Community Cellular Networks: An Experimental Evaluation in the Philippines

Last registered on August 07, 2019

View Trial History

Pre-Trial

Trial Information

General Information

Title

Community Cellular Networks: An Experimental Evaluation in the Philippines

RCT ID

AEARCTR-0004469

Initial registration date

August 07, 2019

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

August 07, 2019, 5:05 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country

Philippines

Region

Luzon

Primary Investigator

Name

Niall Keleher

Affiliation

University of California, Berkeley

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Niall Keleher

PI Affiliation

University of California, Berkeley

Contact Investigator

PI Name

Erin Troland

PI Affiliation

Federal Reserve Board of Governors

Contact Investigator

PI Name

Arman Rezaee

PI Affiliation

University of California, Davis

Contact Investigator

PI Name

Joshua Blumenstock

PI Affiliation

University of California, Berkeley

Contact Investigator

Additional Trial Information

Status

On going

Start date

2016-07-01

End date

2020-06-30

Keywords

Agriculture, Other, Welfare

Additional Keywords

Communications, Mobile Phones, Rural Markets

JEL code(s)

D83, L86, O18, O31

Secondary IDs

Abstract

Despite the rapid expansion of mobile coverage in the developing world, roughly 10% of the world's population still lives beyond the reach of a cell tower and all of the platforms, products, and services that cell towers make possible. This study investigates the impact of a new type of cellular connectivity designed specifically to address the "last mile" problem: the Community Cellular Network. Built on open-source technology, the Community Cellular Network provides community-owned cellular infrastructure at one-tenth of the cost of traditional mobile towers.
Our goal is to develop a deeper understanding of the social and economic impacts of first-time access to a cellular network in isolated communities in the Philippines. We use a two-stage randomized control trial in which we (i) randomly assign villages to receive or not receive a new cell tower; and (ii) within treated villages, randomize the price that households paid to access these networks. This pre-analysis plan describes how we plan to study the impacts of cellular access on social network structure, and several social and economic outcomes.

External Link(s)

Registration Citation

Citation

Blumenstock, Joshua et al. 2019. "Community Cellular Networks: An Experimental Evaluation in the Philippines." AEA RCT Registry. August 07. https://doi.org/10.1257/rct.4469-1.0

Former Citation

Blumenstock, Joshua et al. 2019. "Community Cellular Networks: An Experimental Evaluation in the Philippines." AEA RCT Registry. August 07. https://www.socialscienceregistry.org/trials/4469/history/51405

Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Experimental Details

Interventions

Intervention(s)

1. Community Cellular Networks
The primary intervention is the installation of cellphone towers in selected sites. Researchers from the University of California, Berkeley developed a new mobile phone technology --- the Community Cellular Network (CCN) --- that provides local coverage at one-tenth of the cost of traditional mobile towers. The CCN is designed specifically for rural settings with intermittent power and is designed to be owned and maintained by local community members with modest technical training (Heimerl and Brewer, 2010). Over the past several years, we have been collaborating with those who developed the CCN and researchers at the University of the Philippines, Diliman, to deploy CCNs in several remote and isolated villages in the Philippines.

2. Price Promotions
We offered promotions to customers through a second randomized experiment. The first dimension is a free credit of 100 Pesos loaded directly to the customer's balance. The second dimension is a type of tariff discount provided to customers. 50% tariff discounts were applied to either local on-network calls or long-distance calls. All customers, including those in the promo control group, received five free long-distance text messages. To account for potential sharing of phones within the household, treatment assignment was done at the household level.

Intervention Start Date

2016-09-01

Intervention End Date

2019-09-30

Primary Outcomes

Primary Outcomes (end points)

We will test four families of hypotheses: (i) access to communications, (ii) social networks, (iii) informedness, and (iv) economic outcomes. Each family contains one or more specific outcomes that we hypothesize are likely to be impacted by treatment.

Within the family access to communications, we have one hypothesis:
1. Increased access to communications

Within the family social networks, we have two hypotheses:
2. Increased social connectedness - local
3. Increased social connectedness - long-distance

Within the family informedness, we have two hypotheses:
4. Increased informedness
5. Increased disaster preparedness

Finally, within the family economic outcomes, we have ?ve hypotheses:
6. Increased market access
7. Changed migration
8. Changed remittances and risk-sharing
9. Changed income and expenditures
10. Changed subjective well-being

Primary Outcomes (explanation)

See pre-analysis plan for detailed explanations.

Secondary Outcomes

Secondary Outcomes (end points)

We have two secondary families of outcomes

The first is political economy, which has hypotheses:
1. Increased political participation
2. Increased political manipulation
3. Increased political-economic knowledge

The second is intra-household decision-making, which has hypotheses:
4. Increased role of women in income decisions
5. Increased role of women in fishing/agriculture decisions
6. Increased role of women in control of resources
7. Decreased spousal difference in reported fish/crop prices

Secondary Outcomes (explanation)

See pre-analysis plan for detailed explanations.

Experimental Design

By leveraging a new low-cost technology, and by cultivating strong local relationships, we created a truly unique opportunity to conduct a randomized control trial of Community Cellular Networks in a set of 14 geographically isolated villages in the Aurora province of the Philippines. The randomized control trial had two stages. The first was a village-level treatment, where a "staggered roll-out" design randomly determined which seven of the 14 villages received a CCN between September 2017 and September 2019, as well as the timing of each installation. The second stage was a household-level treatment, through which we reduced the cost of making phone calls and sending text messages to specific households. This second stage will allow us to precisely estimate the demand for cellular connectivity (and all that comes with it), and will create random household-level variation in communication costs, which we can in turn use to identify the impact of connectivity on other outcomes.

Experimental Design Details

Randomization Method

Community Cellular Networks - Matched-pairs randomization done on a computer using the R package, nbpMatching, from Beck, Lu, and Greevy (2016)
Price Promotions - Random assignment to price promotions done in Stata.

Randomization Unit

Community Cellular Networks - randomized at the village/site level
Price Promotions - Randomized at the household level

Was the treatment clustered?

Yes

Experiment Characteristics

Sample size: planned number of clusters

Community Cellular Networks - 14 sites
Price Promotions - 1131 eligible households

Sample size: planned number of observations

Community Cellular Networks - 2,370 households Price Promotions - 6,613 adults

Sample size (or number of clusters) by treatment arms

Community Cellular Networks -
1. Treatment (7 sites)
2. Control (7)

Price Promotions -
1. No promotions (187 households)
2. Free phone credit (191)
3. Local discount (183)
4. Local discount plus free credit (186)
5. Long distance discount (191)
6. Long distance discount plus free credit (193)

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

IRB