The Impact of Mobile Money on Poverty - Research Brief - April 2021 - Bill ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Research Brief The Impact of Mobile Money on Poverty April 2021
DISCLAIMER This work is provided as-is, without any warranty of any kind, and for noncommercial, informational use only. Any further use may require the consent of third-party content owners. © Bill & Melinda Gates Foundation | April 2021 2
THE IMPACT OF MOBILE MONEY ON POVERTY
Mobile money accounts have spread widely
in select regions of the developing world,
particularly in Sub-Saharan Africa.
Over the past decade, evidence has emerged
citing beneficial impacts of mobile money in
developing economies on consumption, poverty, This presentation summarizes
labor outcomes, remittances, and migration. formative experimental and rigorous
non-experimental evidence from the
This has led to a surge in rigorous studies development economics literature.
focusing on the impact of mobile money on
poor and rural households who tend to be
unbanked and have nonexistent or very low
mobile money agent access.
© Bill & Melinda Gates Foundation | April 2021 3DEFINING MOBILE MONEY
“Mobile money enables mobile phone This Research Brief focuses on the individual and
household impacts of mobile money and, thus, takes
owners to deposit, transfer, and a broad definition of mobile money, focusing on the
user experience of account-to-account
withdraw funds without owning a transfers/payments enabled by mobile phones.
bank account. It is therefore distinct The studies covered leverage mobile money platforms
from mobile banking, which allows across a variety of providers and from a variety of
countries—including M-Pesa (Kenya), mKesh
access to one’s existing bank account (Mozambique), bKash (Bangladesh), and Airtel (Uganda,
Malawi, Niger).
via a mobile phone.”
Suri: 2017, Annual Review of Economics
These platforms operate under a diversity of regulatory
and licensing models for bank and non-bank led
approaches to mobile phone-based payments accounts.
Further information on regulation of mobile money
services can be found in the Inclusive Digital Financial
Services: A Reference Guide for Regulators1
1. https://docs.gatesfoundation.org/documents/InclusiveDigitalFinancialServices_ReferenceGuide.pdf
© Bill & Melinda Gates Foundation | April 2021 4EVOLUTION OF THE GLOBAL MOBILE MONEY SERVICES, 2001 TO 2019
Europe and Central Asia
Middle East and North Africa
Latin America and the Caribbean
South Asia
East Africa and Pacific
Sub-Saharan Africa
Source: GSMA State of the Industry Report on Mobile Money 2019
© Bill & Melinda Gates Foundation | April 2021 5THE RISE OF MOBILE MONEY: GLOBAL AND REGIONAL GROWTH IN 2020 Green arrows represent growth in 2020. Source: GSMA State of the Industry Report on Mobile Money 2020. © Bill & Melinda Gates Foundation | April 2021 6
THE IMPACT OF MOBILE MONEY ON POVERTY: KEY FINDINGS (1 OF 2)
Consumption, Risk Sharing, and Poverty Labor Outcomes and Investment
Mobile money had direct impacts on consumption, the ability to cope Mobile money impacted labor outcomes by allowing workers to shift into
with shocks, and extreme poverty. more productive occupations and firms to invest in fixed assets.
• Mobile money increased consumption expenditure by 44% when • In Northern Uganda, self-employment increased from 3% to 6% for
households experienced a flood shock in Mozambique (Batista and Vicente, individuals that lived far away from a bank branch (Weiser et al., 2019).
2020). • Women in Kenya increased self-employment by 2-3%. (Suri and Jack,
• Mobile money users in Kenya who experienced a negative shock saw no 2016)
change in their consumption level, whereas nonusers experienced a 7% • In Malawi, microentrepreneurs worked less in their primary business and
decrease in consumption (Jack and Suri, 2014). more on their farms (Aggarwal et al., 2020).
• Mobile money increased daily per capita consumption by 8% and reduced • In Bangladesh, rural households that had urban migrants were 17% less
the extreme poverty index by 42% when urban migrants remitted income likely to engage in wage labor (Lee et al., Forthcoming).
back to their household in rural Bangladesh (Lee et al., Forthcoming).
• Rural households in Mozambique reduced their agricultural investment by
• In Northern Uganda, mobile money increased food security by 45% for 28%, but saw an increase in their number of migrants, suggesting a shift
households that lived far away from bank branches (Weiser et al., 2019). from rural to urban occupations (Batista and Vicente, 2020).
• Kenyan female-headed households who lived in areas with many agents • Firms in Kenya, Tanzania, and Uganda that use mobile money saw a 16%
saw their long-run consumption grow by 8.5% (Suri and Jack, 2016). increase in the likelihood of investing in fixed assets (Islam et al., 2016).
© Bill & Melinda Gates Foundation | April 2021 7THE IMPACT OF MOBILE MONEY ON POVERTY: KEY FINDINGS (2 OF 2)
Remittances and Migration Savings
Mobile money users were more likely to send and receive remittances Mobile money does not tend to impact the level of savings*; however,
and to have additional household members migrate. there is suggestive evidence that mobile money accounts can be used
as a substitute for informal savings.
• Remittances received by Kenyan households increased their annual income
by 3-4%, following a negative shock (Jack and Suri, 2014). The biggest impact on savings seems to be for migrants and firms with
high cash turnover.
• In Mozambique, mobile money led to a 30% increase in the share of
migrants in a household. (Batista and Vicente, 2020). • Households in Bangladesh that had urban migrants and actively used
mobile money saved 296% more than nonusers (Lee et al., Forthcoming).
• In Bangladesh, mobile money increased the value of remittances by 28%
and the migration rate by 35% (Lee et al., Forthcoming). • 83% of microentrepreneurs in Malawi used mobile money accounts to save
when there were no withdrawal fees, and continued to save via mobile
• In Niger, 50% of households are willing to pay to use mobile money to send
money following the intervention once withdrawal fees increased (Aggarwal
remittances, but there is a lack of agent infrastructure to do so (Aker et al.,
et al, 2020).
2020).
*Weiser et al., 2019; Batista and Vicente, 2019; Jack and Suri, 2014
© Bill & Melinda Gates Foundation | April 2021 8RESEARCH BRIEFS:
THE IMPACT OF MOBILE MONEY
ON POVERTY
Bangladesh
4
Niger
9, 13
Sri Lanka
11
Kenya
Uganda
1, 2, 10, 12
5, 6, 10
Tanzania
7, 10
Malawi
8
Mozambique
3REPORTS SUMMARIZED
Relevance: Low High
Labor Outcomes
Consumption Remittances Migration Savings
Article Authors & Investment
Risk Sharing and Transaction Costs: Suri, Jack
1 Evidence from Kenya’s Mobile Money Revolution (2014)
The Long-Run Poverty and Gender Impacts Suri, Jack
2 of Mobile Money (2016)
Is Mobile Money Changing Africa? Batista, Vicente
3 Evidence from a Field Experiment (2020)*
Poverty and Migration in the Digital Age: Lee, Morduch, Ravindran,
4 Experimental Evidence on Mobile Banking in Bangladesh (Forthcoming) Shonchoy, Zaman
The Impact of Mobile Money on Poor Households: Wieser, Bruhn, Kinzinger,
5 Experimental Evidence from Uganda (2019)* Ruckteschler, Heitmann
Mobile Money and Risk Sharing Against Village Shocks (2018) Riley
6
Payment Mechanisms and Antipoverty Programs: Aker, Bounmijel,
7 Evidence from a Mobile Money Cash Transfer Experiment in Niger (2016) McClelland, Tierney
Cashing In (and Out): Experimental Evidence on the Effects of Mobile Money Aggarwal, Brailovskaya,
8 in Malawi (2020) Robinson
Migration, Money Transfers, and Mobile Money: Aker, Prina, Welch
9 Evidence from Niger (2020)
Does Mobile Money Use Increase Firms’ Investment? Islam, Muzi, Meza
10 Evidence from Enterprise Surveys in Kenya, Uganda, and Tanzania (2018)
Can Mobile-linked Bank Accounts Bolster Savings? De Mel, McIntosh, Sheth,
11 Evidence from a Randomized Trial in Sri Lanka (2018)* and Woodruff
Transaction Networks: Jack, Ray, Suri
12 Evidence from Mobile Money in Kenya (2013)
Mobile Money, Remittances, and Household Welfare: Munyegera, Matsumoto
13 Panel Evidence from Rural Uganda (2016)
*Study is currently a well-known working paper.
© Bill & Melinda Gates Foundation | April 2021 10SUMMARY OF ONBOARDING/TRAINING INTERVENTIONS ACROSS STUDIES
Many of the studies utilize some form of training for onboarding consumers or mobile money agents. Some onboarding interventions also distribute phones
and facilitate actual money transfers to encourage learning by doing. Note that these trainings with account administration and use are often quick, inexpensive, and,
therefore, quite different from many more comprehensive financial education interventions that have been examined elsewhere and shown to have little impact.
Study Date Onboarding in Randomized Control Trials
Can Mobile-linked Bank Accounts Bolster Savings? 2018* Participants were given a mobile phone, the minimum balance to open a savings account with the
Evidence from a Randomized Trial in Sri Lanka government bank, assistance linking the savings account to the mobile phone, and training on how to
make deposits to the savings account using the mobile phone.
Poverty and Migration in the Digital Age: Experimental 2020 Participants were given a 30-45 minutes training session on how to sign up and use mobile money
Evidence on Mobile Banking in Bangladesh (bKash).The session included at least five hands-on transactions.
Is Mobile Money Changing Africa? Evidence from a Field 2020* Participants were trained to deposit money onto the mobile account, make a purchase using mobile
Experiment money, and transfer mobile money to another mobile phone. Free trial money was given to the
participant.
The Impact of Mobile Money on Poor Households: 2019* Mobile money agents were recruited and given equipment, training, and marketing materials. There
Experimental Evidence from Uganda was no direct intervention with consumers to encourage mobile money use.
Cashing In (and Out): Experimental Evidence on the Effects 2020 Participants received a mobile phone and training on how to use a mobile money account.
of Mobile Money in Malawi
Payment Mechanisms and Antipoverty Programs: Evidence 2016 Participants were given a mobile phone and training on how to exchange e-money (mobile money
from a Mobile Money Cash Transfer Experiment in Niger digital currency) for cash.
*Study is currently a well-known working paper.
© Bill & Melinda Gates Foundation | April 2021 1101 02 03 04 05 06 07 08 09 10 11 12 13
Risk Sharing and Transaction Costs
MOBILE MONEY REDUCES VULNERABILITY TO SHOCKS
Authors: Tavneet Suri and William Jack
Journal: American Economic Review, 2014
Research Design: This study Descriptive Descriptive Impact Impact Impact
surveyed households across
Mobile Money Use Mobile Money Agents Consumption and Remittances Better Leveraging of
Kenya on welfare measures,
mobile money use, and Between 2008 and 2009, Between 2008 and 2010, Risk Sharing A negative shock increases Social Networks
remittances. In addition, the the share of Kenyan M-PESA agents increased Consumption levels the likelihood that M-PESA M-PESA users reach
study surveyed the entire households who used from 4,000 to 15,000, of M-PESA users are users receive remittances deeper into their social
network of M-PESA agents in M-PESA increased from whereas bank branches unaffected by negative by 9 percentage points. network to send and
locations where households 43% to 70%. grew by 20%. shocks, whereas non-users receive remittances in the
were interviewed. In the presence of a
experience a 7% drop negative shock, remittances presence of a negative
Survey Dates: in consumption. received by M-PESA users shock.
Households: Sept. 2008 – increases their annual
June 2010 consumption by about 4%.
Agents: March 2010
Households Using M-PESA Percent Growth 2008-2010 Consumption Levels After
Country: Kenya Negative Aggregate Shocks
275%
Sample: 2,017 households Non
and 7,700 agents 70% M-PESA
M-PESA
Users
Context: 70% of Kenya’s adult Users
population had adopted M- 43%
PESA Unaffected
7%
Contribution: Examines the Drop
impact of reducing the 20%
transaction costs of sending
remittances and a household’s
ability to cope with negative 2008 2009 Bank M-PESA
shocks Branches Agents
Jack, William, and Tavneet Suri. "Risk Sharing and Transactions Costs: Evidence from Kenya's mobile money revolution." American Economic Review 104, no. 1 (2014): 183-223.
© Bill & Melinda Gates Foundation | April 2021 1201 02 03 04 05 06 07 08 09 10 11 12 13
The Long-Run Poverty and Gender Impacts of Mobile Money
FEMALE-HEADED HOUSEHOLDS INCREASE FINANCIAL RESILIENCE AND
SAVINGS USING MOBILE MONEY IN KENYA
Authors: Tavneet Suri and William Jack
Journal: Science, 2016
Research Design: This study Descriptive Impact Impact Impact
uses long-term household data
Consumption Impact by the Numbers* Consumption Growth* Occupation*
to examine the impact of
changes in mobile money Kenyans consume approximately Mobile money lifted 2% of Kenyan Female-headed households The change in agent density
agent density1. US$2.50 on average each day. households out of poverty. That is, experienced an 18.5% increase in increased the share of women in
the increased availability of M- consumption due to an increase in business/ sales by 2% and
Survey Dates: 2008 – 2014
Poverty PESA agents “helped raise 194,000 agent density. decreased the share participating in
Country: Kenya households out of extreme poverty farming and secondary occupations
• Extreme Poverty: 43% of the
Sample: 1,608 households and induced 185,000 [female- by 3% and 1%, respectively.
sample live on less than US$1.25 Extreme Poverty*
headed households] to switch into
Context: 96% of Kenyan per day Increases in M-PESA agent density
business or retail as their main Savings*
households had used mobile • General Poverty: 66% of the caused the share of female-headed
occupation.” 2
money since its launch in sample live on less than US$2 households living in extreme Female-headed households
2007. increased their financial savings by
per day poverty to decrease by 21%; that is,
Contribution: Examines the from about 43% to about 34%. 22% due to the increase in agent
long-run impact of mobile
Occupations
194,000 Households density.
money, and in particular, Lifted Out of Extreme Poverty
• 25% of the sample are farmers Welfare Changes to Female Headed
differential impacts by gender. Households Due to
• 18% of the sample run a business Increases in Agent Density
*Impact results represent 185,00 Women HH Living in
the interquartile impact, i.e., Migration Extreme Poverty
Switched into Business or Retail
the difference between • 41% of households had
individuals at the 25th and at least one migrant +18.5%
75th percentiles of the agent
distribution - 21%
Increase in Agent Density Increase in
Consumption
1.Changes in agent density occurred between 2008 and 2010. 2.These results extrapolate the impacts derived from the sample to all households in Kenya. 3. Note: the study did not find impacts on migration. Citation: Suri, Tavneet, and William Jack. "The long-run poverty and
gender impacts of mobile money." Science 354, no. 6317 (2016): 1288-1292.
© Bill & Melinda Gates Foundation | April 2021 1301 02 03 04 05 06 07 08 09 10 11 12 13
Is Mobile Money Changing Africa?
MOBILE MONEY INCREASES THE ABILITY TO COPE WITH SHOCKS
IN MOZAMBIQUE
Authors: Cátia Batista and Pedro C. Vicente
Journal: Working Paper, 2020
Research Design: The Impact Impact Impact Impact
intervention introduced mobile
Consumption Labor Outcomes and Investment Rural to Urban Migration Mobile Money Transfers
money services to randomly
selected rural areas in Mobile money increases Agricultural activity decreased from Mobile money facilitates rural to Households who have access to
Mozambique. Individual and consumption expenditure in the 94% to 89% and agricultural urban migration by: mobile money and experience an
community-wide event of shocks. investment decreased by 28%. • Reducing remittance transaction aggregate flood shock are 11
demonstrations were held to percentage points more likely to
• Aggregate Flood Shock: 44% The reduction in agricultural activity costs
teach participants how to use receive a mobile money transfer
the service.
increase in consumption and investment, combined with the • Improving migration-based
expenditure than households who have access
increase in remittances and insurance possibilities
Intervention Active: July to mobile money and do not
• Household Shock: 21% increase migration, suggest an occupational Mobile money increases the share
2012 – June 2015 experience an aggregate flood
in consumption expenditure shift from rural to urban labor of migrants from exposed
Country: Mozambique shock.
activities. households by 15.8 percentage
Sample: 102 areas where points in the event of an aggregate
2,004 individuals were flood shock. Savings
surveyed Increases in Consumption Agricultural Activity
Expenditure Share of Households Mobile money does not have a
Context: The study took place significant impact on savings overall.
in a migration corridor. Aggregate Household 94%
Flood Shock Shock 89% However—households who have
Contribution: The area access to mobile money services
studied did not previously have
are 58 to 76 percentage points more
any mobile money services.
likely to save using mobile money
Themes: Examines responses compared to households in the
to aggregate and idiosyncratic +44%
unexposed group.
shocks. +21%
Exposed Group Exposed Group Unexposed Exposed
vs. Control Group vs. Control Group
For completeness, the research design also included behavioral games that were played in the field in order to illicit respondent’s marginal willingness to save and remit using mobile money.
Batista, Catia, and Pedro C. Vicente. ”Is mobile money changing rural Africa? evidence from a field experiment.” Working Paper, 2020.
© Bill & Melinda Gates Foundation | April 2021 1401 02 03 04 05 06 07 08 09 10 11 12 13
Poverty and Migration in the Digital Age
MOBILE MONEY REDUCES EXTREME POVERTY FOR FAMILIES OF
MIGRANTS IN BANGLADESH
Authors: Jean N. Lee, Jonathan Morduch, Saravana Ravindran, Abu S. Shonchoy, and Hassan Zaman
Journal: American Economic Journal: Applied Economics, Forthcoming
Research Design: The Impact Impact Impact
randomized control trial
Total Remittances: The intervention induced a Poverty: The intervention led to a 42% decline in Mobile Banking
selected migrant-household
pairs to facilitate and 26% increase in the value of total remittances the extreme poverty index of the exposed
encourage the use of a sent by urban migrants in the exposed group households that actively used bKash compared to Rural Households: Exposed rural households
Bangladesh mobile money compared to the unexposed group. the unexposed group. were 48 percentage points more likely to use
system, bKash, through a This suggests that new remittances were the bKash than the control group.
training intervention. The primary driver in the increase of total remittances
intervention taught participants Migration: The intervention led to a 7% decrease
rather than a substitution away from other means in the average household size of those exposed Urban Migrants: Urban migrants exposed to the
how to use mobile money and
of sending remittances. to the intervention and a 35% increase in the intervention were 47 percentage points more
translated the phone menus
from English to Bangla, the migration rate. likely to use bKash than the unexposed group.
local language. Consumption: Daily per capita expenditure in
Intervention Active: April households exposed to the treatment was 7.5% Labor:
2015 – June 2016 greater than households in the unexposed group.
• Exposed households that actively used bKash
Country: Bangladesh Likelihood of Using bKash
are 17% less likely to engage in wage labor.
Sample: 815 rural household- Remittances Daily Per Capita • For exposed households who actively used
urban migrant pairs Sent Expenditure bKash and engaged in self-employment, the 70% 68%
Context: The areas studied intervention led to a 42% increase in the
are rural, poor, and vulnerable number of self-employed people within the
to seasonal food insecurity household. The intervention did not
22% 21%
during the monga season. significantly induce households not engaged in
+26% self-employment to shift into self-employment.
Contribution: Examines the +7.5%
impact of mobile money as a
Unexposed Exposed Unexposed Exposed
facilitating mechanism
between rural-urban migration Exposed Group Exposed Group Rural Households Urban Migrants
pairs. vs. Control Group vs. Control Group
Lee, Jean N., Jonathan Morduch, Saravana Ravindran, Abu S. Shonchoy, and Hassan Zaman. ”Poverty and migration in the digital age: Experimental evidence on mobile banking in Bangladesh.” American Economic Journal: Applied Economics, Forthcoming.
© Bill & Melinda Gates Foundation | April 2021 1501 02 03 04 05 06 07 08 09 10 11 12 13
The Impact of Mobile Money on Poor Households
MOBILE MONEY INCREASES FOOD SECURITY IN RURAL UGANDA
Authors: Christine Wieser, Miriam Bruhn, Johannes Kinzinger, Christian Ruckteschler, and Soren Heitmann
Journal: World Bank Policy Research Working Paper, 2019
Intervention: The intervention Impact Impact Impact Impact Impact
rolled-out mobile money
Food Insecurity Labor Outcomes Impact by the No’s Usage Remittances
agents to randomly selected
areas in rural Northern Mobile money reduced the Mobile money stimulated In total, 8,576 households Mobile money agents Mobile money decreased
Uganda. share of households with the non-farm self- live in the exposed increased mobile money the costs of remittance
food insecurity. employment rate. communities. These usage. transactions.
Intervention Active: 2016 –
2017 This effect is likely due to This effect is likely due to impacts suggest that within
increased remittance households using their these communities:
Country: Uganda
transfers or the income increased peer-to-peer The rollout of 121 agents
Sample: 658 areas where generated from non-farm transfer receipts and cost provided self-employment
4,541 households were self-employment. savings from remittance to 257 households and
surveyed transfers to invest in self- improved food security for
Context: The regions studied -25%
employment. 1,345 households1.
are rural, poor areas that have
very few existing mobile + 121 Agents +33%
62.9%
money agents, low access to
47.2%
financial services through bank
16.8%
branches, and low remittance
receipts. Decrease in remittance
12.6% transaction costs equal
Contribution: Examines the +88%
Self Employment to saving 10%
impact of rolling out mobile + 257 Households of daily consumption
money to rural areas with low 6.4%
remittance activity (15%) and 3.4%
that are very far from banks.
Improved Food Security
Previous studies examined
country-wide samples with + 1,345 Households
Unexposed Exposed Unexposed Exposed Unexposed Exposed
remittance rates ranging from Areas Areas Areas Areas Areas Areas
40-65%.
1. These results extrapolate the impacts derived from the sample to all households living in the exposed areas. Wieser, Christina, Miriam Bruhn, Johannes Kinzinger, Christian Ruckteschler, and Soren Heitmann. ”The impact of mobile money on poor rural households: Experimental
evidence from Uganda.” The World Bank, 2019.
© Bill & Melinda Gates Foundation | April 2021 1601 02 03 04 05 06 07 08 09 10 11 12 13
Mobile Money Remittances, and Household Welfare
MOBILE MONEY INCREASES REMITTANCES TO RURAL HOUSEHOLDS
WITH MIGRANT WORKERS IN UGANDA
Authors: Ggombe Kasim Munyegera and Tomoya Matsumoto
Journal: World Development, 2016
Research Design: This paper Descriptive Impact Impact Impact Impact
studies the impact of mobile
Mobile money adoption Remittances Remittances Consumption Distance from Agent
money on welfare in rural
Uganda in the absence of increased from 1% in Comparing adopter Households located 1km
shocks. 2009 to 38% in 2012. households that have a away from a mobile money
Mobile money adoption
Survey Dates: 2009 - 2012 increases the probability of migrant worker to adopter agent consume less and
Remittance Receipts receiving remittances by 7 households that do not, the are 2 percentage points
Country: Uganda
percentage points. results show that less likely to receive
Sample: 846 Households households with a migrant remittances.
Contribution: One of the first worker:
papers to study the impact of Mobile money adopters • Increase their likelihood
mobile money on rural 78% receive 36% more in of receiving remittances
household welfare in the remittances than non- by 11 percentage points
65% Mobile money
absence of shocks. adopters, or approximately • Increase their total value adopters increase
US$61. of remittances by 42%. household per capita
50% 50% consumption by 13%
Other evidence suggests compared to non-
that prior to the introduction adopters.
of mobile money there was
no significant relationship
between having a migrant
worker and remittances;
thus, the results above
provide evidence in support
2009 2012 2009 2012 of the impact of mobile
Mobile Money Non- money.
Adopters Adopters
Citation: Munyegera, Ggombe Kasim, and Tomoya Matsumoto. "Mobile money, remittances, and household welfare: panel evidence from rural Uganda." World Development 79 (2016): 127-137.
© Bill & Melinda Gates Foundation | April 2021 1701 02 03 04 05 06 07 08 09 10 11 12 13
Mobile Money and Risk Sharing Against Village Shocks
MOBILE MONEY USERS SMOOTH CONSUMPTION IN THE PRESENCE OF
VILLAGE-LEVEL SHOCKS
Authors: Emma Riley
Journal: Journal of Development Economics, 2018
Research Design: This study Descriptive Impact Impact
uses a household survey to
Remittance Transactions Shocks and Consumption Remittances
examine the impact of mobile
money on household • The consumption level of mobile money users are • In general, mobile money users are 15 percentage
consumption in the presence • 67% of households have sent unaffected by the shock. points more likely to receive remittances compared to
of a village-wide rainfall shock. remittances • Households that live in villages without any mobile non-users.
In particular, the paper
• 82% of households have received money users experience a 7% decrease in • Following a shock, mobile money users receive
investigates the spillover
effects of mobile money to remittances consumption in the presence of a shock. US$10 more in remittances compared to non-users.
non-mobile money users when • Non-users that live in villages with mobile money This is approximately 4% of the median household’s
they reside in the same village users do not significantly differ from non-users that per capita income in 20131.
as mobile money users. Means of Sending Remittances
live in villages without any mobile money users. The results suggest that in the presence of a shock,
Panel Survey Dates: • 40% of households sent remittances mobile money users are not more likely to receive
The results suggest that although mobile money
2008 – 2013 (3 waves) physically via friends and family remittances, but the value of remittances received
users are able to smooth consumption in the
Country: Tanzania presence of a shock, non-mobile money users do not significantly increases.
Sample: 3,265 households in benefit from living in the same village with others that
26 districts use mobile money; that is, there are no spillover
Likelihood to Receive Remittances
Contribution: Examines the effects of mobile money detected.
spillover effects of mobile Consumption Levels After Negative Shock
money; that is, whether mobile
money users share their
35% 33%
of HH used mobile Non-User HH Non-User HH
remittances in the presence of money to send Mobile Money in Villages in Villages
a village-level shock. Sheds Users With Mobile Without Mobile 18%
remittances
light on how new technologies Money Money
affect traditional risk sharing
agreements. 7% 7%
Unaffected
Drop Drop
Users Non-Users
1. Only data from 2013 are used for these results.Citation: Riley, Emma. "Mobile money and risk sharing against village shocks." Journal of Development Economics 135 (2018): 43-58.
© Bill & Melinda Gates Foundation | April 2021 1801 02 03 04 05 06 07 08 09 10 11 12 13
Cashing In (and Out)
MOBILE MONEY LEADS TO A REALLOCATION OF LABOR FROM BUSINESS
TO AGRICULTURE FOR MICRO-ENTREPRENEURS IN MALAWI
Authors: Shilpa Aggarwal, Valentina Brailovskaya, and Jonathan Robinson
Journal: American Economic Association Papers and Proceedings, 2020
Research Design: The Impact Impact Impact Impact
intervention assisted randomly
Savings Labor Supply Deposits Interpersonal Transfers
selected micro-entrepreneurs in
opening mobile money 83% of exposed micro- Mobile money led micro-entrepreneurs exposed to the Micro-entrepreneurs The mobile money
accounts. Training modules on entrepreneurs reported intervention to work less in their primary business and more exposed to the intervention accounts led to a 25%
mobile money features were using mobile money on their farm. were 55-80% more likely to increase in the share of
provided, withdrawal fees were accounts for long-term make a deposit. exposed micro-
waived, and firms were savings and 12% for short- entrepreneurs making
encouraged to save using the The share of exposed micro-entrepreneurs working in their
term money storage, primary business decreased by 8.5% relative to the control The value of deposits transfers to people outside
accounts.
compared to 32% of group. Exposed: 75%. Unexposed: 82%. increased 67-83% for of the household.
Dates: July 2017 – Aug. 2019 unexposed micro- exposed micro- • Exposed: 55%
Country: Malawi entrepreneurs for any entrepreneurs relative • Unexposed: 44%
The share of exposed micro-entrepreneurs working on their
form of savings. to the control group.
Sample: 480 Micro- farm increased by 110% relative to the control group. Exposed micro-
entrepreneurs Exposed: 44%. Unexposed: 21%. entrepreneurs sent on
Exposed micro- average US$11 and
Context: The sample consisted Productive Activities
entrepreneurs made on received US$9.50,
of micro-entrepreneurs in urban Use of MM Accounts
Malawi that had less than 3
Primary Business Farming average 11 deposits compared to average
employees. Additionally, mobile 83% 82% amounting to US$90, deposits of US$90.
money use in Malawi is modest. 75% relative to their average
+110% daily profits of about
Contribution: One of the first -8.5% Post-Intervention
mobile money randomized
44% US$2.50.
32% There continued to be
experiments among micro-
12% 21% substantial usage in mobile
entrepreneurs. Additionally, one
of the only studies to find money accounts even after
impacts driven by savings rather Any Form Short- Long- the withdrawal fee waiver
Savings Term Term
than interpersonal transactions. Savings Savings Unexposed Exposed Unexposed Exposed was removed.
Unexposed Exposed
Aggarwal, Shilpa, Valentina Brailovskaya, and Jonathan Robinson. "Cashing In (and Out): Experimental Evidence on the Effects of Mobile Money in Malawi.” AEA Papers and Proceedings 110 (2020): 599-604.
© Bill & Melinda Gates Foundation | April 2021 1901 02 03 04 05 06 07 08 09 10 11 12 13
Migration, Money Transfers, and Mobile Money
DEMAND FOR MOBILE MONEY IN NIGER
Authors: Jenny C. Aker, Silvia Prina, and C. Jamilah Welch
Journal: American Economic Association Papers and Proceedings, 2020
Research Design: This study Descriptive Descriptive Descriptive Descriptive Key Insight
surveys households on
Mobile Phone Ownership Remittances Remittances Migration Patterns Willingness to Pay Mobile
migration, remittances, and
willingness to pay for mobile 84% of households in 68% of households How do respondents 54% of households had at Money Fees (via
money1. In addition, surveys the sample own mobile received remittances. receive remittances? least one seasonal migrant Behavioral Game Theory
are conducted on all money phones, and in general, and 17% had a permanent Experiment)
• Friend or Family
transfer service providers in 9% of households in Niger migrant. Approximately 50% of the
Member: 74%
Niger. have used mobile money. sample is willing to pay the
• Domestic Money
Dates: 2017 actual cost of sending the
Transfer Provider: 34%
transfer, yet only 3% use
Country: Niger • Bus: 8% this channel.
Sample: 460 households and • Mobile Money: 3%
45 money transfer service
providers
Context: Niger is one of the
most financially excluded
countries in sub-Saharan 84%
Africa. 54%
Contribution: Explores mobile 17% 50%
money adoption patterns in
Niger and provides evidence
on the willingness to pay for
mobile money services.
9% 68% 3%
Mobile Mobile Willing Actually
Phone Money to Pay Use
Ownership Use
1. Willingness to pay was observed using a behavioral game. Aker, Jenny C., Silvia Prina, and C. Jamilah Welch. “Migration, Money Transfers and Mobile Money: Evidence from Niger.” AEA Papers and Proceedings 110 (2020): 589-93.
© Bill & Melinda Gates Foundation | April 2021 2001 02 03 04 05 06 07 08 09 10 11 12 13
Does Mobile Money Use Increase Firms’ Investment?
DOES MOBILE MONEY USE INCREASE FIRMS’ INVESTMENT?
EVIDENCE FROM ENTERPRISE SURVEYS IN KENYA, UGANDA, AND TANZANIA
Authors: Asif Islam, Silvia Muzi, and Jorge Luis Rodriguez Meza
Journal: Small Business Economics, 2018
Research Design: This study Descriptive Descriptive Impact Impact
used the World Bank’s
Adoption of Mobile Money Reasons Firms Adopt Mobile Money Investment Ways Mobile Money is
Enterprise Surveys to examine
the relationship between 54% of firms in the sample used Mobile money use by Used and Investment
mobile money use and firm mobile money to conduct a manufacturing and service
Kenya Of the firms who adopt
outcomes. financial transaction. firms is associated with
mobile money there is a:
Year: 2012 Main Reason for Main Reason for a 16% increase in the
Adopting Not Adopting likelihood of investing. • 27% increase in the
Countries: Kenya, Uganda, likelihood of investing for
and Tanzania Satisfy Customers Payments firms that used mobile
Sample: 1,228 firms 54% Request Too Large money to pay suppliers
Context: The sample of firms Used Mobile • 21% increase in
are in the manufacturing and Money likelihood of investing for
Tanzania
service sector and have 5 or firms that receive mobile
more employees. Main Reason for Main Reason for payments from
Adopting Not Adopting customers
Contribution: Examines the
relationship between mobile Adopter Characteristics Reduce Customers +16% in the • 17% increase in the
money use by firms and Transaction Costs Do Not Use likelihood likelihood of investing for
On average, firms that adopt
private investment, and does firms that make payments
mobile money are: of investing
so by comparing across
to employees using
countries. • Smaller Uganda mobile money
• Younger
Main Reason for Main Reason for
• Concentrated in Adopting Not Adopting
the service sectors
Reduce Suppliers
• Located in the main business Transaction Costs Do Not Use
or capital cities.
Islam, Asif, Silvia Muzi, and Jorge Luis Rodriguez Meza. "Does mobile money use increase firms’ investment? Evidence from Enterprise Surveys in Kenya, Uganda, and Tanzania." Small Business Economics 51, no. 3 (2018): 687-708.
© Bill & Melinda Gates Foundation | April 2021 2101 02 03 04 05 06 07 08 09 10 11 12 13
Can Mobile-linked Bank Accounts Bolster Savings?
CAN MOBILE-LINKED BANK ACCOUNTS BOLSTER SAVINGS?
EVIDENCE FROM A RANDOMIZED TRIAL IN SRI LANKA
Authors: Suresh De Mel, Craig McIntosh, Ketki Sheth, and Christopher Woodruff
Journal: NBER Working Paper, 2018
Research Design: The Impact Impact Empirical Insight
randomized intervention
Partner Bank and Other Formal Bank Deposits Total Savings What if the intervention
introduced a novel savings
account mobile-deposit service The intervention led to a 44% increase in the amount of total savings deposited to the Total savings (formal and targeted women or those
provided by a partnering bank. partner bank. Mobile deposits accounted for less than half of this increase2. informal) were unaffected by who lived 2-5 km from a
Randomly selected individuals the intervention suggesting bank?
were mailed offer letters to that percentage gains in
participate. Those who accepted formal savings, as well as
were provided assistance opening
Women and Savings
Amount Deposited to Partner Bank via the partner bank, were not
a bank account, as well as given The intervention could
Mobile Money in Local Currency meaningful increases.
a mobile phone, SIM card, and potentially increase total
demonstration of the service. savings by 23% for women
Funds could be deposited without Transaction Fees relative to the unexposed
a transaction fee1. group.
Additional randomization
Intervention Active: December
assigned individuals to one
2011 – May 2013
of four exposed groups that Distance and Savings
Country: Sri Lanka (Central) differed by transaction fee
Households who live 2–5 km
Context: Formal savings are (0-8%).
widely available in Sri Lanka;
away from a bank branch saw
The level of transaction fee a 79% increase in the amount
however, informal saving methods
are commonly used.
(ranging from 0–8%) did not deposited to the partner bank
lead to differences in the and a 26% increase in formal
Sample Size: 1,908 individuals
demand for the mobile- deposits relative to the
Contribution: One of the first deposit service. unexposed group.
experiments to use mobile phone-
linked bank accounts to Both frequent and infrequent mobile-deposit users preferred the traditional
encourage savings, and in
method of deposits, implying that transaction costs are not a barrier to the use
particular formal savings.
of savings accounts.
1. Additional randomization varied based on the level of transaction fee, but most impact results compare participants without a transaction fee to the non-exposed group. 2. The intervention only provided the mobile-deposit service for the partner bank.
De Mel, Suresh, Craig McIntosh, Ketki Sheth, and Christopher Woodruff. ”Can Mobile-Linked Bank Accounts Bolster Savings? Evidence from a Randomized Controlled Trial in Sri Lanka.” NBER Working Paper, 2018.
© Bill & Melinda Gates Foundation | April 2021 2201 02 03 04 05 06 07 08 09 10 11 12 13
Transaction Networks
MOBILE MONEY LEADS TO MORE RECIPROCAL TRANSACTIONS IN KENYA
Authors: William Jack, Adam Ray, and Tavneet Suri
Journal: American Economic Review, 2013
Research Design: This study Descriptive Impact Impact Impact
surveyed households across
All M-PESA Transactions Remittances Reciprocity Types of Transactions
Kenya on detailed remittance
information, such as means of Reciprocal Transactions: 21% Receive The likelihood of M-PESA users to Households that use M-PESA are
and reason for the transfer. In Non-reciprocal Transactions: 61% of M-PESA users receive conduct a reciprocal transfer is 17%, more likely to send remittances for
addition, the entire network of 79% remittances compared to 24% of compared to 4% of non-users. regular support, credit
M-PESA agents were arrangements, and emergency help.
non-users
surveyed.
22% of all M-PESA User Send Composition of Transactions
Survey Dates:
Transactions are Reciprocal. 63% of M-PESA users send • 50% of an M-PESA user’s
Households: Sept. 2008 – transactions are sent as regular
Dec. 2009 This 22% is composed of: remittances compared to 29% of
Regular Support: 42% non-users support, vs. 61% of a non-user’s.
Agents: March 2010
Credit Arrangements: 14% • 11% of an M-PESA user’s
Country: Kenya transactions are credit
Emergency Help: 11% Likelihood of Conducting
Sample: 2,017 households Remittances Reciprocal Transfer arrangements, compared to
and 7,700 agents No Particular Reason: 19% 6% of a non-user’s.
Context: 70% of Kenya’s adult Other: 14% 17% • An M-PESA user’s transactions
population had adopted M- that are sent as emergency help
PESA 11% of Non-M-PESA User 63% 61%
do not significantly differ from a
Contribution: This paper Transactions are Reciprocal. non-user’s. 11% of a nonuser’s
extends the evidence on This 11% is composed of: 29% 24% transactions are sent as
M-PESA mobile money 4%
Regular Support: 53% emergency help.
transactions and risk sharing
by investigating the Emergency Help: 13% This suggests that M-PESA users
characteristics of interpersonal Credit: 4% Non- M-PESA Non- M-PESA are shifting away from regular
Non-M-PESA M-PESA
transactions and examining No Particular Reason: 22% M-PESA User M-PESA User
User User
support transfers and toward credit
the types of transactions that User User transfers, and possibly emergency
are conducted. Other: 8%
Receive Send support transfers.
Jack, William, Adam Ray, and Tavneet Suri. "Transaction networks: Evidence from mobile money in Kenya." American Economic Review 103, no. 3 (2013): 356-61.
© Bill & Melinda Gates Foundation | April 2021 2301 02 03 04 05 06 07 08 09 10 11 12 13
Payment Mechanisms and Anti-Poverty Programs
PAYMENT MECHANISMS AND ANTI-POVERTY PROGRAMS:
EVIDENCE FROM A MOBILE MONEY CASH TRANSFER EXPERIMENT IN NIGER
Authors: Jenny C. Aker, Rachid Bounmijel, Amanda McClelland and Niall Tierney
Journal: Economic Development and Cultural Change, 2016
Research Design: The intervention varied the Impact Impact Impact
delivery mechanism of an unconditional cash
Uses of the Transfer Food Security Children and Nutritional Status
transfer program in Niger following the
2009/2010 drought and food crisis. The delivery Households that received the cash transfer Households that received the cash transfer Although children in the mobile money
mechanisms varied as follows: cash delivered in via mobile money purchased a more diverse via mobile money had a more diverse diet transfer group ate slightly larger and more
an envelope, received a mobile phone along with set of goods compared to the other exposed than both groups that received the cash diverse meals, their nutritional status was
having the cash delivered in an envelope, and groups1: transfer manually. unchanged.
cash delivered via mobile money transfer. All
participants in the intervention received a cash • Compared to households that received This score was .28-.51 points higher. • Children in households that received the
transfer, so there was no pure unexposed group. the cash manually, mobile money transfer Households that received cash manually cash transfer via mobile money ate an
recipients purchased .78 more types of had a diet diversity score of 3.17 out of 12. additional 1/3 of a meal compared to both
Intervention Active:
May 2010 – May 2011
goods. groups that received the cash transfer
• Compared to households that received a manually. Children in the group that
Country: Niger received cash manually ate 3.17 meals
mobile phone, along with the cash in Increase in Diversity of Food
Sample: 1,152 Households in 96 villages by Mobile Money per day on average.
hand, mobile money transfer recipients
Context: Within Niger, there is high rainfall purchased .85 more types of goods. • Children in mobile money transfer
variability, which has led to at least 7 droughts households also ate more diverse meals
between 1980 and 2010. During the 2010
• Households that received cash manually
purchased 4.32 types of goods on relative to the group that received a
drought, 2.7 million people were classified as
average. +0.51 mobile phone along with the manual
vulnerable to extreme food insecurity. Agriculture
is the primary income source for 97% of +0.28 cash transfer. Their diet diversity score
households. was 12-14% higher.
Contribution: Disentangles the impact of
Relative to Relative to
technology from the transfer mechanism.
Households that Households that
Received Transfer Received Transfer
Manually Manually
+
Received a Mobile
Phone
1. Households were surveyed on which goods and services they purchased, but not on a full expenditure and income module. Aker, Jenny C., Rachid Boumnijel, Amanda McClelland, and Niall Tierney.
"Payment mechanisms and antipoverty programs: Evidence from a mobile money cash transfer experiment in Niger." Economic Development and Cultural Change 65, no. 1 (2016): 1-37.
© Bill & Melinda Gates Foundation | April 2021 24You can also read
