1 of 18

©JICA

High Frequency Monitoring of Agricultural Commercialization in Malawi using SWIFT

Nobuo Yoshida

Lead Economist, the World Bank

2 of 18

Introduction of MA-SHEP

2

  • Project Name: Malawi Market-Oriented Smallholder Horticulture Empowerment and Promotion Project (MA-SHEP) by JICA

​

  • Project Purpose: Raise income through behavioral changes and improving marketing skills by training farmers

Project Target Areas

2nd Batch Target

1st Batch Target

4th Batch Target

3rd Batch Target

3 of 18

High Frequency Monitoring

3

Round 1

Round 2

Round 4

Round 3

February 2022

June 2022

November 2022

February 2023

4 of 18

　High-Frequency Monitoring for MA-SHEP with SWIFT

Quasi-experimental Study

​

  • Treatment group (1,080 households): MASHEP farmers (M)
  • Control Group (1,040 households): Non-MASHEP farmers (NM)
  • Balancing treatment and control groups: Their pre-program status of household welfare was similar

4

5 of 18

Key challenges for high frequency survey

5

Complex questions

Cost

Time

SWIFT

6 of 18

SWIFT (Survey of Wellbeing via Instant and Frequent Tracking)

6

  • Traditional poverty surveys are complex, slow, and costly
  • SWIFT is a fast, low-cost tool by the World Bank.
  • Uses 10–15 simple questions to estimate poverty.
  • Trains models with machine learning on survey data.
  • SWIFT has been implemented for 200+ applications in 70+ countries

7 of 18

The advantages of SWIFT

CHEAP ($15 or less per household)

TIMELY (2 – 5 minutes of interview)

USER-FRIENDLY (automated system)

COSTLY ($ 200+ per household)

SLOW (1 hour or more of interview)

COMPLEX – requires expertise and resources

Traditional approach

SWIFT approach

8 of 18

Poverty Reduction: Overall Trends

8

  • Remember that both groups started roughly at the same household welfare
  • In Round 1 (R1), the poverty rate of non-MA-SHEP farmers (NM) is 10% higher than MA-SHEP farmers (M)
  • Average gap grew to 14 % in Round 4 (R4)

9 of 18

　Poverty Reduction: Regional Differences and Stability

9

  • Rural Central region: the longest exposure to the project (3 years+)
  • Round 1: Gap in poverty was 11 percent
  • Round 4: Gap grew to 21 points

10 of 18

Resilience to Shocks

10

  • The tropical storm Ana hit Malawi just before Round 1

​

  • MA-SHEP farmers (M) recovered in one year.

​

  • The non-MA-SHEP farmers (NM) did not

Recovery from tropical storm Ana

11 of 18

Food Security (Coping strategy Index, rCSI)

11

12 of 18

Lessons learned

12

  • SWIFT made high frequency monitoring of poverty feasible and affordable
  • Food security indicator, like rCSI, often does not need SWIFT type of simplification
  • The survey unravels MA-SHEP
    • reduced poverty
    • improved food security, and
    • strengthened resilience to shocks

​

13 of 18

Future of SWIFT

13

SWIFT

AI -SWIFT

Project monitoring

Increasing frequency of poverty monitoring

Shock-responsive targeting mechanism

14 of 18

Annexes

14

15 of 18

Reference

  • This project is a joint initiative of JICA and World Bank teams. The presentation is prepared based on the following JICA-World Bank joint report:

Cardona, Lina Marcela, Harue Kitajima, Tomoya Masaki, Aya Tamura, Michino Yamaguchi, Nobuo Yoshida, and Kazusa Yoshimura. 2025. “Monitoring the Smallholder Horticulture Empowerment Promotion (SHEP) Project in Malawi Using SWIFT (the Survey of Well-Being via Instant and Frequent Tracking).” JICA–World Bank Joint Report. https://www.jica.go.jp/activities/evaluation/__icsFiles/afieldfile/2025/03/26/MASHEPSWIFTrepot.pdf.

  • If you are interested in more details, please download this report. And if you have any questions, please feel free to contact me (nyoshida@worldbank.org). I will connect you to team members or other experts on SHEP, SWIFT, and data collections.

15

16 of 18

Expansion of the SHEP Approach

  • The expansion of the SHEP approach is one of the Gov. of Japan’s national commitments through the TICAD process
  • “We would like to transition away from agriculture that enables the farmer to eat to agriculture the farmer to earn money.” by former Prime Minister Mr. ABE@TICAD V in 2013
  • Commitment: Expansion to 10 countries, capacity building of 1,000 skilled agricultural trainers, support to 50,000 farmers

⇒They were achieved by TICAD VI (2016)

16

©Gov. of JAPAN

17 of 18

Annex: Pre-program conditions and Re-weighting

17

  • Before reweighting, the pre-program status of household welfare was similar between MA-SHEP farmers and non-MA-SHEP farmers
  • After reweighting, the pre-program conditions of the MA-SHEP and non-MA-SHEP farmers are identical
  • This survey will tell us how much differences occurred after the implementation of MA-SHEP package

​

​

​

Rural South

Variables

Pre-reweighting

Post-reweighting

NM

M

NM

M

Floor = mud

0.78

0.80

0.80

0.80

Roof = grass

0.48

0.47

0.47

0.47

lighting = battery

0.81

0.73

0.73

0.73

Cooking fuel=collected

firewood

0.94

0.85

0.85

0.85

Pit latrine with slab

0.31

0.30

0.30

0.30

Mortar

0.56

0.49

0.49

0.49

Bed

0.20

0.21

0.21

0.21

Radio

0.35

0.30

0.30

0.30

Iron

0.15

0.14

0.14

0.14

Bicycle

0.49

0.41

0.41

0.41

Table

0.31

0.27

0.27

0.27

# of rooms

2.25

2.16

2.16

2.16

18 of 18

rCSI measure

18

Strategy

Description

Standard Weight

1. Rely on less preferred and less expensive foods

Eating cheaper or less preferred foods

1

2. Borrow food or rely on help from friends/relatives

Asking others for food or assistance

2

3. Limit portion size at mealtime

Reducing meal sizes so food lasts longer

1

4. Restrict consumption by adults so small children can eat

Adults eat less or skip meals

3

5. Reduce number of meals eaten per day

Skipping meals

1