Governance Structures and the Promotion of the Public Interest: Evidence from Brazilian Resource-Scarce Communities

Last registered on July 19, 2021

View Trial History

Pre-Trial

Trial Information

General Information

Title

Governance Structures and the Promotion of the Public Interest: Evidence from Brazilian Resource-Scarce Communities

RCT ID

AEARCTR-0002765

Initial registration date

March 07, 2018

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

March 07, 2018, 3:53 PM EST

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

Last updated

July 19, 2021, 1:10 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Country

Brazil

Region

Rio de Janeiro

Primary Investigator

Name

Leandro Pongeluppe

Affiliation

University of Pennsylvania

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Anita McGahan

PI Affiliation

University of Toronto

Contact Investigator

Additional Trial Information

Status

In development

Start date

2018-03-16

End date

2022-07-31

Keywords

Governance, Labor

Additional Keywords

JEL code(s)

Secondary IDs

Abstract

The goal of this research is to examine the performance of public, private (NGO) and hybrid (public-private) governance (Mahoney, McGahan, & Pitelis, 2009; McGahan, Zelner, & Barney, 2013; Quélin, Kivleniece, & Lazzarini, 2017) in a resource-scarce environment (Baker & Nelson, 2005; George, McGahan, & Prabhu, 2012). Therefore, this research aims to develop an understanding of how different governance structures deploy resources to promote the public interest by evaluating a job training program targeting residents of Brazilian favelas. The research objective is to analyze alternative forms of commissioning resources under public versus private versus hybrid governance mode. Specifically, we analyze how governance structure influences both the fulfillment of the public interest and the distribution of value among stakeholders engaged in the deployment. Our setting allows a comparison of how much a non-profit organization can foster social prosperity – measured by the increased level of employment (formal and informal), income, the confidence level of individuals, the optimism level of individuals, among other outcomes. The approach is to compare these outcomes between individuals who undertook the training and those who did not. Also, the research design enables us to understand whether the training is more effective when performed through the public organization, the NGO, or through a hybrid form.
References
Baker, T., & Nelson, R. E. (2005). Creating Something from Nothing: Resource Construction through Entrepreneurial Bricolage. Administrative Science Quarterly, 50(3), 329–366.
George, G., McGahan, A. M., & Prabhu, J. (2012). Innovation for Inclusive Growth: Towards a Theoretical Framework and a Research Agenda. Journal of Management Studies, 49(4), 661–683. http://doi.org/10.1111/j.1467-6486.2012.01048.x
Mahoney, J. T., McGahan, A. M., & Pitelis, C. N. (2009). The interdependence of private and public interests. Organization Science, 20(6), 1034–1052. http://doi.org/10.1287/orsc.1090.0472
McGahan, A. M., Zelner, B. A., & Barney, J. B. (2013). Entrepreneurship in the Public Interest: Introduction to the Special Issue. Strategic Entrepreneurship Journal, 7(1), 1–5.
Quélin, B. V., Kivleniece, I., & Lazzarini, S. (2017). Public-Private Collaboration, Hybridity and Social Value: Towards New Theoretical Perspectives. Journal of Management Studies, 54(6), 763–792. http://doi.org/10.1111/joms.12274

External Link(s)

Registration Citation

Citation

McGahan, Anita and Leandro Pongeluppe. 2021. "Governance Structures and the Promotion of the Public Interest: Evidence from Brazilian Resource-Scarce Communities." AEA RCT Registry. July 19. https://doi.org/10.1257/rct.2765-3.4000000000000004

Former Citation

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Intervention Start Date

2018-03-19

Intervention End Date

2018-12-21

Primary Outcomes

Primary Outcomes (end points)

Primary outcome variables include: formal employment, informal employment, wages (income) level, perceived confidence level, perceived optimism level, perceived self-stereotyping level.

Primary Outcomes (explanation)

Primary outcome explanation: Other than Labor Market Outcomes, an Instrument (questionnaire) will measure the perceived confidence, optimism, and stereotype of individuals. The instrument is based on the literature of behavioral economics and social-psychology (Chen, Gully, & Eden, 2001; John & Srivastava, 1999). Finally, the instrument is composed of 23 questions plus a “Ten Item Personality Inventory” questionnaire.
References
Chen, G., Gully, S. M., & Eden, D. (2001). Validation of a New General Self-Efficacy Scale. Organizational Research Methods, 4(1), 62–83. http://doi.org/10.1177/109442810141004
John, O. P., & Srivastava, S. (1999). Big Five Inventory (Bfi). Handbook of Personality: Theory and Research, 2, 102–138. http://doi.org/10.1525/fq.1998.51.4.04a00260

Secondary Outcomes

Secondary Outcomes (end points)

Secondary Outcomes (explanation)

Experimental Design

The NGO will recruit participants from various favelas in Rio de Janeiro city (Brazil) to perform the training program in 2018. The research design is the following:
“Hybrid Mode” (Public and Private Organizations in partnership):
The NGO will recruit individuals interested in participating in the training program with the support of two public social service agencies. Among those interested in the training program, the NGO will select approximately 280 individuals, who will be “candidates for a spot” in the program. These candidates will provide some basic information on socioeconomic dimensions. A stratified randomized assignment (Bruhn & McKenzie, 2009) will follow, i.e., individuals will be stratified considering four elements: i) their age (median age cutoff), ii) their income level (only income source is the Brazilian social security, "Bolsa Familia", cutoff), iii) their social service agency (CRAS) filiation, and iv) the period they want to perform the training (morning or afternoon). Then, within each strata, the randomization will occur. Therefore, "half" of the members of the strata will be allocated to the NGO treatment (training program) and the other "half" to the control. The intervention will then follow. This procedure is the most adequate for the research objective given the sample size limitation (Bruhn & McKenzie, 2009).
“Pure Private Mode” (only NGO)
The NGO will also recruit, at least, additional 280 individuals from other favelas, in at least four different locations. These individuals will also receive the training treatment. However, the selection of these individuals will be performed through a self-selection process, i.e., the individuals will be enrolled in the program up to the point that there are no more spots available, based on a “first-in-first-serve” mechanism. Therefore, by the end of the intervention, the researchers will match each individual from the “Pure Private Mode” with one of the pairs in the “Hybrid Mode” and check whether there is any difference among the distinct modes of governance. Note that in this “Pure Private Mode”, the allocation is not random. Thus possible selection bias exists.
References
Bruhn, M., & McKenzie, D. (2009). In Pursuit of Balance: Randomisation in Practice in Development Field Experiments. American Economic Journal: Applied Economics, 1(4), 200–232. http://doi.org/10.1257/app.1.4.200

Experimental Design Details

Randomization Method

The stratified randomization will be performed on a computer, on March 16th. The NGO will be informed about the result right away. The intervention will then start on March 19th.

Randomization Unit

Individual level randomization.

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

N/A (clustering will not be used)

Sample size: planned number of observations

Up to 280 individuals in the “Hybrid Mode” (RCT), considering 140 treated and 140 control. At least additional 280 individuals in the “Pure Private Mode” (non-randomly selected), will be compared.

Sample size (or number of clusters) by treatment arms

"Hybrid Mode" (RCT)
Up to 140 treated individuals, and up to 140 control individuals.

"Pure Private Mode" (Non-randomized)
Up to 280 treated individuals

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

As our sample size of 280 individuals is fixed (given the NGO’s fixed number of spots), power calculations show that we will be able to detect from 0.15 to 0.25 standard deviations in the best and worst case scenarios, respectively, when controlling for other covariates. The power calculations were performed using the software Optimal Design Plus (Raudenbush, Spybrook, et al., 2011) References Raudenbush, S.W. Spybrook, J., Congdon, R., Liu, X., Martinez, A., Bloom, H., & Hill, C. (2011). Optimal Design Plus Empirical Evidence.

Supporting Documents and Materials

IRB