesses
Observations
306 businesses
Arms
151 businesses treatment and 155 businesses control
Primary outcomes
The primary outcomes of interest are platform sales (diversified source), total business sales (including non-diversified source) and the ratio between these variables. Business average profits are also examined, as well as the number of platform sellers within ZIP codes.

Cited by

No citation data available for this registration yet.

Keywords