Mobile-izing Savings with Automatic Contributions: Experimental Evidence on Dynamic Inconsistency and the Default Effect in Afghanistan∗

Automatic payroll deductions represent one of the most effective means of increasing savings in developed countries. We design and experimentally evaluate a mobile phone-based automatic payroll deduction system in Afghanistan, a country with limited formal financial infrastructure. We find that employees initially assigned a default contribution rate of 5 percent are 40 percentage points more likely to contribute to the account than individuals assigned to a default contribution rate of zero. We also randomize employer matches, and estimate that a 50% match would be required to replicate the default effect through financial incentives alone. To better understand why the default effect is so pronounced, we conduct several additional experimental interventions designed to induce employees to make a non-default selection, and use behavioral protocols to elicit measures of time inconsistency. We conclude that the effect of default assignment is driven largely by dynamic inconsistency, and specifically by the tendency to procrastinate over the task of finding an optimal plan. ∗Authors’ Note: We are grateful to Karim Khoja and the Roshan team for partnership in the design and implementation of the M-Pasandaz mobile savings product. We thank Mohammad Isaqzadeh, Shahim Kabuli, Nasir Mahmoodi, Galen Murray, Maria Qazi, and Hugo Gerard for excellent research assistance and Katy Doyle for outstanding project management. Eli Berman, Marianne Bertrand, Stefano DellaVigna, Dean Karlan, Craig McIntosh, Brigitte Madrian, Sendhil Mullainathan, Rohini Pande, Jacob Shapiro, Charles Sprenger, Christopher Woodruff and many other colleagues provided insightful feedback. We acknowledge funding from the Citi/IPA Financial Capability Research Fund (FCRF), the Consortium on Financial Systems and Poverty (CFSP), the Empirical Studies of Conflict Project (ESOC), the Institute for Money, Technology, and Financial Inclusion (IMTFI), the UC San Diego Policy Design and Evaluation Lab (PDEL), and the USAID Development Impact Lab (USAID Agreement AID-OAA-A-13-00002). This work is also based on prior work supported by the U.S. Department of Defense Minerva research initiative through the Air Force Office of Scientific Research under award FA9550-09-1-000001. †University of Washington, School of Information, [email protected]. ‡Harvard University, Kennedy School of Government, michael [email protected]. §Princeton University, Woodrow Wilson School, [email protected].