Communication, Search, and Mobile Phones: A Telephone Directory Intervention in Tanzania
How to cite
Dillon, Brian; Aker, Jenny; Blumenstock, Joshua. 2021. "Communication, Search, and Mobile Phones: A Telephone Directory Intervention in Tanzania." AEA RCT Registry. AEARCTR-0008042. https://doi.org/10.1257/rct.8042-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
The expansion of information and communications technology (ICT) throughout the developing
world is among the most profound instances of technological change in modern economic history.
Yet, for all the promise of this technological transformation, our understanding of the economic
significance of these changes remains surprisingly rudimentary. This is particularly the case
regarding the potential for ICT to provide pathways out of rural poverty.
We propose to study the extent to which maize farmers in Tanzania directly benefit from an
exogenous reduction in the cost of searching for firms relevant for agricultural production or trade.
Our primary intervention is a telephone directory that lists contact information for enterprises. The
rationale is simple: while phones reduce the costs of communicating with known contacts, it is not
clear how households can expand their existing networks without complementary goods that lower
the cost of using ICT to search for new contacts.
Because this intervention generates an exogenous increase in firm exposure, we will
measure impacts on both households that receive the directory and firms that are listed. Our
findings will contribute to the understanding of both smallholder market participation and SME
performance in an era of rapidly changing information and communication technologies.
Intervention
Our primary intervention is a telephone directory that lists contact information for enterprises in eight sectors relevant to agriculture. The RCT is two-sided. We randomized enterprises into trial versions of the directory at the subvillage-sector level, and we randomize distribution of the directory to households at the village level. For the enterprises, treatment is defined as being listed. For the households, treatment is defined as receiving a copy of the directory.
Experimental design
We began this two-sided randomized control trial (RCT) with a census of all enterprises in eight sectors relevant to farming households. The census covered 1,494 enterprises operating in 49 villages and cities in a roughly 5,000 square mile area. We randomly selected a little over half of these enterprises to be listed in a printed phone book. The phonebook was organized by location and sector, and contained information on each enterprise's priority goods and services, which allowed for differentiation and specialization. During the next planting season we distributed 3,290 printed phone directories via community meetings in randomly selected villages. Because of the high likelihood of spillovers through directory sharing, we randomized distribution at the village level. The experiment then ran for nearly one year, covering a full cycle of agricultural activities.
Design details
Cited by
No citation data available for this registration yet.