Risk sharing and the economics of M-PESA

William Jack
Georgetown University

Tavneet Suri
MIT Sloan With support from the Consortium on Financial Systems and Poverty
Impact and Policy Conference August 30 – September 1, 2012 Bangkok

The solution: problem:

Jack - M-PESA

M-PESA as a risk spreading tool
• Formal insurance is limited • Informal insurance exists, but is often incomplete…….why? • Moral hazard: information asymmetries • Limited commitment: contract enforcement • Transaction costs
Jack - M-PESA

Summary of findings
• The consumption of households who don’t use M-PESA falls by about 7% - 10% when they suffer negative shocks • Lower transaction costs allow households who use M-PESA to smooth these risks perfectly

The M-PESA concept
• Remote account storage accessed by simple SMS technology • Cash-in and cash-out services provided by MPESA agents

Jack - M-PESA

Millions

16 14 12

Customer and Agent growth
2011

?

30,000

25,000

10

20,000

Customers

8 6

Customers 2008 Agents

15,000

10,000

4 2 0
Oct-06 Apr-07 Nov-07 Jun-08 Dec-08

2007

5,000

0
Jul-09 Jan-10 Aug-10 Feb-11 Sep-11

Jack - M-PESA

Agents

2009

2010

Lake Victoria Nairobi

Mombasa

June 2007
Jack - M-PESA

Note: partial data only

Lake Victoria Nairobi

Mombasa

Dec 2007
Jack - M-PESA

Note: partial data only

Lake Victoria Nairobi

Mombasa

June 2008
Jack - M-PESA

Note: partial data only

Lake Victoria Nairobi

Mombasa

Dec 2008
Jack - M-PESA

Note: partial data only

Lake Victoria Nairobi

Mombasa

June 2009
Jack - M-PESA

Note: partial data only

Lake Victoria Nairobi

Mombasa

Dec 2009
Jack - M-PESA

Note: partial data only

Lake Victoria Nairobi

Mombasa

June 2010
Jack - M-PESA

Note: partial data only

Our household survey
• 3,000 households across most of Kenya • Four rounds: 2008, 2009, 2010, 2011
Uganda

Somalia

Nairobi

Tanzania Indian Ocean
Jack - M-PESA

Who is using M-PESA?
100%

75%

50%

Households outside Nairobi
25%

Median consumption ~$2 per day
0% 2008 2009 2010 2011

>$2/day

$1.25-$2/day

<$1.25/day

Jack - M-PESA

Banking for the unbanked?
100%

75%

50%

Households outside Nairobi
25%

Median consumption ~$2 per day
0% 2008 2009 2010 2011

Unbanked

Banked

Jack - M-PESA

100%

How do people use M-PESA?
Share of households

80%
60% 40%

20%
0%

Transactions

2009 data
Jack - M-PESA

How often do people use M-PESA?
Less often Once a year Every 6 months Every 3 months 4% 4% 14% 43% 6% 5% 24%

Monthly
Every 2 weeks Weekly Daily 0%

2%
10% 20% 30% 40% 50%

Transaction Costs
1,400 1,200 1,000 Tariff 800 600 400 200 0 0 5,000 10,000 15,000 20,000 25,000 Amount deposited and sent M-PESA: Reg to reg 30,000 35,000 40,000

Postapay

Western Union

Empirical strategy
c = a +  Shock + User + bUser * Shock + controls Consumption User Shocks don’t hurt users so much (b)
Users are richer ()

Non-user (a) Shocks hurt ()

Shock

No shock

Shock status

Basic Results
M-PESA User Negative Shock User*Negative Shock OLSA 0.553*** [0.037] -0.207*** [0.038] 0.101** [0.050] -0.105*** [0.033] -0.207*** [0.038] PanelB -0.090** [0.036] 0.241** [0.116] 0.176*** [0.050] 0.052* [0.028] -0.069** [0.032] PanelC -0.016 [0.047] 0.232 [0.169] 0.156** [0.062] 0.055 [0.035] -0.068 [0.043] Without NairobiC -0.008 [0.049] 0.120 [0.141] 0.150** [0.065] 0.050 [0.037] -0.056 [0.045]

Shock, Users Shock, Non-Users

A: Full sample with time Fes; B: Full sample with controls + interactions C: Full sample, controls + interactions, time and time x location FEs
Jack - M-PESA

Improving Agent Access
Distance to the 3.5 closest agent 3 (km)
2.5 2 1.5 1 0.5 0 4

22% Change 14% Change 33% Change
Round 1 Round 2

40% Change

28% Change

Mean Distance (km)

5th Percentile

25th Percentile 50th Percentile 75th Percentile
Jack - M-PESA

Using Agent Roll Out
Agents w/in Agents Agents w/in Agents Distance to 1km w/in 2km 5km w/in 20km Agent Negative Shock Agents Agents*Shock 0.152 [0.152] -0.022 [0.039] 0.055*** [0.019] 0.122 [0.153] -0.003 [0.031] 0.050*** [0.015] 0.148 [0.160] 0.018 [0.024] 0.021** [0.010] -0.176 [0.140] -0.002 [0.006] -0.002 [0.005] 0.619*** [0.203] 0.051 [0.054] -0.058*** [0.019]

Jack - M-PESA

Mechanisms
• Consumption smoothing could be effected through
– Remittances – Savings – Information/communication

• We find remittances are the dominant factor
– More likely, More often, More – Larger network
Jack - M-PESA