Using Monitors to Increase Savings: Evidence from a Field Experiment

In developed countries, technologies, such as savings reminders and automatic enrollment in retirement savings plans, help individuals overcome common obstacles to meeting their savings goals; however, in developing countries, these solutions are far harder to implement, especially for poor and rural clients.

Emily Breza's research project explores how existing peer networks might help individuals in developing countries overcome common psychological barriers to savings.

Her experiment will open savings counts for villagers in three districts in Karnataka, India to test the effectiveness of different peer and incentive structures. Villagers will decide on a savings goal and will either be in a control group or will be randomly assigned to a savings monitor. Each savings monitor will be informed of the saver's goal as well as their weekly progress towards attaining that goal. A subsample of savers will choose the identity of their savings monitor. In another treatment, savings monitors will receive a small monetary incentive upon successful attainment of the saver's goal. 

This study will be a departure from previous work on this topic since Breza will be able to identify not only average peer effects, but she will also be able to determine those network relationships that create the most effective peer monitors. Furthermore, studies of peer savings mechanisms have never been completed before using high quality network data. Using extensive covariate and network data previously collected from these villages, Breza will also be able to identify those qualities that characterize the most effective savings referees.

Breza's project seeks to advance understanding about the barriers to savings faced by rural and poor households. Breza plans to offer specific policy recommendations for structuring savings devices. She believes peer-level intervention, like savings monitors, has many attractive characteristics as a potential solution for savings. Most notably, peer level financial devices operate at little or no cost to financial institutions, which already have a high fixed cost hurdle in providing services for lower income and rural clients.

Research Questions: 
  • Can peer pressure be harnessed to increase savings balances?
  • How does the identity of the peer monitor impact savings behavior? Is it the case that network characteristics, such as social distance or relative importance, make monitors more effective?
  • What is the optimal savings referee for each individual in the study?
  • What are the policy implications of our results? Can we suggest a better savings product design or an independent, locally run savings account overlay?
Data Notes: 

Survey Size:

  • 1250 individuals from 50 villages

Sample:

  • Individuals from three districts in Karnataka, India

Intervention:

  • Individuals will save with or without a monitor
  • The monitor will be a peer, either randomly selected or chosen by the saver
  • Some monitors will receive incentives, based on the saver's attainment of goals

Data Utilized:

  • Network data and survey data from Banerjee et al. (2010) and Jackson et al. (2010)
  • This network data contains a detailed set of links between individuals on 13 dimensions
  • The survey data also contains information about caste, occupation, household amentities, and geographic information

Experimental Time Frame:

  • June 2011: Train field team and collect baseline data
  • July 2011: Randomize sample into treatments
  • August 2011 - January 2012: Execution of experiment
Partners: 
AgriGold
Status: 
Completed