FINAL REPORT
Table of Contents
Executive Summary............................................................................................................................ i
1 Introduction...............................................................................................................................1
1.1 Background of the Problem..........................................................................................................1
1.2 The SAFETI Project ........................................................................................................................1
1.3 Purpose of the Evaluation.............................................................................................................3
1.4 The Evaluation Questions.............................................................................................................3
1.5 Limitations of the Study................................................................................................................4
1.6 Report Structure ...........................................................................................................................4
2 Evaluation Methodology and Implementation.............................................................................4
3 Findings .....................................................................................................................................7
3.1 Key Findings and Performance Analysis .......................................................................................7
3.1.1 Achievements in performance indicators.............................................................................7
3.1.2 System productivity ............................................................................................................15
3.1.3 Management.......................................................................................................................16
3.1.4 Benefit-cost.........................................................................................................................18
3.2 Project Management, Partnership and MEL...............................................................................19
3.2.1 Project Management ..........................................................................................................19
3.2.2 Monitoring, Evaluation and Learning (MEL) .......................................................................20
3.3 Overall Assessment.....................................................................................................................21
3.3.1 Effectiveness.......................................................................................................................21
3.3.2 Relevance............................................................................................................................21
3.3.3 Efficiency .............................................................................................................................22
3.3.4 Impact .................................................................................................................................23
3.3.5 Sustainability.......................................................................................................................23
4 Lessons Learned and Way Forward ...........................................................................................24
5 References...............................................................................................................................27
Annexure.........................................................................................................................................29
Acronyms
ADO Aquaculture Development Officer
AFF Aquaculture Field Facilitator
BB Bagda Bagda
BBS Bangladesh Bureau of Statistics
BCR Benefit-Cost Ratio
BDT Bangladesh currency (Taka)
BFRI Bangladesh Fisheries Research Institute
BG Bagda Golda
BMP Best Management Practice
BSFF Bangladesh Shrimp and Fish Foundation
CAPI Computer Aided Personal Interview
CATI Computer Aided Telephonic Interview
CBO Community Based Organization
CBSG Capacity Building Service Group
CODEC Community Development Centre
DFID Department for International Development
DoF Department of Fisheries
FGD Focus Group Discussion
FIQC Fish Inspection of Quality Control
GTZ Gesellschaft für Technische Zusammenarbeit (German Technical Cooperation Agency)
GP Golda Carp Polyculture
IEC Information Education and Communication
IFP Increasing Farming Productivity
INGO International non-governmental organization
KII Key Informants Interview
LGD Local Government Development
MEL Monitoring, Evaluation and Learning
MoC Ministry of Commerce
MoFL Ministry of Fisheries and Livestock
MoU Memorandum of Understanding
MTE Mid-Term Evaluation
NBR National Board of Revenue
NGO Non-Governmental Organization
PCR Polymerase Chain Reaction
PI Program Indicator
PL Post Larvae
PM Project/Program Manager
PSM Propensity Score Matching
SAFETI Safe Aqua Farming for Economic and Trade Improvement
SDC Swiss Agency for Development and Cooperation
SPF Specific Pathogen Free
SPSS Statistical Package for the Social Sciences
STATA Software for Statistics and Data Science
TMSS Thengamara Mohila Sabuj Sangha
UAEM Upazila Aquaculture Extension Manager
USA United States of America
USDA United States Department of Agriculture
USG United States Government
WSSV White spot syndrome virus
Acknowledgements
This Mid-Term Evaluation report is the result of a cooperative, transparent and mutually supportive 
working relationship between CBSG and the SAFETI project. 
The Evaluation team gratefully acknowledges the contribution of all SAFETI team members, but in 
particular Dr. Mohammad Mokarrom Hossain, Deputy Chief of Party, Mr. Md. Abdul Gaffar, MEL 
Specialist, Mr. Anisur Rahman, Shrimp Specialist, for remaining instrumental and supportive throughout 
the evaluation process including data collection tools preparation, methodology finalization, data 
analysis and report preparation and for making this mid-term evaluation possible. 
CBSG expresses special gratitude to Dr. John A Dorr, Chief of Party, SAFETI Project and David James 
Currie - Aquaculture Production and Trade Specialist, SAFETI project; Dr. Mesbah Ul Alam - Senior 
Adviser, Policy Analysis and Government Liaison and Country Representative, Winrock International, for 
providing strategic guidance for conceptualization and valuable comments on the oral field findings 
sharing meeting on October 2, 2019. Discussion and guidance from both Winrock International and 
SAFETI project team was eye-opening and CBSG is proud to be the evaluation partners of the project. 
We gratefully acknowledge the openness and interest of the Director General of Department of 
Fisheries Ms. Quazi Shams Afroz to listen to the presentation of our international consultant, Dr. D.
Ramraj and the other evaluation team members during the short presentation session at DoF 
conference room, Dhaka, held on October 2, 2019. There are many other participants in the 
presentation session on that day, who deserved to be mentioned but allow us to refrain from doing so, 
because we might forget one or two names. 
We are grateful to the farmer respondents who allowed us to collect information. We also appreciate all 
the SAFETI project partners, participants, and other stakeholders who extended their helping hand 
during our data collection period. 
Finally, while we acknowledge valuable inputs from mentioned altruistic persons, none of the individuals 
are in any way responsible for final text and numbers in this report.
The Mid-Term Evaluation Team –
Core members:
Dr. Moin Us Salam, Technical Team Leader
Dr. Habibur Rahman Khondaker, Technical Specialist
Dr. D Ramraj, International Consultant, Shrimp and Hatchery Sector Expert
Mr. Joyanta Roy, CBSG Team Leader
Associate members:
Ms. Musahida Akter, Research Associate
Mr. H M Simon, Research Officer
Mr. Al-Amin Islam, Data Management Specialist
Dr. Rita Sen, Gender Specialist
Final Report: Mid-term Evaluation of SAFETI Project i | P a g e
Executive Summary 
Background
Agriculture in the lower coastal belt of Bangladesh is centered on shrimp/prawn-based aquaculture 
which include the Black Tiger Shrimp (Penaeus monodon, locally referred to as ‘bagda’) and the Giant 
Freshwater Prawn (Macrobrachium rosenbergi, locally referred to as ‘golda’). This sub-sector provides 
livelihood to about 0.8 million people and contributes to the fifth highest export earnings. In spite of 
right edaphic and climatic conditions, shrimp/prawn productivity has become stagnant and external 
trade has declined in recent years. 
In the above context, Safe Aqua Farming for Economic and Trade Improvement (SAFETI) project was 
developed in 2016 with $15.70 million funding from the United States Department of Agriculture (USDA) 
to address the ongoing issues related to productivity, farm incomes, trade, and food safety across the 
shrimp/prawn value chains in Bangladesh. The life of the project was set for five years (October 2016 –
September 2021), though field activities were delayed by one year due to funding complexities. Two 
broad objectives of the project are: (i) to increase agricultural productivity by improving production and 
management of hatcheries and farms in the shrimp and prawn value chains; and (ii) to expand trade of 
agricultural products by improving handling and sanitary controls, reducing unnecessary layers in the 
value chain, promoting extension services, and building the capacity of trade associations and 
government trade promotion bureaus.
The Mid-Term Evaluation
The mid-term evaluation (MTE) was commissioned in August 2019 to (i) assessing the effectiveness of 
SAFETI project activities; (ii) looking at the project’s critical interventions and accomplishments against 
the mid-term targets; (iii) assessing and recommending any successful (pilot) interventions that could be 
scaled up for greater impact in the future across the specific activity or sector; and (iv) discussing the 
obstacles and challenges SAFETI had been facing, and how these could through learning, be adapted and 
addressed during implementation for the remainder of the project.
The MTE followed inclusive, participatory and transparent exercises to gain insights and alternative 
perspectives. Methodologies included document review (project documents and secondary literature)
quantitative sample survey, qualitative data collection such as Key Informants’ Interview with 
stakeholders, Focus Group Discussion with both control group and direct project farmers, and 
consultative meeting and workshop with the project team and other stakeholders such as associations 
related to shrimp industries. MTE used multiple primary and secondary information sources and 
stakeholder perspectives to validate as well as triangulate key issues and findings. 
SAFETI engaged shrimp and prawn farmers in several manners as follows: demonstration farmers who 
were selected from groups of 25 farmers referred to as ‘demo’ farmers and other farmers in those 
groups referred to as ‘direct’ farmers. Both demo and direct farmers received various types of 
assistance directly from the project. A third group of farmers drawn from the overall farming sector in 
SAFETI’s operational areas were those who SAFETI engaged through indirect mechanisms such as farmer 
field days, communication activities, etc. and are referred to as ‘indirect’ farmers. More specific 
definitions of how farmers were classified and credited to SAFETI’s targets can be found in the project’s 
Performance Monitoring Plan (PMP). Farmers denoted as ‘non-beneficiaries’ are those who have not 
received any direct or indirect assistance from SAFETI and are not counted as demo, direct or indirect 
Final Report: Mid-term Evaluation of SAFETI Project ii | P a g e
farmers/beneficiaries. Farmers included in the control group engaged
1 during the mid-term evaluation 
are classified as non-beneficiaries. 
Progress Achieved and Issues
The SAFETI planning and monitoring framework provided both mid-term and end of project targets on 
all key indicators. The MTE evaluated project progress against 16 specified targets including issues, 
challenges and scope for scale up. The following summary provides key accomplishments as well as 
issues: 
i. SAFETI has been promoting its trademark ‘6-Steps Towards IFP’ (improving farming practices) 
procedure among the farmers to invigorate shrimp/prawn farming practices and enhance 
productivity and income with emphasis on shrimp/prawn high-health PL, biosecurity, disease 
management, pond preparation, water quality management and postharvest management. 
Overall, the project exceeded its mid-term target for Indicator 1 by about 10% as 11,163 direct 
farmers had so far adopted and applied various steps of SAFETI prescribed procedures in farm 
management. However, it is important to note that not all farmers adopted all the 6-steps, or to 
the same extent. Adoption varied with farmers’ perception on individual technologies and 
individual capability to adopt those.
ii. Among the four key technology pillars in ‘6-StepsTowards IFP’, the biosecurity measures were 
those adopted most widely by the farmers; 87% by SAFETI demonstration (‘demo’) farmers and 
79% by non- demonstration farmers who were directly assisted by the project (referred to as 
‘direct’ farmers’). The recommended shrimp/prawn high-health ‘Post Larvae’ (PL) practices
were most commonly adopted, followed by disease and water quality management related 
practices. In contrast, harvest and postharvest management procedures were the least 
frequently adopted technology. The low frequency of harvest and postharvest technology may 
have been caused in part by the strategic approach of SAFETI to first emphasize technologies to 
increase production, and later concentrate on postharvest quality improvement issues. SAFETI’s 
two project partners, CODEC and TMSS, offered micro-financing opportunities to the farmers to 
allow them to invest in and apply SAFETI-promoted improved farming methods and technology. 
To date, TMSS and CODEC have provided 1421 loans (44% of the total number of loans 
disbursed), while the remaining 1833 loans (56%) were provided by five other MFIs. The farmers 
received a combined total of $1,342,109 in loans from the partners and MFIs, which was more 
than double the mid-term target. The average loan amount was slightly more than $400, which 
exceeded the average amount that the project had initially estimated, thereby resulting in 
overachievement of the loan value target but underachievement of the target for number of 
loans. The primary reason for the shortfall in number of loans was that TMSS and CODEC did not 
have the institutional capacity to expand their loaning activities into SAFETI areas that are 
outside of their current geographical area of operations. Other MFIs were able to service some 
additional areas/farmers, but the majority (82%) of SAFETI-assisted farmers were unable to 
access loaning services. 
iii. The productivity of SAFETI-assisted bagda and golda farmers exceeded that of non-beneficiary 
farmers. The project also exceeded the cumulative mid-term targets for production volume and 
sales. The cumulative production volume for the two species combined was 3,445 MT, which is 
9% more than the mid-term target. The cumulative value of sales was 1% above the mid-term 
1 Control group famers were selected from non-project Upazilas (sub-districts) who possessed similar socio￾economic characteristics of direct project farmers when the baseline survey was conducted. 
Final Report: Mid-term Evaluation of SAFETI Project iii | P a g e
target of $28,756,403. When disaggregated by species, the cumulative value of sales was nearly 
equal for bagda and golda, with both species contributing equally to the mid-term target. 
However, with respect to production, golda exceeded its mid-term sub-target by 33%, whereas 
bagda fell short of its sub-target by 8%.
iv. SAFETI investment has been benefiting farmers and market actors (inputs suppliers, feed sellers, 
hatchery and nursery operators and processors). Farmers realized Benefit-Cost-Ratios (BCR) of 
1.9 to 3.8 per annual cycle, depending on the type of culture system (i.e., mono-/polyculture; 
single/multiple crop) and farmer (demonstration, direct, non-beneficiary). Moreover, gross 
margins of SAFETI-assisted demo and direct farmers increased over non-assisted control farmers
(Bagda monoculture systems: demo farmers received BDT 491,344 ha-1
, direct farmers BDT 
328,568 ha-1 and control farmers BDT 268,458 ha-1
; Bagda-golda rotational system: demo 
farmers received BDT 572,869 ha-1
, direct farmers BDT 383,052 ha-1 and control farmers BDT 
505,248 ha-1
; Golda-carp polyculture system: demo farmers received BDT 747,766 ha-1
, direct 
farmers BDT 652,619 ha-1 and control farmers BDT 322,102 ha-1
.
v. PL stocking density affects productivity, with low density stocking being preferable to high 
densities. In line with SAFETI recommendations, project-assisted direct farmers stocked lower PL 
densities than the non-assisted control farmers, but not as low as the project-assisted ‘demo’ 
farmers.
vi. Harvest size is an important factor in sale price, with large shrimp/prawn generally fetching 
higher prices than smaller animals. Harvest frequency affects size in that more frequent 
(multiple) harvesting within an annual growing cycle reduces the period of growth, the size the 
animal can attain and thereby reduces the price. However, harvest frequency is also linked to 
risk management, i.e., production losses. Demo and direct farmers had considerably reduced 
harvest frequencies in comparison with control farmers. 
vii. The project reached 17,950- direct beneficiary farmers (about 97% of the mid-term target) with 
variety of inputs and services. Training was one of the most important support services. Farm 
management technologies had been disseminated mainly through training and information 
dissemination based on ‘6-StepsTowards IFP’ approach. The project so far trained 31% more 
farmers than the planned. Most of the additional training went to the ‘golda’ farmers and to the 
women beneficiaries. However, the project has yet to complete the all the expected number of 
trainings to stakeholders - government officials, industries, processors, service providers and
input dealers.
viii. On the policy and regulatory front, the project achieved policy stages 1 and 2 for all 5 targets
and stage 3 for 4 of the 5 policy targets. SAFETI is on schedule for achieving the remaining 
stages that are scheduled for Years 4 and 5. It may be noted that the project supported drafting 
the “Development of SPF Black Tiger Shrimp’s Hatchery Protocol” and submitted to the Ministry 
of Fisheries and Livestock (MoFL).
ix. Capability building of the Fish Inspection of Quality Control (FIQC) laboratory of DoF with 
training and equipment is an important ongoing project activity that will ultimately allow the 
Government to help provide training to the farmers and business owners on food safety. 
x. The average farm productivity project farmers (demo and direct) exceeded the mid-term target. 
The MTE estimated production of 664 kg ha- 1 and 724 kg ha-1 per annum for bagda and golda, 
respectfully, against the mid-term target of 450 kg ha-1
for each species. 
xi. Among the three culture systems, bagda monoculture proved to be most productive for the 
project farmers followed by bagda-golda rotation and golda-carp polyculture systems. The demo 
farmers outperformed the project farmers achieving 80% of the bagda monoculture and bagda-
Final Report: Mid-term Evaluation of SAFETI Project iv | P a g e
golda rotation production target and 70% the golda-carp polyculture system target. The 
performance of the demo farmers proved the potential of the SAFETI farming technology.
xii. The MTE noted that in the bagda monoculture farms, when two crops were run, productivity 
was lower in the second crop. The team also noted that the productivity of bagda in the bagda￾golda rotational farms was lower than in the bagda monoculture farms. Additionally, that the 
productivity of golda in the golda-carp polyculture farms was higher than in the bagda-golda
rotational farms. The likely reasons for these differences are: the lower water salinities due to 
monsoon rains in the second crop in each case; inadequate pond preparation for the second 
crop that increased biosecurity risk; unavailability of good quality feed and other inputs; and, 
the shorter duration of the second crop. The project should consider the relevance of the 
production environment when recommending production systems. 
xiii. Among other factors, increased production was made possible due to the availability of high￾health PL produced from Specific Pathogen Free (SPF) bagda broodstock, through active SAFETI 
motivation and support to hatcheries. On the other hand, production of improved high-health 
golda PL was not always possible due to failure of ensuring disease-free PL from natural 
broodstock, given that most golda PLs are still sourced from nature. 
xiv. Demo farms provide examples of new technologies and help promote sustainability in 
technology adoption. However, there remains a mixed perception on the potential and 
effectiveness of new technology and farm practices among the demo farmers. In some 
instances, demo farms did not play their extension role as effectively as they could. Although 
demo farms performed better than direct farmers, some demo farms lacked the comprehensive
farm management practices being promoted in SAFETI’s ‘6-StepsTowardsIFP’. 
Key Issues and Way Forward
SAFETI project has made good progress on many fronts but can also adapt based on what it has learned, 
to help meet its targets. The MTE identified key learning points that can help SAFETI more fully attain its 
targets. The following summary provides key lessons learned from the mid-point implementation of the 
SAFETI project and suggests ways forward for the remainder of project implementation. 
i. SAFETI always discourages multiple stocking and multiple harvest, and suggests that farmers 
harvest above 40-g sized shrimp. However, not all project beneficiaries are following this advice. 
Therefore, the project should continue its campaign and re-emphasize the benefits of delaying 
harvest to larger-sized shrimp to all beneficiaries including demo farmers.
ii. The second crop with bagda has not been always successful even when stocked with SPF PLs, 
due to increased biosecurity risks during the rainy season, water quality and environmental 
changes (compared with the dry-season crop). SAFETI may suggest an alternate to bagda (e.g. 
growing rice instead of bagda) for the farmers, because the yield from the second crop of bagda 
is lower and disease outbreaks are more pronounced than then first crop. MTE understands that 
SAFETI is working on promoting this initiative. 
iii. However, in some areas where salinity remains high, a second crop of bagda has the potential to 
perform better than in areas where rain reduces salinity. This would require the adoption of 
appropriate farming technology adoption; SAFETI may explore this opportunity.
iv. The golda crop in the bagda-golda’ rotational culture is below par resulting from the remnant 
salinity in the soil and water and shortened duration of the production cycle. Tidal flushing of 
pond water prior to stocking golda may prove effective, but may not be feasible in many cases.
v. In complex multiple-cropping situations, SAFETI should promote practical approaches such as 
strictly following black soil removal, pond drying where applicable, liming and beaching.
Final Report: Mid-term Evaluation of SAFETI Project v | P a g e
vi. Reduction of harvesting frequency of animals (i.e., increasing the growing period) would likely 
result in increased size of the products and attract higher selling price and thereby impact 
profitability - assuming the risk of crop failure during the extended time period can be 
minimized. The project should consider sending this message to the farmers.
vii. SAFETI should identify a small number of high performing demo farmers who can serve as 
champions, form linkages with direct and indirect farmers, and help demonstrate and promote 
their successful farming methods.
viii. Among the indicators of new technology adoption, postharvest operations remained poorly 
adopted by the farmers; this is because SAFETI’s priority was to address the issue of low 
productivity first, then looking into postharvest issues. SAFETI should emphasize promoting 
improved post-harvest technology in the rest half of the project.
ix. SAFETI needs to determine if the farmers who have not yet been able to access loan services 
desire loans, and if so, attempt to facilitate an expansion of loan services to their areas. 
Final Report: Mid-term Evaluation of SAFETI Project 1 | P a g e
1 Introduction 
Background of the Problem 
Bangladesh is considered as one of the most suitable regions for fisheries in the world, with the world's 
largest flooded wetland and the third largest aquatic biodiversity in Asia after China and India 
(Shamsuzzaman et al., 2017). The fisheries in the country can broadly be classified into three categories: 
inland capture fisheries, inland aquaculture and marine fisheries, of which the inland aquaculture sector 
is contributing more than 55% of the total production (DoF, 2016). Two types of aquaculture practices 
are going on in Bangladesh - freshwater and coastal aquaculture. The ‘Chingri’-based (shrimp locally 
called ‘bagda’, Penaeus monodon and prawn locally called ‘golda’, Macrobrachium rosenbergii) coastal 
aquaculture, practiced in ‘ghers’ (coastal ponds or enclosures), has significant social and economic 
importance in Bangladesh. About 15 million people are directly or indirectly engaged in shrimp/prawn 
related activities in hatchery, nursery, farming, processing industry, depot, landing centre, feed industry 
and transportation (MoFL, 2014). Livelihoods of about 800,000 environmentally vulnerable smallholders 
largely depend on this industry. Shrimp is one of the major valuable export items and ranking as the fifth 
largest source of earning foreign currency in the country (BBS, 2018). Small-scale commercial 
shrimp/prawn farming started in Bangladesh in the 1980s. Since then, it has gradually spread to larger 
areas. Initially, this farming spread was unplanned by releasing seeds in the ghers and raising with 
home-sourced feed; the same practice is still being practiced by many smallholders. Coastal Bangladesh 
has right conditions (soil, salinity and climate) for this farming in more than 275,000 pond/gher areas, 
and in a position to be the next country making a major jump in commodity productivity (SAFETI 
Document 02). Bangladesh is one of the most suitable countries in the world for freshwater prawn 
farming, because of its favorable biophysical resources (Ahmed et al., 2008 In spite of this potential, 
production remains almost unchanged and export volume has been declining since 2011 (SAFETI 
Document 02 and 03). This is because, in recent years, the industry has encountered challenges that 
include disease, compliance with quality standards in exporting markets, and inequitable terms of 
exchange among value chain actors (Belton et al., 2011).
The SAFETI Project 
The Safe Aqua Farming for Economic and Trade Improvement (SAFETI) project was developed to address 
the ongoing critical issues of constraining productivity, farm incomes, and trade and food safety across 
the shrimp/prawn value chains in Bangladesh. The United States Department of Agriculture (USDA)
funded $15.70 million for the project to carry out in five years (October 2016 – September 2021) (Award 
Agreement #: FCC-388-2016/004-00). SAFETI set two broad objectives to achieve outcomes; those are 
to:
i. Increase agricultural productivity by improving production and management of hatcheries and 
farms in the shrimp and prawn value chains to contribute to both improved incomes, food 
quality and safety, as well as to improve livelihoods and environmental sustainability in the 
Bangladesh prawn and shrimp industry; and 
ii. Expand trade of agricultural products by improving handling and sanitary controls, reducing 
unnecessary layers in the value chain, promoting extension services, and building the capacity of 
trade associations and government trade promotion bureaus.
Given the substantial potential for shrimp/prawn industry to positively impact smallholder producers 
and other direct and indirect beneficiaries, SAFETI, in collaboration with project partners, has been
implementing activities to address issues of related to under production and food safety and working to 
expand trade of shrimp and prawn. Elements of these project activities include investing in quality 
Final Report: Mid-term Evaluation of SAFETI Project 2 | P a g e
production inputs, promoting good aquaculture and hygienic practices, facilitating quality improvement 
and working with processors to expand trade.
Initially, the Cooperative Agreement with USDA lists 11 Activities for SAFETI. However, with USDA 
approval, portions of Activity 8 were either discontinued or merged with Activity 7. Thus, incorporating 
10 activities (as per the Cooperative Agreement with USDA), SAFETI addressed three critical issues in 
relation to boosting shrimp-prawn industry in Bangladesh: Critical Issue 1 (CS-1) - productivity and farm 
income; Critical Issue 2 (CS-2) - food safety; and Critical Issue 3 (CS-3) - trade. The 10 activities are – (A1)
Inputs: Develop Agro-dealers and/or other Input Suppliers- to improve shrimp and prawn quality, 
increase hatchery efficiency and reduce the incidence of disease in hatcheries; (A2) Financial Services: 
Facilitate agricultural lending through financial services to support adoption of project-promoted inputs, 
practices and technologies; (A3) Training: Improved agricultural production technologies extend to 
shrimp-prawn producers, hatcheries and processors; (A4) Capacity Building of Government (GOB) 
Institutions: Build the capacity of government extension agents and testing laboratories and more 
importantly to support industry certification efforts; (A5) Capacity Building to Promote Improved Policy 
and Regulatory Framework: Support GOB efforts to rigorously establish food safety in the shrimp/prawn 
value chain; (A6) Training on Sanitary and Good Practice Standards and Certification: Primarily for 
selected processors; (A7) Training on Post-harvest Handling: Primarily for processors; (A8) Training on
Improved marketing and branding: Working with selected processors to provide support to help them 
identify and access markets, meet quality requirements of buyers, and strengthen their company brands 
as names recognized for reliable quality; (A9) Market Access to Facilitate Buyer-seller Relationships: 
Work with agro-based service centers who will coach farmers on pursuing market opportunities while 
assisting lead intermediary and end market firms to develop sustainable business models; and (A10) 
Capacity Building to Trade Associations: Strengthen capacity of key export trade facilitators (government 
agencies, trade associations) through organizational capacity building and strategic market development 
training. MTE observed direct linkage of four activities (A1, A2, A3 and A9) related to increasing 
productivity and farm income (CS-1), two (A4 and A5) to ensuring food safety (CS-2) and four (A6, A7, A8 
and A10) to expanding trade (CS-3). However, a number of activities cross-linked to CS1 and CS2.
SATETI focused on development of three shrimp/prawn ‘semi-intensive’ culture systems for the 
smallholders in coastal Bangladesh: (i) Bagda monoculture; (ii) Bagda-Golda rotational culture; and (iii) 
Golda-Carp polyculture. The ‘semi-intensive’ culture system is defined as maintenances of stocking 
density @ 3-4 for bagda and 2-3 for golda PL per square meter with no use of aerator, and providing 
regular and partial feeding emphasizing on natural productivity, proper pond preparation, water quality 
and biosecurity management. This system is similar to the ‘improved extensive system’, defined by DoF. 
In doing so, the project has promoted six steps aimed at increasing farm productivity, referred to as the 
‘6-StepsTowardsIFP’ of bagda and golda in the three culture systems. These are (i) Maintaining optimum 
water depth of 3–5 feet and cleaning (i.e., removing mud, and drying) pond bottom; (ii) Disinfecting 
water and removing all potential threats for disease (iii) Stocking appropriate density with high-health 
(SPF) or PCR tested PL for bagda and healthy PL for golda; (iv) Ensuring appropriate biosecurity to 
prevent cross- contamination by outside disease carriers; (v) Using good quality feed administered on an 
optimum feeding schedule; and (vi) Ensuring good water quality and shrimp/prawn health management.
In implementing its activities, the SAFETI partnered with BSFF (regarding policy and regulations), Auburn 
University USA (regarding golda seed and feed technology), WorldFish (regarding disease investigation 
and feed technology), and TMSS and CODEC (to work with farmers and facilitate microfinancing). 
Additionally, intense collaboration was established with DoF and BFRI, formalized through MoUs.
For field implementation of the productivity-related activities, SAFETI organized beneficiary farmers into 
‘groups’. Each group numbered around 25 farmers, one being the demo farmer and the rest direct 
Final Report: Mid-term Evaluation of SAFETI Project 3 | P a g e
farmers. The concept of demo farmer was designed to peer leadership, whereas the rest of the group to 
be the followers. Each demo farmers received in-kind need-based support from the project worth 
around BDT 20,000. The groups of demo and direct farmers were trained on the basics of SAFETI 
promoted technologies (three modules) by the Upazila Aquaculture Extension Manager (UAEM). 
Following this, they were further trained, coached and supervised by Aquaculture Field Facilitators 
(AFFs), the lowest-tier field staff of the two partner organizations. Each AFF was responsible to look 
after, on average, 10 groups, who himself/herself was supervised by the UAEM. The technical 
assistances were delivered to beneficiary farmers via UAEM by the Aquaculture Development Office 
(ADO), the core SAFETI staff stationed at district-level. The ADO was also responsible for capability 
development of aqua-inputs dealers in the district.
Purpose of the Evaluation 
The purpose of this mid-term evaluation (MTE) was to
i. assess the effectiveness of programmatic approaches of SAFETI project activities carried out so 
far;
ii. explore and look at the project’s key critical interventions and accomplishments against the mid￾term project targets;
iii. assess and recommend any successful (pilot) interventions that can be scaled up for greater 
impact in the future across the specific activity or sector; and
iv. Discuss the obstacles and challenges SAFETI faces, and how these can through learning, be 
adapted and addressed during implementation for the remainder of the project. 
In addition, the evaluation intended to give the SAFETI team and partners, collaborating institutions, 
USDA, and other stakeholders (e.g., Solidaridad, World Bank, DFID, Netherlands, SDC and GTZ) who had 
worked in this sector an understanding of:
i. progress on the implementation of activities and key interventions;
ii. project performance with respect to intermediate results and indicator targets to date;
iii. key implementation gaps if they exist and what causes them;
iv. evidence of where project human, financial, and material resources are being deployed 
efficiently and effectively and where adjustments may be appropriate;
v. unforeseen opportunities or challenges that may require revisions of the project strategy or 
redirection of project resources; and
vi. Implementation strengths and what they mean for the underlying project logic and for ongoing 
implementation.
The Evaluation Questions 
The MTE addressed six salient evaluation questions, as defined below.
Relevance: The alignment of project interventions to the needs of the project beneficiaries, the 
country’s agriculture and/or development investment strategy, and with USDA and the USG’s 
development goals, objectives, and strategies.
Effectiveness: The extent the project had achieved its objectives according to the results of program 
indicators.
Efficiency: How the project resources (inputs) had led to the desired results (outcomes) by investigating 
insights into efficiency of resources in achieving desired results by looking at the cost-benefit of the 
interventions and activities.
Final Report: Mid-term Evaluation of SAFETI Project 4 | P a g e
Impact:The radiation of effects beyond outcome targets, i.e., beyond targeted direct beneficiaries 
and/or beyond pre-projected results. 
Sustainability: The likelihood that the benefits of the project would endure over time after completion 
of the project by addressing sustainability by looking at what measures had been taken to ensure 
project impacts would endure after end of the project ended.
Learning: Knowledge, skills and lessons gained from project implementation by evaluating how the 
SAFETI MEL system allowed for project learning loops and continuous improvement and innovation.
Limitations of the Study 
The evaluation had to be completed in a tight timeframe. The team needed to spend considerable time 
in identifying and selecting control farmers. The baseline data (e.g., production per unit area) from the 
original survey in 2017-2018 were developed looking at intensive (aerated) and semi-intensive farms -
but separately. The adjustment was then made only to the semi-intensive data set, eliminating a few 
outliers like 400 ha farms in Chattogram with very low productivity. However, eliminating outlier values 
associated with very large farms with profiles that did not match those of SAFETI-targeted farmers, and 
whose production greatly exceeded that of the usual small farmer and therefore skewed the data and 
sample statistics. While the MTE provides a recommendation for this baseline data, we recognize the 
limitations in making this determination.
Report Structure 
Apart from the executive summary, this report consists of five parts: 
Introduction: Encompasses the background of the study, overview of the SAFETI project, assessment 
context, assessment questions, and study limitation(s).
Evaluation methodology: Provides a short description of principles of data collection and tools, 
estimation of sampling and sample size, implementation of assessments, and techniques of data 
banking and analysis. It also presents the characteristics of the surveyed households, in particular their 
income and educational status.
Findings: Presents (i) the key findings of the evaluation with respect to the project performance 
indicators and system productivity; (ii) overall assessment of the project in relation to relevance of the 
project to context, project design, activities and implementation and to gender; SAFETI’s effectiveness in 
the line of program indicators; efficiency of the project; impact of SAFETI quantifying how the project 
meeting its targets; and sustainability highlighting how the impacts would endure after end of the 
project.
Learning: Points out the issues and lessons so far learned on project activities and implementation. 
Conclusions and Recommendations: Contains summary of the SAFETI’s performance followed by 
recommendations.
In addition to those sections, this report presents several reference documents and analysis tables in the 
annex. These annexes contributed to a deeper analysis allowing for more accurate evidence-based 
conclusions and recommendations.
2 Evaluation Methodology and Implementation 
MTE followed inclusive, participatory and transparent exercises to gain insights and alternative 
perspectives in the study. The following methodologies were adopted in the evaluation processes; the 
details are presented in Annex-6.
Final Report: Mid-term Evaluation of SAFETI Project 5 | P a g e
Document Review: This included project reports, communication and training materials, monitoring 
data on quality and outputs of activities, baseline survey and other available study data. In addition, 
research and other related documentation were reviewed. The list of the documents reviewed is 
presented in the ‘Reference’ section (Section 6).
Quantitative Sample Survey: The evaluation carried out a sample survey to ascertain program 
effectiveness on the key participants i.e. farmers. The survey was implemented on three ‘Farmer-types’: 
‘demo’, ‘direct’ and ‘control’ farmers. Demo farmers were direct SAFETI project beneficiary farmers who 
received ‘inputs-support’ in addition to training and technical support. Direct farmers received training 
and technical supports, but not ‘inputs-support’, whereas, control farmers did not receive any support 
from SAFETI. Furthermore, the ‘control’ farmers were out of SAFETI locations, but in ecologically similar 
areas with respect to soil, water and environmental conditions. 
Qualitative Data Collection: The evaluation adopted qualitative methods including Focus Group 
Discussion (FGD), in-depth Key Informants Interview (KII) and Direct Observation methods to collect 
qualitative information from a wide range of project participants and stakeholders. The list of meetings 
and meeting participants is presented in Annex-1.
Consultative Meeting and Workshop: The evaluation team arranged a consultation meeting cum 
workshop with the participation of project management and important stakeholders (of both treatment 
and control areas) to validate and further generate recommendations.
Table 1 presents the sample size for the field survey, 
FGDs and KII (details of the specific locations are 
appended in Annex-2). The sample survey questionnaire 
was developed based on SAFETI project result 
framework and outcome indicators as well as applicable 
Food for Progress indicators. The questionnaire was field 
tested and refined as needed. The questionnaire was
setup in the SurveyCTO platform for recording data 
electronically. Data collection was carried out by 24 
professional enumerators who have prior and reliable 
experience with similar studies in the fisheries sector. 
The KII and FGDs were conducted by the senior 
consultants of CBSG, while an international consultant 
conducted a few strategic KIIs and FGDs. The field data 
collection was carried out between September 4 and 17, 
2019. CBSG used SPSS, STATA, 
and MS Excel for data analysis.
For the characteristics of the 
surveyed households (i.e., the 
best or most relevant units 
related to the Pond/Gher 
information summarized in 
Table 2), the average land size 
of the respondent farmers was 
91 decimals, of which 12.3% 
was dyke area (Table 2). The 
demo farmers had smaller 
ghers (59 decimals) than direct 
Table-1: Sample size by data collectionmethods
Category of 
Farmer
Bagda 
Mono
Bagda￾Golda 
Rotational
Golda￾Carp Poly Total
Demo 9 7 7 23
Direct 164 190 S 525
Control 74 65 64 203
Total 247 262 242 751
Sample Survey Respondents
Focus Group Discussions (FGD):
Project (Treatment)– 6; Control-3
Key Informants Interview (KII):
Hatchary-4: Nursary-2; DoF officials-13, Partners 
NGO-2, Other stakeholders - 14
Table 2: Physical specifications and land tenure of the best production
units (Ghers) of respondent farmers in the field surveyed area
Pond / Gher information
Farmer-type
Demo Direct Control All
Land size (decimal) 59 92 92 91
Dyke area (% of land area) 9 12 12 12 Water depth main area (feet) 4 3 3 3 Water depth canal (feet) 6 5 5 5
Leasing land size (decimal) 49 89 102 91
Leasing cost (BDT decimal-1
) 755 612 365 551
Tenure of lease (year) 3 2 2 2
Final Report: Mid-term Evaluation of SAFETI Project 6 | P a g e
or control farmers (92 decimals). Interestingly, the proportionate dyke area for demo farmers was 9% 
compared with 12% for direct farmers. Overall, farmers maintained 3-foot water depths in main 
production areas and 5-foot depths in the canals. Demo farmers maintained relatively deeper water 
depths (4 feet) in comparison with the project and control farmers. On average, the lease tenure of the
ghers was 2 years, but longer (3 years) for demo farmers for the reported ghers. The leasing cost - on 
average BDT 551 for the duration of leasing - varied according to farmer type: BDT 755, BDT 612 and 
BDT 365 for demo, direct and control farmers, respectively. In terms of physical condition, the quality of 
the ghers operated by the control 
farmers was not as good as that of 
the demo and direct farmer ghers.
Across the survey respondents, the 
average annual income was BDT 
415,092 per household; by farmer
type it was BDT 406,691, BDT 390,953
and BDT 478,471, respectively for 
demo, direct and control farmers. The 
average national rural household 
income of Bangladesh is BDT 202,724
(BBS, 2019) indicating the survey 
respondents were much higher 
income earners than the national 
average. Farmers earned major share of income from fisheries sector (46.6% for demo, 58.1% for direct 
and 52.5% for control farmers, (Table 3). The source of income from crop sector was 17.4%, 12.8% and 
14.8% for demo, project and control farmers, respectively. Percent share of rural household income 
from agriculture (crop, livestock and fisheries sectors) nationally stands as 38.2% compared with 76.9% 
for the respondents, indicating the farmers in the study area were heavily involved in agricultural 
activities. Also, among the three types of farmers, 
demo farmers had high proportion of income from 
services and remittance (21.5%) compared with 6.8% 
for direct farmers and 7.0% for control farmers. 
Therefore, the demo farmers should have better 
opportunities than other farmers for investing in 
improved shrimp/prawn farming and/or coping with 
adversities in shrimp/prawn farming such as natural 
shocks and unanticipated mortality.
About 96%, 92% and 87% of demo, direct and control
farmers, respectively, received some sort of formal 
education (Table 4); the literacy rate of all farmer-type was greater the national average of 70.2% (BBS, 
2019. A higher number of respondents from control group (13.3%) attended university compared to 
direct (6.5%) and demo (4.3%) farmers. However, a reverse trend was observed in the farmer-type who 
attended or completed secondary education (school-years 6-10). In general, the absolute majority of the 
respondents under the three types of farmers were almost equally educated. About 75% of the 
respondent demo and direct farmers enrolled as SAFETI project beneficiaries in 2017-18 production 
season and the remainder enrolled in 2018-19 production season.
Table 3: Share of annual income of demo, direct and 
control farmers from various income-sources
Demo Direct Control
Crops 17.4 12.8 14.8
Bagda 24.6 28.2 27.6
Golda 15.4 16.5 10.0
Other fish 6.6 13.4 15.0
All fish 46.6 58.1 52.5
Livestock 6.8 5.8 10.6
Service 10.8 5.8 5.2
Remitance 10.7 1.0 1.8
Others 7.6 16.5 15.1
Annual household income source (%)
Farmer-type
Table 4: Education level of respondent demo, direct and 
control farmers in the study area
Demo Direct Control
Primary 8.7 25.3 22.2
Secondary 78.3 48.4 44.8
Higher secondary 4.3 10.1 6.9
University 4.3 6.5 13.3
Religious 0.0 1.5 0.0
No 4.3 8.2 12.8
Farmer-type
Education level of respondent (%)
Final Report: Mid-term Evaluation of SAFETI Project 7 | P a g e
3 Findings 
The findings of this MTE are presented under three broad headings : (i) Key Findings and Performance 
Analysis, which critically investigated and evaluated the performance of SAFETI against set ‘performance 
indicators’; the performance of bagda monoculture, bagda-golda rotational culture and golda-carp 
polyculture; adoption of SAFETI promoted management steps (6-StepsTowards IFP); and, farmers’ 
benefit-cost in shrimp/prawn cultivation; (ii) Project Management and Implementation pointing out 
issues of concern; and, (iii) Overall Assessment highlighting the evaluation questions of effectiveness, 
relevance, impact and sustainability.
Key Findings and Performance Analysis 
3.1.1 Achievements in performance indicators 
The SAFETI project has 16 PMP performance indicators concentrated in two broad areas – farm 
productivity and income, and trade; food safety is embedded on both the areas. This section of the 
report describes the status of each indicator in relation to set target for mid-term period (the 
compilation is attached in Annex-3). 
Performance indicator 1: Number of individuals who have applied new techniques or technologies.
Figure 1 shows that against the set target of 
10,128 individual who applied new techniques 
or technologies, MTE found the achievement be 
at 11,163 individuals, which was around 10% 
above the target. This indicator was further 
disaggregated in adoption categories by gender, 
shrimp/prawn high health PL, biosecurity, 
disease management, and other issues such 
postharvest. The participation and adoption of 
new technologies by women remained 
satisfactory, MTE estimation found it to be 
around 69 percent more (856) than the set 
target (507) for mid-term (Table 5, Annex-5)).
This indicator reflected the adoption and 
practice of six steps recommended and 
promoted by SAFETI (’6-StepsTowardsIFP’), which remained as the core of the intervention. Maintaining 
required pond depth was the first and foremost step among all other steps, but MTE found around 20%
of the demo and direct farmers still did not maintain the recommended 3-5 feet pond depth, and cases 
were observed where the demo farmers still did not, or could not meet the criteria. 
Figure 1: Number of individuals who applied 
new techniques or technologies
10,128 11,163 
22,200 
Mid-term
target
Mid-term
achievement
End-project
target
Table 5: Number of individuals who applied improved techniques or technologies on genetics, 
biosecurity, disease management and others 
Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
Shrimp/Prawn genetics 1,300 4,879 4,800 
Biosecurity management 8,096 9,185 17,150 
Disease management 8,096 8,132 17,150 
Others (Water treatment and harvesting-post harvesting) 2,032 1,978 5,050 
Final Report: Mid-term Evaluation of SAFETI Project 8 | P a g e
Among the six steps, SAFETI was extremely successful on adopting shrimp/prawn high-health PLs, in 
prima facie view MTE found that the achievement reached to 275% above the set target for mid-term
(adoption target 1,300 individuals versus achievement 4,879). However, careful analysis of this set 
target revealed that the SAFETI project had assumed that only about 13% of the targeted project 
beneficiaries would have access to and/or adopt high-health PLs of bagda/golda. This notable 
performance achievement was due to the support and collaboration of SAFETI with two hatcheries –
‘MKA’ and ‘Desh Bangla’ – which had resulted in a marked increase in SPF bagda PL production. In 
contrast to the adoption of high-health PLs, looking carefully at other areas of new technology adoption, 
in none of the cases were the demo and direct farmers found to have fully accepted all of the new 
technologies being promoted by SAFETI. Among the four indicators of new technology adoption, 
postharvest operations remained poorly adopted by the farmers, due in part to SAFETI’s strategic 
decision to address production constraints as a priority during the first half of the project.
SAFETI directed its beneficiary farmers to hatcheries to receive supplies of SPF golda PL. However, 
production of golda PLs has not been widely established in Bangladesh yet, with farmers still relying on 
supplies of PLs captured in the wild. Given this situation, SAFETI advocated and campaigned on the use 
of ‘healthy’ golda PLs. However, the judgement of what constitutes a ‘heathy’ golda PL still remains 
based on subjective visual observation. 
The farmers were highly receptive to the importance of biosecurity management which is evidenced by 
the target of SAFETI for this category was exceeded by 13.46% (target 8,096 versus achievement 9,185). 
Regarding addressing the disease problem, the project was on-target for the mid-term figure. SAFETI 
was not fully successful in attracting farmers on other management issues such as prescribed method of 
pond preparation. The reason for this was grounded in some structural and behavioral practices such as 
by land owners preventing farmers from deepening their ponds (to allow later culture of rice) and 
traditional practices involving shallow pond depths. Regarding this issue, farmers had been cultivating 
mostly on leased land for a duration of 2-3 years, and some farmers were evaluating the benefit-cost for 
investing in the deepening of ponds.
Performance indicator 2: Number of individuals receiving financial services.
The project fell short of the Mid-term target (3,750) for the indicator by 13.2%. According to the most 
recent data available from SAFETI MEL and partner organizations, TMSS provided 1246 (38% of total 
who received) and CODEC 175 (5% of total who received) loans to individuals. The remaining 1,833 loans 
(56% of total who received) were received from other MFIs. The MTE noted from FGDs that many 
farmers who received loans from MFIs had prior contract with loan-providing institutions prior to their 
involvement with SAFETI. The KII with TMSS and CODEC revealed that both partners were less interested 
in distributing loans to project farmers in areas outside of TMSS and CODEC traditional areas of loaning 
operations, but were willing to support 
farmers’’ on-ground production related 
activities. MFIs operate under the principle 
that the client should be permanently 
residing within 10 kilometers of MFIs’ service 
stations to facilitate collection of loan 
installment payments. Also, TMSS and 
CODEC did not have sufficient loaning 
services based in the SAFETI project areas. The overall figure for loan recipients indicates that total loan￾receiving farmers represented only 18.0% of the demo and direct farmers (considering the total of 
17,950) (Table 6). 
Table 6: Number of individuals receiving loans and number 
of loans disbursed Program indicator 
Mid-term 
target
Mid-term 
achievement
End-project 
target
Number of loans receiving 3,750 3,254 6,000
Number of loans disbursed 3,750 3,254 6,000
Value of loans provided $656,250 $1,342,109 $1,050,000
Final Report: Mid-term Evaluation of SAFETI Project 9 | P a g e
Performance indicator 3: Number of loans disbursed.
The SAFETI MEL PMP has established the same target for this indicator as for the previous one (Indicator 
2). There would have been a different result if the project partners had a dedicated program of loan 
disbursement to the beneficiary farmers.
Performance indicator 4: Value of loans provided.
According to SAFETI MEL information, the loans provided to beneficiary farmers were valued $1,342,109 
against the mid-term target of $656,250, which was more than double the target. The share of the value 
by TMSS was $481,862 (36% of the total number of loans disbursed), CODEC was $69,463 (5% of the 
total number of loans disbursed), and other MFIs was $790,884 (59% of the total number of loans
disbursed). SAFETI’s mid-term loan value target ($656,250) was exceptionally low compared to what 
farmers received ($1,342,109). SAFETI most likely underestimated the value of the loan amounts needed 
by individual households. The average value of the loans that farmers received was $387, $397 and $431 
from TMSS, CODEC and MFIs, respectively. Also, once the farmers received their required amount of 
loan from MFIs, they would generally continue to seek loans from them because they are already in the
MFIs’ systems and thereby were able to avoid going through the formal eligibility review process. The 
two MFIs that have been working with SAFETI need to customize their loan packages in accordance with 
the needs of smallholder farmers. Otherwise, the goal of ensuring access to adequate finance for the 
smallholder farmers may not be fully achieved. 
Performance indicator 5: Number of private 
enterprises, producer organizations, water 
users, women’s groups, trade & business 
associations, & community-based organizations 
(CBOs) that applied improved techniques and 
technologies.
Several distinct components are aggregated 
under this performance indicator, but are 
disaggregated by types of techniques or 
technologies. According to SAFETI MEL data, as 
of the mid-term period, achievement was 951 
against set target of 1023, which was 7.04% less 
than the mid-term target (Figure 2). Given the 
current momentum of the project’s progress, it 
seems that when it concludes the project will be able to achieve the life-of-project target.
Figure 2: Number of individuals of private sector 
actors used new techniques or technologies
1023 951
1504
Mid-term
target
Mid-term
achievement
End-project
target
Table 7: Number of individuals of private sector actors used new 
techniques or technologies
Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
Private enterprises 300 228 500
Producer organizations 720 718 1000
Trade & business associations 3 5 4
Final Report: Mid-term Evaluation of SAFETI Project 10 | P a g e
Performance indicator 6: Value of sales by project beneficiaries.
MTE found that value of sales by the project beneficiaries increased by 1% compared to the mid-term 
target. The project 
projected the yearly target 
of sales value at 
$28,556,599, whereas MTE
empirical estimation 
showed the worth value 
$28,756,403 (Table 8). For 
the two species, a 
downward achievement was observed for bagda, and an upward achievement for golda. The probable 
reason could be the higher number of training the farmers received from the project on bagda 
compared to golda (note performance indicator 8). The annual sales value for bagda was $14,178,727
compared with the mid-term target of $19,037,733 (25.52% shortfall), and for golda it was $14,577,676
compared with the mid-term target of $9,518,866 (53.15% over-achievement). Based on the overall 
achievement for this indicator, the MTE assumes that SAFETI’s intervention had been fully working. 
Performance indicator 7: Volume of commodities sold by project beneficiaries.
As with the increased value 
of sales, MTE estimated an
increase in the volume of 
bagda and golda sold by 
beneficiary (demo and direct) 
farmers against the mid-term 
target of 3,173 metric tons; 
the MTE calculated the value 
of sales to be 3,445 metric tons (Table 9), an over-achievement of 8.59%. However, when disaggregated 
by species, production volume showed a shortfall for bagda and an over-achievement for golda. The 
annual production value for bagda was 1,754 MT compared with the mid-term target of 1,903 MT 
(7.81% shortfall), and for golda was 1,691 MT compared with the mid-term target of 1,270 MT (33.16% 
over-achievement).
Table 8: Yearly value of sales by project beneficiaries
Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
Sale value of 'Bagda' and 'Golda' $28,556,599 $28,756,403 $88,240,810
Sale value of 'Bagda' only $19,037,733 $14,178,727 $58,827,207
Sale value of 'Golda' only $9,518,866 $14,577,676 $29,413,603
Table 9: Project beneficiaries ‘Bagda’ and ‘Golda’ sold volume (MT)
Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
sold volume of‘Bagda’ and ‘Golda’ 3,173 3,445 9,805
sold volume of ‘Bagda’only 1,903 1,754 5,881
sold volume of ‘Golda’ only 1,270 1,691 3,924
Final Report: Mid-term Evaluation of SAFETI Project 11 | P a g e
Performance indicator 8: Number of individuals who have received short-term agricultural sector 
productivity or food security 
training.
With SAFETI’s rigorous activities, 
this indicator achieved more than 
the mid-term target (31.28% over￾achievement) (Figure 3). The related technologies were predominantly related to farm management (‘6-
StepsTowardsIFP’). When disaggregated, the over-achievement was observed more with females
(98.41%) than males (27.75%) (Table 10). SAFETI training offered to golda farmers exceeded the 
midterm target by 82.83%, whereas the project was a little below the target (by 0.68%) for bagda 
farmers. The project was unable to provide training, due to engaging manpower more on capacity 
building of project (demo and direct) farmers, to expected number of governments, industries and 
organizations (36.67% shortfall), people in firms - processors, service providers, input dealers (22.33%
shortfall) and academic institutions (9.09% shortfall). However, the project was able to provide training 
to the targeted number of farmers who were the prime focus of the current intervention, and with the 
current momentum of the project, it seems the project should meet the life-of-project targets during its 
remaining time. 
Performance indicator 9: Number of 
individuals benefiting directly.
Achievement of this indicator lagged slightly 
behind the mid-term target, with 17,950
enlisted beneficiaries (demo and direct 
farmers) against the mid-term target of 18,470 
beneficiaries (Figure 4). A detailed analysis 
showed that the beneficiary women in this 
sector exceeded the mid-term target, with
about 49% more women benefitting than 
targeted.
Performance indicator 10: Number of 
individuals benefiting indirectly.
Figure 3: Number of individuals received short term agricultural 
sector productivity or food security training
18,470
24,247
26,300
Mid-term target Mid-term
achievement
End-project target
Figure 4: Number of directly beneficiaries form SAFETI project 
18,470 17,950
26,300
Mid-term target Mid-term
achievement
End-project target
Table 10: Number of individuals received short term agricultural 
sector productivity or food security training Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
Producers- farmers 18,000 23,876 25,000
'Shrimp/Bagda' Producers-farmers 10,800 10,700 15,000
'Prawn/ Golda' Producers-farmers 7,200 13,149 10,000
Input dealers 300 233 500
Policymaker 60 38 300
NGOs 110 100 500
Final Report: Mid-term Evaluation of SAFETI Project 12 | P a g e
This indicator measures the unassisted scaling effect 
of the project. For this indicator, SAFETI exceeded by 
34% the number of targeted indirect project 
beneficiaries; target 75,000 versus achieved 100,809
(Figure 5). This resulted mainly from distribution of 
posters highlighting management issues (‘6-
StepsTowardsIFP’), dissemination of SAFETI key 
messages through SAFETI-trained private 
organizations (such as feed companies), and 
interactions with agro-dealers and input suppliers. It 
is to be noted that the ‘shrimp app’ that is under 
development had not been released to the public; this 
would likely have increased the number of indirect beneficiaries.
Performance indicator 11: Number of policies, regulations and/or administrative procedures in each of 
the following stages of development.
The project achieved its mid-term target for formulating policies, regulations and/or administrative 
procedures. In relation to this indicator, BSSF-led and SAFETI-facilitated the ‘Development of SPF Black 
Tiger Shrimp’s hatchery Protocol’ was finalized and waiting for formalizing by the ministry, thereby 
reaching Stage 3 of the FFPr Policy Indicator’ (detailed in Annex-5). This protocol will guide production 
quality of PLs by hatcheries. The project greatly contributed to capability building of the Fish Inspection 
of Quality Control (FIQC) laboratory of DoF through human resource development and improved 
equipment. In turn, this helped provide more training (i.e., trainers) to farmers and shrimp/prawn 
related businesses owner on food safety. Disease diagnostics at the field level gained momentum led by 
WorldFish and collaborating partners – BFRI, DoF and Auburn University in the USA. Auburn had also 
been helping to improve feeding systems for bagda. The policy entitled ‘Amendment of Access to 
Institutional Finance for Shrimp and Fish Farmers through Amendment of Bangladesh Bank Agricultural 
and Rural Credit Policy and Program 2018-19’ achieved Stage 4 of the FFPr Policy Indicator. The policy 
entitled ‘Study of the Present Status of SPF Black Tiger Shrimp Hatcheries in Bangladesh’ achieved Stage 
3 of the FFPr Policy Indicator. The policy entitled ‘Develop Methodologies for Data Collection and 
Dissemination of Official Data to Facilitate the Availability of Key Market Information to Producers and 
Exporters achieved Stage 3 of the FFPr Policy Indicator. And, the policy entitled ‘Improvement of ‘Fish 
Hatchery Acts 2010 and Rules 2011’ achieved Stage 2 of the FFPr Policy Indicator. Details of the policies 
are attached in Annex-5.
Performance indicator 12: Average production of targeted commodities in kg per hectare in the 
targeted areas.
The MTE-estimated the 
productivity, on an annual 
cycle, of bagda at 664 kg ha-1
and for golda at 724 kg ha-1
, 
i.e. 47.60% and 60.65% (and 
on average 54.73%) more 
than the mid-term target)
(Table 11). The SAFETI MEL 
Figure 5: Number of indirectly beneficiaries form 
SAFETI project 
75,000
100,809
250,000
Mid-term
target
Mid-term
achievement
End-project
target
Table 11: Average production of bagda and golda (kg ha-1
)
Program indictor
Mid-term 
target
Mid-term 
achievement
End-project 
target
Average production of bagda and golda 450 696 500
Average production of bagda only 450 664 500
Average production of golda only 450 724 500
Final Report: Mid-term Evaluation of SAFETI Project 13 | P a g e
data showed this achievement in Year 2 (2017-18) to be 405 kg ha-1 and 626 kg ha-1
, respectively for 
bagda and golda. The MTE measured (figures cited above) the bagda productivity in the bagda 
monoculture system and the golda productivity in the golda-carp polyculture system. The high 
performance of bagda was achieved through farming of a second crop during the annual cycle that was 
practiced by 69% of the respondent farmers. Similarly, the high performance of golda was achieved 
through farming a second crop during the annual cycle that was practiced by almost all the respondent 
farmers. Although as has been previously noted, farming a second crop of bagda or golda comes with 
risks and also requires harvest of the first crop at smaller sizes, this result underlines the fact that when 
proper farming practices and management are exercised, a second crop can make an important 
contribution to the overall annual productivity of a farm. 
Final Report: Mid-term Evaluation of SAFETI Project 14 | P a g e
Performance indicator 13: Number of Specific Pathogen Free (SPF) and PCR screened Shrimp/PLs sold 
by hatcheries.
For bagda, disease is one of the 
major factors for low productivity. 
The risk of disease can be reduced by 
through the use of SPF PLs and/or 
PCR screened quality PLs. The data 
for this indicator revealed that 
hatcheries sold 738 million SPF and 
PCR-screened PLs against the mid￾term target of 500 million, despite the price of SPF PLs being considerably higher than non-SPF PLs
(Table 12). This was made possible because of SAFETI’s support, through advisory and connecting to 
government officials, to hatcheries, primarily MKA and Desh Bangla, who produced 340 million more 
SPF PLs in their most recent production cycle compared with the preceding cycle (i.e., prior to SAFETI 
interventions, as determined by KIIs). 
Performance indicator 14: Number of input suppliers/dealers that sell or supply improved 
inputs/services because of USDA assistance.
As promoted by the SAFETI project, the number of input sellers selling and/or supplying improved 
inputs/services (e.g., bleaches, feed 
etc.) increased in the market (Table 
13). This increase was 98% more 
than the mid-term target. Adoption 
of SAFETI activities triggered the 
increased demand from farmers, and 
in parallel, improved the 
understanding of the input sellers 
through SAFETI-delivered training 
that motivated service providers to sell improved inputs.
Performance indicator 15: Number of events to improve linkage between Buyers and Sellers.
The farmers maximize their benefits when they can sell their products directly to the buyers. The SAFETI 
project has 
successfully 
conducted four 
events to meet the 
mid-term target, but 
more such events are 
required (Table 14).
As part of SAFETI 
advocacy, a processing plant – Crimson Rosella - has agreed to collect shrimp/prawn directly from the 
farmers as long as the farmer (groups) can provide at least 100 kg from a single (group) location.
Performance indicator 16: Number of capacity building events for trade associations.
The trade associations are tertiary-level beneficiaries of the project. As of the mid-term period, the 
project had only conducted two capacity-building events with the trade associations against the mid￾term target of four events. Regarding, arranging capacity-building events for trade associations, the 
project needs to take rapid actions given that only half of the target has been achieved.
Table 13: Number of input suppliers/dealers that sell or supply 
improved inputs/services
Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
Input suppliers/dealers 60 179 100
Aqua Input suppliers 50 179 85
Shrimp PL nurseries 10 2 15
Table 14: Number of events to improve linkage between buyers and 
sellers Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
Improve linkage between Buyers and Sellers 4 4 10
Capacity building events for trade associations 4 2 10
Table 12: SPF and PCR tested PLs sold by hatcheries (million)
Program indicator
Mid-term 
target
Mid-term 
achievement
End-project 
target
SPF and PCR tested PLs 500 738 1,235
SPF PLs 400 703 1,100
PCR tested PLs 100 35 135
Final Report: Mid-term Evaluation of SAFETI Project 15 | P a g e
3.1.2 System productivity 
Figure 6 shows that there was a 193 and 
57 kg ha-1 productivity increase by direct 
farmers in bagda monoculture, and 
golda-carp polyculture system, 
respectively, over the control farmers. 
However, in one instance, the project’s 
direct farmers showed lower
productivity by 54 kg ha-1
for the bagda￾golda rotational system in comparison 
with the control farmers. This 
unexpected result dictates a follow-up 
investigation to determine why (i.e., 
causes for) the production of the direct 
farmers was less than the control 
farmers.
The performance of the demo and direct 
farmers considerably exceeded the 
overall mid-term target of average 450 
kg ha-1 production (for individual and 
combined species) for all three 
production systems. In contrast, with the 
exception of the bagda-golda rotational system, the production of control farmers was always lower 
than that of the demo and direct farmers. As a result, SAFETI has been able to show the potential for 
smallholder farmers of Bangladesh to increase the productivity of shrimp/prawn following the project￾promoted semi-intensive cultivation system.
The MTE analyses demonstrated 
consistently lower productivity with 
bagda during the wet season (213 – 294 
kg ha-1
) in the bagda monoculture 
system across all farmer-types (Figure 
7). The common reason was that it was 
more likely that low water salinity 
prevailed during the wet season as a 
result of the monsoonal rains adding 
freshwater to the farming system. The 
SAFETI 5th Semi-Annual Report (SAFETI 
Document 08) also suggested this 
correlation. Bagda does not grow well in 
a low-saline water environment. In 
addition, other causes such as improper 
pond preparation inviting biosecurity 
risk, lack of availability of good quality 
feed and other inputs, and shorter duration of the second crop in the systems might have impacted
bagda productivity. 
Figure 6: Annual productivity of bagda and/or golda in Bagda 
monoculture, Bagda-Golda rotational and Golda-Carp 
polyculture systems comparing the performance of three 
farmer-type, demo, direct and control in the SAFETI study 
area. 
761 791
713
660
481 430 467
535
373
0
200
400
600
800
1000
Bagda
monoculture
Bagda-Golda
rotational
Golda-Carp
polyculture
Culture system
Annual productivity (kg ha
-1
)
Demo farmer Direct farmer Control farmer
Figure 7: Comparative productivity of bagda in the Bagda
monoculture system under three farmer-type in the SAFETI 
study area. 
503
366
257 254
294
213
0
100
200
300
400
500
600
Demo Direct Control
Farmer-type
Seasonal productivity (kg ha
-1
)
Dry season Wet season
Final Report: Mid-term Evaluation of SAFETI Project 16 | P a g e
The productivity of bagda under bagda 
monoculture and bagda-golda rotational systems
varied, with higher productivity achieved in the 
latter system (Figure 8); the reason is likely due to 
the production environment having low salinity 
that was not suitable for bagda culture. As noted 
above, these results warrant further investigation. 
Golda performed much better in golda-carp
polyculture compared to the bagda-golda 
rotational system for all three farmers-types
(Figure 9). The reason might be due to the
production environment having a level of salinity 
that is not suitable for golda culture, as well as a 
reduced duration of the cropping period. 
In general, the ranking of productivity performance of 
shrimp/prawn under the three farmer-type was demo > direct > 
control. 
3.1.3 Management 
Adoption of farm management practices was 
more intense with demo farmers, followed by 
direct and lastly control farmers. This was 
probably reflected on productivity 
performance achieved by the farmer￾categories (reference Table 12). The MTE 
indexed the adoption of SAFETI promoted 
management (‘6-StepsTowardsIFP’) under 
four groups – biosecurity, disease, water, and 
harvest and postharvest. These steps 
accommodated nine issues in biosecurity
management, five issues in disease 
management, 10 issues in water 
management and five issues in harvest and 
postharvest management. Each issue was 
given a maximum score of 100, and the index 
for each management group under each 
farmer-type was calculated as the ratio of 
summation of the percent respondent’s 
affirmation of each issue by the summation 
of the maximum score of each issue. The 
indices for all management parameters was 
calculated in the same manner. The 
maximum index for individual and across all 
management groups was set as 1 (one). It 
may be noted that understanding the 
perceptions and priorities of the demo, direct 
and indirect farmers will help SAFETI 
Figure 9: Comparative productivity of golda, in the Bagda￾Golda rotational and Golda-Carp polyculture systems 
under three farmers-type in the SAFETI study area.
217
160
251
713
430
373
0
100
200
300
400
500
600
700
800
Demo Direct Control
Farmer-type
Culture system productivity
(kg ha
-1
)
Bagda-Golda rotational Golda-Carp polyculture
Figure 10: Adoption of SAFETI promoted management issues on four broad 
aspects – biosecurity, disease, water, and harvest and post-harvest –
indexing 0 (none) to 1 (maximum) under three farmer-type in the SAFETI 
study area.
0.87
0.73
0.64
0.43
0.70
0.79
0.63
0.59
0.37
0.63
0.69
0.54
0.54
0.42
0.56
0.0
0.2
0.4
0.6
0.8
1.0
Biosecurity Disease Water Harvest and
postharvest
Overall
Index of adoption of SAFETI promoted 
management (0 = None; 1 = Possible 
maximum) 
Broad management group
0 Demo Direct Control
100
200
300
400
500
600
0 20000 40000 60000 80000 100000
Productivity (kg ha
-1
)
PL density (number ha-1
)
Demo Direct Control
Figure 11: Relationship between PL density and productivity in 
bagda and golda, in the Bagda monoculture and Bagda-Golda 
rotational systems, under three farmers-type in the SAFETI study 
area.
R
2 = 0.29
Figure 8: Comparative productivity of bagda, in the Bagda monoculture and Bagda-Golda rotational systems, under 
three farmers-type in the SAFETI study area.
503
366
254
574
321 284
0
100
200
300
400
500
600
700
Demo Direct Control
Farmer-type
Culture system productivity
(kg ha
-1
)
Bagda monoculture Bagda-Golda rotational
Final Report: Mid-term Evaluation of SAFETI Project 17 | P a g e
determine what underrated elements of the six-step need to be emphasized during the remainder of 
the project
Across the three farmer-type, the adoption of biosecurity issues indexed the highest, followed by 
disease and water, and interestingly, the lowest on harvest and postharvest issues (Figure 10). This is 
likely because the farmers were at early stage of exposure of SAFETI technologies and concentrated on 
productivity rather than postharvest quality improvement issues. Results indicate around half of the 
control farmers had been considering SAFETI recommended farming practices in their farm 
management. It would be worth investigating the extent to which they have incorporated these 
measures. 
Regarding the current 
perception assessed by 
MTE, none of the 
farmer-types put priority 
weight on any one of the 
‘6-StepsTowardsIFP’
(See Table 15). The 
demo farmers gave 
highest priority to the 
maintenance of good
water quality and healthy shrimp/prawn (34.8% of the respondents), while for the direct farmers 
ensuring optimum water depth and cleaning the pond bottom was the most important of the six steps 
(40.2% of respondents). The control farmers rated stocking density and healthy PL as the prime method 
of productivity improvement (36.5% of respondents)..
Surprisingly, preventing cross-contamination by outside disease carriers was rated low in relative 
importance by demo 
and direct farmers 
(≤7.8%). SAFETI, on the 
other hand, remained 
quite certain that 
beneficiary farmers 
were aware of the 
threats from cross￾contamination of 
diseases, and surprised
that few interviewed farmers interviewed was aware of this. SAFETI needs to address this concern. 
The most difficulty the farmers experienced with the field application of the ‘6-StepsTowardsIFP’ was 
with the stocking density and healthy PL (expressed by 38.4%, 34.0% and 54.9% of demo, direct and 
control farmers, respectively) (Table 16). The next highest difficulty they encountered was related to 
biosecurity (i.e., prevent cross-contamination by outside disease carriers) and maintenance of good
water quality and healthy shrimp/prawn. A few farmers also faced the problem of quality feed as well. 
Table 15: Perception of smallholder farmers on SAFETI promoted steps towards
increased shrimp/prawn productivity in the field surveyed area SAFETI promoted Steps Demo Direct Control
Good water quality and healthy PL 34.8 14.1 24.6
Optimum depth and cleaning pond bottom 26.1 40.2 14.3
Stocking density and healthy PL 21.7 28.4 36.5
Quality feed 13 4.4 3.9
Prevent cross-contamination by outside disease carriers 4.3 4.8 14.8
Disinfecting water and removing potential threats of diseases 0 7.4 2
Farmer-type and respondent (%) Table 16: Difficulties on adopting SAFETI promoted by smallholder farmers in the
field surveyed area.
 SAFETI promoted steps Demo Direct Control
Good water quality and healthy PL 34.8 34 54.9
Optimum depth and cleaning pond bottom 21.7 14 15.4
Stocking density and healthy PL 17.4 16.5 13.8
Quality feed 17.4 8.8 3.1
Prevent cross-contamination by outside disease carriers 4.3 15.2 8.7
Disinfecting water and removing potential threats of diseases 0 6.9 4.1
Farmer-type and respondent (%)
Final Report: Mid-term Evaluation of SAFETI Project 18 | P a g e
The productivity of bagda and golda is related to PL 
stocking density; the higher the density stocked by 
the farmers using semi-intensive farming techniques, 
, the lower productivity they achieved (Figure 11). In 
general, the control farmers stocked the highest
density followed by direct then demo farmers. The 
SAFETI promoted management was semi-intensive 
and it recommended a stock density of around 140 PL 
per decimal, which was largely adopted by the demo 
farmers. It appears SAFETI, for smallholders farming 
environments, rightly recommended this PL density. 
Harvest size of shrimp/prawn is an important factor 
for attracting increased unit prices for the product. 
The frequency of harvest, within a growing season, 
likely affects the harvest size. Increased frequency of 
harvest results in smaller size at harvest, resulting in a 
reduced unit price. MTE observed this pattern in bagda selling price for both bagda monoculture and 
bagda-golda rotational culture (Figure 12). The demo farmers practiced less frequent harvests than 
direct farmers, and the control farmers harvested more frequently than the project farmers. In the 
bagda monoculture and bagda-golda rotational culture systems, the average frequency of harvest by the 
control farmers was alarmingly high at 12.5 and 9.9 harvests per cropping season, respectively (Figure 
12). Control farmers harvested this frequently mainly due to avoid the risk of crop failure, which was not 
uncommon due to the prevalence of diseases in farmed bagda.
3.1.4 Benefit-cost 
SAFETI promoted production 
technologies (‘6-StepsTowardsIFP’) 
need more investment from farmers 
in their production units. In return, 
the technology-adopting farmers 
can expect increased benefits from 
their investment. MTE analysis 
revealed that SAFETI benefitted 
farmers (both demo and direct 
farmers) exceeded benefits over 
costs (benefit-cost ratio, BCR >1) 
(Table 17). However, in both bagda 
monoculture and bagda-golda 
rotational systems, the BCR was 
lower for demo farmers than for 
direct farmers. Moreover, SAFETI
demo and direct farmers had 
smaller BCRs than control farmers.
Further analysis of these findings 
revealed that production cost largely related to, and explained the variation in the BCR (Figure 13). The 
demo farmers invested more, followed by direct farmers and then control farmers, whereas the exactly 
Table 17: Annualized and seasonal benefit-cost ratio in the three 
shrimp/prawn production systems under three farmer-type in the 
SAFETI study area 
Culture type Season Demo Direct Control
Dry season 1.6 2.2 3.6
Wet season 2.3 2.7 4.0
Annual 1.9 2.4 3.8
Dry season 2.3 2.8 4.6
Wet season 2.2 2.9 4.1
Annual 2.3 2.8 4.3
Dry season 2.9 2.7 -
Wet season 4.8 4.0 4.5
Annual 3.8 3.2 -
Golda-Carp polyculture
Farmer-type
Bagda monoculture
Bagda-Golda rotational
Figure 12. Relationship between the frequency of 
harvest of bagda and selling price in Bagda monoculture 
and Bagda-Golda rotational systems. Demo (835) Direct (666) Control (605) Demo (825) Direct (628) Control (629)
0
300
600
900
0 3 6 9 12 15
Selling price (BDT kg
-1
)
Number of harvest within a growing season
Bagda-Golda rotational Bagda monoculture
0
1
2
3
4
5
6
 - 50,000 100,000 150,000 200,000 250,000 300,000
Benefit-Cost Ratio
Production cost (BDT ha-1
)
'Control' 'Demo' 'Project'
R
2 = 0.62
Figure 13. Relationship between production cost and benefit-cost 
ratio in the three shrimp/prawn production systems under three 
farmer-type in the SAFETI study area.
Final Report: Mid-term Evaluation of SAFETI Project 19 | P a g e
reverse occurred with the BCR. This suggests that SAFETI-supported farmers should try to achieve 
further improvements in their production efficiency (i.e. increased productivity) to realize increased 
benefits from their investment.
However, 
investments of 
demo and direct 
farmers yielded 
more gross 
return compared 
with control 
farmers. In the 
bagda
monoculture system, demo farmers received a gross margin of BDT 491,344 ha-1
, compared with BDT 
328,568 ha-1 by direct farmers and BDT 268,458 ha-1 by control farmers. In the bagda-golda rotational
system, demo farmers received a gross margin of BDT 572,869 ha-1
, compared with BDT 383,052 ha-1 by 
direct farmers and BDT 505,248 ha-1 by control farmers. In the golda-carp polyculture system, demo 
farmers received a gross margin of BDT 747,766 ha-1
, compared with BDT 652,619 ha-1 by direct farmers 
and BDT 322,102 ha-1 by control farmers (Table 18).
The selling price of bagda differed 
between farmer-type, with demo 
farmers receiving a much higher price 
(+197 BDT kg-1
) than direct and 
control farmers for bagda 
monoculture, because of size of the 
produce. A similar pattern of selling 
price variation for bagda was 
observed in the bagda-golda 
rotational system, benefitting demo 
farmers (+169 BDT kg-1 more than 
direct farmers and +230 BDT kg
-1
more than control farmers).
The bagda price was lower in the wet-season culture than during dry-season culture (Table 19). This 
situation was probably due to restricted growth (echoed by the farmers in FGDs) because of a less than 
optimal environment due to reduced salinity level in the ponds. This may pose a question regarding the 
sustainability of wet-season bagda culture in such environments. One mitigative response might be not 
growing bagda in the season and delaying the dry-season crop to avoid disease risk.
The variation in the price of golda by farmer-type or growing season was small, suggesting that demo
farmers could not attract higher prices for this commodity than direct farmers.
Project Management, Partnership and MEL 
3.2.1 Project Management 
The Winrock International offices in the USA oversee the SAFETI project implementation and provides 
necessary guidance, concurrences and approvals to support the project’s central office based in Dhaka. 
An expatriate CoP leads the project team from the Dhaka office supported by a national deputy CoP, an 
expatriate Aquaculture expert, and a Director for Finance and Administration. Additionally, there are 
Table 18: Gross margin income in the three shrimp/prawn 
production systems under three farmer-type in the SAFETI study 
area 
Culture type Demo Direct Control
Bagda monoculture 491344 572869 747766
Bagda-Golda rotational 328568 383052 652619
Golda-Carp polyculture 268458 505249 322102
Farmer-type
Table 19: Sale Price (BDT kg-1
) of fish species in dry and wet seasons in 
three shrimp/prawn production systems under three farmer-type in the 
SAFETI study area Culture type Season Demo Direct Control
Dry season 825 628 629
Wet season 651 623 571
Dry season 835 666 605
Wet season 597 677 687
Dry season 676 702 664
Wet season 159 154 180
Farmer-type
Bagda monoculture
Bagda-Golda rotational
Golda-Carp polyculture
Final Report: Mid-term Evaluation of SAFETI Project 20 | P a g e
few program support professionals and a Monitoring, Evaluation and Learning Specialist. The SAFETI 
project has been implementing 10 different activities in association with implementation partners 
including NGOs, Worldfish, and Auburn University, USA. 
SAFETI has separate project implementation and technical support partners. Two implementation 
partners, TMSS and CODEC are responsible for field level implementation in two regions though their 
technical and implementation capacity varies. CODEC is a regional NGO with organizational experience 
mainly in working with fisherman in the coastal areas while TMSS is a national organization 
implementing multi-dimensional projects all over the country. TMSS operates in about 70% of the 
project areas. Each implementing partner has a Project Manager (PM) who is fully responsible for the 
SAFETI field activities; the PM has vehicle support (motorbike and fuel expenses) and mobile phone bill 
support. The staff member at the upazila level (UAEM) level also has a similar logistic support. The AAFs 
distributed in the upazilas are provided with bicycles and mobile phone bill support. The AFFs need to 
travel about 25-30 km to project areas. CODEC has two-day weekend, whereas TMSS has one-day 
weekend. 
The field staff turn-over rate observed was high (about 40% in the last one year) for both the 
implementing partners, which resulted in time and costs needed to hire and train new staff on a regular 
basis. However, TMSS and CODEC Project Managers felt this rate of turnover was not unexpected, given 
the circumstances in the current job market for project-based staff. It takes about 25-30 days to recruit, 
train and complete a staff position replacement.
SAFETI has partnerships with DoF, BFRI, Worldfish and Auburn University, USA for technical support 
including disease diagnosis. The evaluation team is aware of the technical collaboration but it seemed 
too early to assess collaboration effectiveness. It should be noted that the project greatly contributed to 
capability building of the Fish Inspection of Quality Control (FIQC) laboratory of DoF through human 
resource development and improved equipment.
The project signed collaboration MoUs with the DoF and BFRI in 2018. SAFETI has also Work Plans that 
are reviewed quarterly by a Project Technical Committee appointed by DoF and BFRI. At the field level, 
DoF official participates in farmers’ training program and provides technical assistance on requests. 
3.2.2 Monitoring, Evaluation and Learning (MEL) 
SAFETI uses a framework for conducting monitoring and evaluation activities. The project has 16 specific 
indicators for performance assessment, and therefore, a routine function of the MEL team is to keep 
track on project progress across these 16 indicators. To do so, SAFETI conducts annual sample surveys,
as well as compiling data from partner progress reports. 
During the process of activity implementation, a large volume of data is generated at various levels. 
Demo farmers are required to maintain a record book that they update with the assistance of the field 
staff when needed. Direct farmers are also provided with a record book, but maintaining it is not 
mandatory but rather intended simply to introduce them to record-keeping and encourage them to do 
so by providing them with the template that the demo farmers are using. The evaluation team observed 
that a large number of direct farmers are not aware of what is documented in the farmer’s record book 
which is quite comprehensive, and provides scope for self-check on farm activities. According to SAFETI 
plan, direct farmers are not required to maintain a record book – primarily because SAFETI does not 
have a sufficient number of field staff to help 25,000 farmers do this. 
Training is one of the main activities of the project. It provides step-by-step technical training to the 
farmers over a period of six months. As part of the training evaluation, farmers provide feedback based 
on which follow-up trainings are planned and implemented. The evaluation team observed that the 
Final Report: Mid-term Evaluation of SAFETI Project 21 | P a g e
farmer feedback had not been stored in the SAFETI database. The evaluation team expects that SAFETI 
would preserve this feedback as data for future reference.
Overall Assessment 
This section analyzes/synthesizes the findings and asseses the project performance in terms of 
effectiveness, relevance, impact and sustainability. 
3.3.1 Effectiveness 
The SAFETI project was launched in a time when the prawn/shrimp industry of Bangladesh was 
struggling in comparison with the progress made in freshwater fish production. The project assessed 
relevant stakeholders who had direct or indirect effect on the overall contribution of this sector in the 
economy. The project identified the factor that application of science in the production of shrimp/prawn 
would be a prime solution, and subsequently developed six steps that would help drive farmers towards 
more optimal productivity. Through both qualitative and quantitative surveys, the MTE found that the 
farmers have generally accepted the idea of six steps – although the adaptation rate by each category 
(i.e. farmer-type) is not uniform. There are several reasons for the less than full and disparate adoption 
of these six steps: among them, control over the pond and access to cash are impediments for the small 
holding farmers. Farmers appeared convinced that ensuring SPF PL was the primary factor for getting 
more production, and in this regard, the technical support of SAFETI two hatcheries had been able 
increase their production of SPF PL. However, still some direct farmers were stocking more PLs than the 
recommended density of 100-120 PLs per decimal for reasons based on anticipated mortality. The 
project engaged MFI-based project partners to ensure access to cash for the farmers, but the result was 
only partially successful. MTE data showed that the production of both golda and bagda has exceeded 
the mid-term targets, and based solely on the production of the demo farmers, it shows that achieving 
the project LOP target of 500 kg ha is likely and that in some cases (farms) 1,000 kg ha might be 
achievable. The project was able reach to the indirect farmers mainly through BCC and IEC materials. As 
result of SAFETI activity, demand for quality input-items has increased and more suppliers were found to 
be selling items of SAFETI- recommended quality. The project emphasized capacity-building of both 
farmers and other actors in the value chain. Farmers were found to be attending the training sessions, 
but MTE observed that some of them were unable to grasp the essence of the training up to the desired 
level of comprehension; examples of this were direct farmers who did not fully understand what was 
recorded in their log books. 
3.3.2 Relevance 
Relevance to context: Shrimp and prawn farming significantly contribute to Bangladesh’s national 
economy. But in recent years, the productivity of the sector has stagnated and external trade has 
declined. Bangladesh Government initiatives to address these concerns are reflected in the National 
Shrimp Policy (DoF, 2014), which clearly identifies the need to address pressing issues impeding 
improved productivity, ensured food safety and expanded trade. 
USDA’s ‘Program and Services’ emphasize on helping developing countries improve their agricultural 
systems and build their trade capacity. Its ‘Food for Progress Program’ has been designed to help 
developing countries and emerging democracies to modernize and strengthen their agricultural sectors
by adopting two principal objectives: to improve agricultural productivity and to expand trade of 
agricultural products (https://www.usda.gov/our-agency/about-usda).
The SAFETI project is thus fully in line with Bangladesh government’s policy plan and USDA’s broad 
strategy and objectives. 
Final Report: Mid-term Evaluation of SAFETI Project 22 | P a g e
Relevance to project 
design, activities and 
implementation: The field 
survey conducted by MTE, 
and validated in FGDs, 
revealed that the shrimp/prawn farmers were in dire need of receiving information related to their 
farming. This was because only few of them had previous had opportunity to undertake training on this 
topic (Table 20). Under these circumstances, the SATETI promoted ‘6-StepsTowardsIFP’ was fully
relevant. Given the present-day economic circumstances of the smallholder shrimp/prawn farmers, 
introducing and practicing expensive high-tech technologies such as mechanical aeration systems in 
small ponds would be a cost-burden. Thus, adoption of semi-intensive farming, as promoted by SAFETI, 
was the correct initial step under the conditions of limited information, training, techniques and capacity 
to adopt high-end (expensive) farming technology.
The MTE, in consultation with the beneficiary farmers during FGDs, received feedback that the field 
implementation structure of the SAFETI activities was relevant to socio-environmental circumstances of 
marginal landholders. The SAFETI activities include, but are not limited to: gaining and exchanging 
knowledge within the peer ‘group’ rather than waiting for the extension agents to help them; receiving 
technical help within the group rather than waiting for external help; and, and buying PLs under the 
banner of the ‘group’. The number of government fisheries extension agents serving at upazila level is
very limited and hinders frequent Government door-step services to farmers. This limitation on service 
delivery by Government extension agents is compounded in remote locations where travelling is not 
difficult and time consuming. 
Of the three critical issues (CS) that SAFETI addressed, this evaluation understood that productivity and
farm income (CS-1) and food safety (CS-2) would potentially contribute to trade (CS-3); whereas CS-2 
would have impact on CS-1. Recognizing the weak shrimp/prawn industry of Bangladesh, MTE believes
that the three critical issues pillared in the SAFETI design were fully relevant. However, in SAFETI’s short 
timeframe, it was impossible to achieve all targeted results, especially from CS-3. MTE noted that the 
while the project began in October 2016, field activities did not commence until September 2017, which 
left four-years (out of five) to produce the expected outputs. The outcome from CS-3, notably with 
respect to trade outside Bangladesh as emphasized by USDA, relies on CS-1 and significantly, on CS-2, 
which will require adequate transitional time. MTE feels that SAFETI should, in parallel, emphasize trade 
expansion within Bangladesh (internal trade). 
Relevance to gender: The MTE estimated women’s representation in the SAFETI project as direct 
beneficiaries at 7.7% (Table 17). This figure is similar to SAFETI MEL records (‘Program Indicator’ # 9, 
Annex-2). On the national scale, rural women own about 4% of the holdings, and the share of farm￾holding rural women (in proportion to men) engaged in agricultural work is reported as 27% (BBS, 2010). 
These figures for women’s representation are likely to be much smaller, compared with the national 
average, in shrimp/prawn farming communities. Therefore, the proportional benefits of SAFETI to 
women were reasonable. Furthermore, the evaluation did not find any evidence that project benefits 
were male-biased i.e. intentionally accruing differently to women. 
3.3.3 Efficiency 
In achieving its goal, SAFETI invested resources on in four areas: (i) general training and supply of 
training and technology materials; (ii) on-site technical support; (iii) assessment and monitoring of water 
quality and diseases; and (iv) connection with inputs and marketing. Those investments have mutually 
benefitted farmers and market actors (inputs suppliers, seed sellers, hatchery and/or nursery owners 
Table 20: Background capacity and Women beneficiaries (%) 
Background capacity and Women beneficiaries (%) Demo Project Control
Farmer training received other than offered by the SAFETI project 0.0 16.4 6.4
Direct project beneficiaries - Women of total beneficiaries 8.7 7.6 9.4
Farmer-type
Final Report: Mid-term Evaluation of SAFETI Project 23 | P a g e
and processor). Qualitative results, through KIIs of input suppliers and seed sellers, fully justified 
resources expended on them, notably through training, which helped them gain modern knowledge on 
the appropriate kinds and items that should be for shrimp/prawn culture. This resulted in win-win 
situation for them (i.e., their businesses) and for farmers (i.e., their commodity productivity). Processors, 
as revealed in KIIs, lacked a sufficient supply of raw shrimp/prawn in their factories to fill-in their unused 
production capacity; increased productivity in faming ponds is required to fill-in this gap.
The approved budget of SAFETI, which includes administrative (non-program) and program expenditures 
is USD 15,894,260.80. The administrative budget is 26.39% of the total budget. The administrative 
budget is paid for in part, from the CCC contribution to the budget. Because the administration of the 
project will run for 5 full years, the originally scheduled budget for administrative costs remains valid. 
The total expenditure up to June 30, 2019 was USD 7,484,231.13 resulting in a burn ratio of 47.07%. 
Therefore, at the mid-term stage, the project’s expenditure was on track, and the investment of project 
money resulted in a USD 3,607,791 return to beneficiary farmers.
3.3.4 Impact 
By the end of the project, SAFETI has targeted engaging 25,000 demo and direct beneficiaries, which 
would cover about 3% of the estimated 800,000 extensive shrimp/prawn farmers of Bangladesh. 
Changes in the productivity in those project beneficiary farms would have only a small impact on 
increased national production for the industry and expended trade. However, a large-scale expansion of 
the adoption of SAFETI methods and selected activities could make a measurable difference at a 
national level. For example, the 250,000 indirect beneficiaries, combined with the 25,000 demo and 
direct beneficiaries would account for about 28% of the estimated total number of farmers, which with 
respect to their contribution to the sector and the project’s outcome, could be significant. Ensuring the 
sustainably of SAFETI’s initiative could help make that happen.
3.3.5 Sustainability 
Ensuring sustainability should encompass two primary elements: (i) sustainable methods, and (ii) 
sustainable partnerships. Sustainability of the methods will depend on two main aspects: (i) perception 
of the farmers, and (ii) financial capacity of the farmers to implement. Farmers’ current perception on 
weighing the ‘6-StepsTowardsIFP’was confusing. Neither the demo nor the direct farmers fully 
understood the importance of the holistic view of the SAFETI-promoted steps on improving on-farm 
shrimp/prawn productivity. Demo farmers believed that ‘good water quality and healthy shrimp/prawn’ 
should be the best management option, followed by ‘optimum water depth and cleaning pond bottom’, 
and ‘stocking density and healthy PL’. In contrast, few direct farmers thought maintenance of ‘good
water quality and healthy shrimp/prawn’ would help boosting their shrimp/prawn productivity. 
Biosecurity and ‘disinfecting water and removing potential threats of diseases’ were poorly rated in 
relative importance by both demo and direct farmers. Based on these observations, MTE believes that 
the farmers’ perception of factors affecting farm productivity of shrimp/prawn may still not be clear to 
them. If this is the case, then it would raise serious concerns related to the sustainability of farm 
productivity. SAFETI needs to address this concern. 
The demo farmers demonstrated the potential for improved productivity for both bagda and golda 
cultivation using SAFETI promoted science-based production activities (‘6-StepsTowardsIFP’). Those 
farmers could be the potential future leaders at the cessation of the project. How to continue the role of 
the demo farmers in the absence of the project support is a future planning issue of SAFETI in the post￾mid-term stage. 
Final Report: Mid-term Evaluation of SAFETI Project 24 | P a g e
The study found, through field survey, FGDs and KIIs, the trust level of farmers to market actors –
hatchery, nursery, PL trader, input retailer, feed seller, ‘aratder’, and processor – was extremely high 
(90-100% affirmation from respondents). The project has contributed building this trust by connecting
the stakeholders through series of meetings and trainings. Ensuring good information and/or inputs flow 
from those actors will help build the sustainability of SATETI’s methods. This will require continued and 
increased partnerships with private sector, notably inputs and feed suppliers, hatchery owners and 
processors
On the issue of partnership, building and leaving strong partnerships with government (DoF and BFRI) 
when SAFETI concludes will also be a strong pillar of sustainability. This is especially important for 
diagnostic and disease management services, and channeling extension services to SAFETI built-up 
‘groups’.
4 Lessons Learned and Way Forward 
Bangladesh has made phenomenal success in production of fresh water fish production during 2008-
2017 period, whereas the marginal gain in the production of shrimp and prawn is minimal with exports 
declining during the 2017-18 period. The SAFETI project embarked on its activities when Bangladesh was 
in need of supporting the sector, particularly the small holder farmers whose livelihood is largely 
dependent on shrimp and prawn. Similar favorable climatic, soil, salinity condition, and geometric 
progression in our neighboring country India gives Bangladesh hope.
At the midpoint of the project, SAFETI has produced some good examples of progress towards improved 
productivity in the shrimp/prawn farming for smallholders in coastal Bangladesh. Notable ones are:
i. Raising the hopes of farmers that improved productivity of shrimp/prawn farming in 
smallholdings could become a reality.
ii. Upgrading farmers’ knowledge on avenues for increasing shrimp/prawn productivity through 
various methods and technology.
iii. Experiencing the mutual benefits of solving difficulties of farming activities through group￾environment (‘groups’).
iv. Achieving the target of 500 kg ha-1 annual bagda and golda productivity under semi-intensive 
(synonymous with ‘improved extensive’ system defined by DoF) culture system.
v. Quantifying the adoption rate of SAFETI-promoted farm management (production activities (‘6-
StepsTowardsIFP’) among farmer-types (demo, direct and control).
vi. Highlighting the importance of local service providers in disseminating knowledge and supplying 
appropriate inputs to farmers. 
vii. Increasing SPF PL production by hatcheries.
viii. Promoting mutual benefits of farmers and market actors through improved farm productivity 
activities.
Based on overall observations, the MTE suggest the following points as ways forward during the 
remaining period of project implementation:
i. Control of stocking is a critical factor and multiple stocking in the same pond should be 
discouraged. For a given pond infrastructure and practices, a stocking density of 100-120 PL per 
decimal should be emphasized.
ii. In the bagda monoculture system, the second (wet-season) bagda crop was not productive even 
with SPF stocks; this is due to a not-fully-favorable environment for bagda growth and increased 
Final Report: Mid-term Evaluation of SAFETI Project 25 | P a g e
biosecurity risk in the rainy season. Under such circumstances, failures with SPF seed could 
affect the credibility of SPF concept. SAFETI might suggest alternate cropping for the farmers in 
this season, such as growing rice.
iii. The first crop of bagda could begin from March, because high temperatures would provide 
natural protection against White Spot disease, assuming other environmental conditions are 
satisfactory. 
iv. The SPF bagda nursery supported by SAFETI helped minimize linkages and ensure supply of 
genuine SPF bagda seed to farmers. However, biosecurity in the nurseries needs strengthen and 
should be structure as an independent activity; duration of PLs in the nursery needs to be 
increased. Nurseries are an integral element of farming, but if nursery biosecurity is 
compromised, it can have devastating effects on the supply of SPF seeds to farmers.
v. Partial harvests are done at 25 g and onwards and at multiple times. SPF bagda grow best after 
30 g and coupled with low stocking density and a single harvest at > 40 g provides the best 
output and financial returns. SAFETI could make this a part of the BMP, where risk is not a 
serious issue.
vi. Farming a second crop with bagda is not as productive as the first crop, even with SPF stocks, 
due to increased biosecurity risk in the rainy season. SAFETI should suggest alternate crops (such 
as rice) for the farmers, given that the second crop yield is reduced and disease outbreaks are 
more pronounced.
vii. The golda crop in bagda-golda rotational farming is below par. This may be caused by residual 
salinity in the soil and water. Tidal flushing of pond water prior to stocking golda may prove 
effective. Where tidal flushing is not possible, SAFETI should explore other practical options, 
such as strictly following the practices of black soil removal, drying ponds as much as possible, 
liming and bleaching.
viii. Several products are currently suggested to the farmers to improve pond water quality. In the 
current farming system only feed, seed and water quality management should be promoted; 
there is a need to avoid introducing products that only drain the farmer’s finances. SAFETI could 
incorporate this in the BMP.
ix. Currently, there is no comprehensive quality control mechanism at the hatcheries. There is a 
need to establish ‘Process Controls’ and verification at the hatcheries for seed quality assurance 
up to the point of stocking in the farms.
x. To reduce production costs, SAFETI, along with BSFF and other organizations involved in shrimp 
market development, should continue dialogue with government (MoFL, DoF, NBR, and MoC) to 
(i) reduce import tax on feed and feed ingredients, machinery and other inputs used in shrimp 
and prawn hatcheries and farms; and (ii) synchronize shrimp/prawn farm electricity prices with 
prices in other agricultural sectors.
xi. The project should continue to facilitate and encourage establishment of local supplies of high￾quality shrimp feed to reduce the cost and dependency on imported feed.
xii. The SPF bagda nursery, promoted by SAFETI, will act as intermediator between hatcher and 
farmer, and ensure supply of genuine SPF ‘bagda’ seed to farmers. However, biosecurity in the 
nurseries should be improved, maintained and promoted as an essential independent entity. 
Nursery duration should be 30-40 days for both shrimp and prawn. Farmers themselves should 
adopt nursery practices, if facilities are available to them.
Final Report: Mid-term Evaluation of SAFETI Project 26 | P a g e
xiii. SAFETI would attract more indirect farmers by arranging more field days, vitalizing hatchery 
service for supplying SPF and PCR-tested PLs (validated through KII), and building the capacity of 
DoF extension staff. 
xiv. The project could establish a learning laboratory for case-based scientific investigation of both 
successes and failures, with all the demo farms serving as a learning unit. 
xv. SAFETI should organize groups of farmers to consolidate and sell their product directly to the 
buyers, through which the role of intermediaries will decline and farmers will obtain more 
benefit by saving time and transport costs.
xvi. Among the indicators of new technology adoption, postharvest operations remained poorly 
adopted by the farmers; SAFETI should emphasize disseminating this technology in the rest half 
of the project. 
xvii. SAFETI should encourage its partners to develop a dedicated loaning program and increased 
areas of operations to allow more farmers to access loans. 
xviii. SAFETI can encourage the Government to ensure ‘Process Controls’ and verification in 
hatcheries for seed quality assurance up to the point of stocking in the producer’s farms.
xix. Farmer training through the use of audio-visual aids could be more effective, given the farmers’ 
habitual weakness in reading.
xx. Given the high turnover of field-based staff, SAFETI should encourage its partners, TMSS and 
CODEC to improve their strategies for retaining field staff. 
xxi. According to SAFETI plan, direct farmers are not required to maintain a record book – primarily 
because SAFETI does not have a sufficient number of field staff to help 25,000 farmers do this. 
But, SAFETI certainly can look at ways of expanding or encouraging direct farmers to use the 
book. For example, SAFETI can ask the demo farmers to promote this use and use them as 
trains for their group of direct farmers. If done, these data can easily be entered to the project 
monitoring system. 
xxii. SAFETI should encourage project partners to engage fisheries-educated individuals as 
Aquaculture Field Facilitators (AFFs). 
xxiii. The SAFETI project might initiate dialogue with local government (LGD) on creating and/or 
developing local water supplies and drainage facilities.
xxiv. SAFETI project partners should involve LEAFs located near project areas. The LEAFs may be 
trained on SAFETI technologies to support local farmers in concert with the AFFs.
Final Report: Mid-term Evaluation of SAFETI Project 27 | P a g e
5 References 
Aakkula J, Artell J and Heini T. 2013. Strengths and weaknesses of Counterfactual Evaluation. MTT 
Agrifood Research Finland. INEA, ROME: ENVIEVAL Grant Agreement Number 312071.
Ahmed N, Demaine H and Muir JF. 2008. Freshwater prawn farming in Bangladesh: history, present 
status and future prospects. Aquacult. Res. 39, 806–819.
Alauddin M and Hamid MA. 2014. Shrimp Culture in Bangladesh with Emphasis on Social and Economic 
Aspects. https://www.researchgate.net/publication/43483037. 
Ayaz M, Sumara K, Vadher KH and Ishakani AH. 2015. Effect of higher salinities on growth and survival of 
pacific white shrimp, Litopenaeus vannamei (Boone, 1931). Journal of Applied and Natural 
Science, 7: 885 – 888.
Belton B, Karim M, Thilsted S, Murshed-E-Jahan K, Collis W and Phillips M. 2011. Review of aquaculture 
and fish consumption in Bangladesh. Studies and Reviews 2011-53. The WorldFish Center. 
November 2011.
BBS 2010. Census of Agriculture 2008 Structure of Agricultural Holdings and Livestock Population 
Volume-1. Bangladesh Bureau of Statistics Division, Ministry of Planning Government of The 
People’s Republic of Bangladesh Dhaka, Bangladesh.
BBS 2018. Bangladesh Statistics 2018. Bangladesh Bureau of Statistics Division, Ministry of Planning 
Government of The People’s Republic of Bangladesh Dhaka, Bangladesh.
BBS 2019. Report on Agriculture and Rural Statistics 2018. Statistics and Informatics Division, Bangladesh 
Bureau of Statistics, Ministry of Planning, Government of Peoples Republic of Bangladesh.
BOBP 1992. Feeds for artisanal shrimp culture in India Their development and evaluation. Studies 
conducted in collaboration with Central Institute of Brackishwater Aquaculture, Madras, and 
Directorate of Fisheries, Andhra Pradesh. Bay of Bengal Programme (BOBP). Madras, India.
Broju GP and Vogl CR. 2011. Impacts of Shrimp farming in Bangladesh: Challenges and Alternatives. 
Ocean and coastal Management, 54: 201-211.
Chowdhury MA and Khairunb Y. 2014. Farmers ‘Local knowledge in Extensive Shrimp Farming Systems in 
Coastal Bangladesh. APCBEE Procedia, 8:125-130.
CIBA 2011. Better Management Practices (BMPs) and Group Management for Empowering Small Scale 
Farmers: Scaling Up Strategies. National Workshop Report - EU-FP7 Project “ASEM Aquaculture 
Platform” ICAR/MPEDA/MOA/EU-ASEM/NACA. Network of Aquaculture Centers in Asia-Pacific. 
Central Institute for Brackish water Aquaculture (CIBA), Chennai, India.
DoF 2016. National fish week, compendium (In Bengali). Dhaka: Department of Fisheries, Ministry of 
Fisheries and Livestock, Government of Bangladesh.
DoF 2018a. Year Book of Fisheries Statistics of Bangladesh 2017-18. www.fisheries.gov.bd. 
DoF 2018b. Annual Report 2017. Department of Fisheries, Ministry of Livestock and Fisheries, 
Bangladesh. www.fisheries.gov.bd .
MoC 2018. Export Policy, 2018. Ministry of Commerce. https://mincom.gov.bd/
MoFL 2011. Fish Feed Rules, 2011. Ministry of Livestock and Fisheries, Bangladesh. https://mofl.gov.bd/ 
MoFL 2017. Fish and Fish product (inspection and quality control) Ordinance, 1983.Ministry of Fisheries 
and Livestock. https://mofl.gov.bd/
MoFL 2011. Fish hatchery Rules, 2010.Ministry of Fisheries and Livestock. Ministry of Fisheries and.
MoFL. 1950. Fish Protection and Conservation Act, 1950 and rules1985.Ministry of Fisheries and 
Livestock. Ministry of Fisheries and Livestock. https://mofl.gov.bd/
Final Report: Mid-term Evaluation of SAFETI Project 28 | P a g e
MoFL.2018.Fisheries Quarantine Act, 2018. Ministry of Fisheries and Livestock. Ministry of Fisheries and 
Livestock. https://mofl.gov.bd/
MoFL.2014. National Shrimp Fish Policy, 2014. Ministry of Livestock and Fisheries, Bangladesh. 
https://mofl.gov.bd/
European Commission. 2013. Design and Commissioning of Counterfactual Impact Evaluations - A 
Practical Guidance for ESF Managing Authorities. Directorate General for Employment, Social 
Affairs and Inclusion Unit A3. European Commission. http://europa.eu. 
Khan Firoz AKM, Mustafa MG and Naser MN. 2016. Effective Supervision Inland captures fisheries of 
Bangladesh and its hurdles in managing the resources. Bandung Journal the Global South, 3:17.
Kumar B, Sharma R, Lakra WS, Sharma A, Prakesh S and Sharma MM. 2016. Economic Assessment of 
Shrimp Farming (Litopenaeus vannamei) in Gujarat - A Profitable Venture. International Journal 
of Innovative Research in Science, Engineering and Technology, 5:15334-15342. 
Nesar A and Diana JS.2015. Coastal to inland: Expansion of Prawn farming for adaptation to climate 
change in Bangladesh. Aquaculture reports, 2: 67-76.
Nesar A, Troell M, Allison EH and Muir JF.2010. Prawn post larvae fishing in coastal Bangladesh: 
challenges for sustainable livelihoods. Marine Policy, 34:218-227. 
Paul AK and Roskaft E. 2013. Environmental degradation and loss of traditional agriculture as two causes 
of conflicts in shrimp farming in the southeastern coastal Bangladesh: present status and 
probable solutions.
SAFETI Document 01: Planning documents for reaching indirect farmers, Supplied by SAFETI project.
SAFETI Document 02: Power Point Presentation 1 for MTE team, Supplied by SAFETI project.
SAFETI Document 03: Power Point Presentation 2 for MTE team, Supplied by SAFETI project
SAFETI Document 04: Power Point Presentation for CBSG enumerators, Supplied by SAFETI project.
SAFETI Document 05: Project Mid-Term Evaluation-SOW, Supplied by SAFETI project.
SAFETI Document 06: Project Performance Indicators with Annual Targets, Supplied by SAFETI project.
SAFETI Document 07: SAFETI Project Evaluation Plan, Supplied by SAFETI Project.
SAFETI Document 08: 5th Semi-Annual Performance Report (October 1 – March 31, 2019), Supplied by 
SAFETI Project.
Shamsuzzaman MM, Islam MM, Tania NJ, Al-Mamun MA, Barman P and Xu X. 2017. Fisheries resources 
of Bangladesh: present and future direction. Aquaculture and Fisheries, 2: 145-156.
Srinivas D, Venkatrayulu C and Swapna B. 2016. Sustainability of exotic shrimp Litopenaeus vannamei
(Boone, 1931) farming in coastal Andhra Pradesh, India: Problems and Issues. European Journal 
of Experimental Biology, 6:80-85. www.pelagiaresearchlibrary.com.
The Daily Star. 2014. New shrimp farming system to boost yields. https://www.thedailystar.net/new￾shrimp-farming-system-to-boost-yield. 
The WoldFish Centre.2013. Greater Final Program Performance Report on Harvest and Economic 
Returns from Shrimp (GHERS). worldfishcenter@cgiar.org
The WorldFish Center.2011. Review of aquaculture and fish consumption in Bangladesh. 
worldfishcenter@cgiar.org 
USDA. 2018. USDA Strategic goals and their objectives. https://www.usda.gov/our-agency/about-usda.
Final Report: Mid-term Evaluation of SAFETI Project 29 | P a g e
Annexure 
Annex-1: List of persons met and/or consulted during mid-term evaluation of the SAFETI project
[list can be made available to USDA upon request to Winrock International]
Draft Report: Mid-term Evaluation of SAFETI Project 30 | P a g e
Annex-2: Distribution of sampling units, used for field data collection for mid-term evaluation of SAFETI, across unions and upazilas in the 
SAFETI operated areas. 
District Upazila Union Environment
Cultur
e 
system
Demo 
farmer
Project 
farmer
Control 
farmer
Khulna Dumuria Rudaghara Fresh water - Low saline BB 1 0
Khulna Dumuria Sobhana Saline BB 1 17
Khulna Paikgachha Chandkhali & Laskar Saline BB 1 17
Satkhira Assasuni Budhhata Fresh water - Low saline BB 1 23
Satkhira Assasuni Budhhata Brackish water - Medium saline BB 1 23
Satkhira Debhata Noapara Fresh water - Low saline BB 0 1
Satkhira Debhata Noapara Brackish water - Medium saline BB 2 40
Satkhira Kaliganj Nalta Brackish water - Medium saline BB 1 22
Satkhira Shyamnagar Shyamnagar Saline BB 1 21
Bagerhat Mongla Chandpai 
&Mithakhali
Dry medium saline, wet rainwater low 
saline
BB 41
Satkhira Satkhira Sadar Dhulihor Brackish water - Medium saline BB 33
Bagerhat Bagerhat Sadar Shat Gambuj Brackish water - Medium saline BG 2 42
Bagerhat Rampal Ujalkur Fresh water - Low saline BG 0 2
Bagerhat Rampal Rampal Saline BG 1 22
Khulna Dumuria Rudaghara Brackish water - Medium saline BG 1 22
Khulna Dumuria Sobhana Saline BG 0 5
Khulna Paikgachha Chandkhali & Laskar Saline BG 1 27
Satkhira Assasuni Budhhata Brackish water - Medium saline BG 1 21
Draft Report: Mid-term Evaluation of SAFETI Project 31 | P a g e
District Upazila Union Environment
Cultur
e 
system
Demo 
farmer
Project 
farmer
Control 
farmer
Satkhira Debhata Noapara Brackish water - Medium saline BG 0 4
Satkhira Debhata Noapara Fresh water - Low saline BG 0 1
Satkhira Kaliganj Nalta Brackish water - Medium saline BG 1 21
Satkhira Kaliganj Nalta Fresh water - Low saline BG 0 23
Bagerhat Mollarhat Udaypur Very low saline BG 2
Bagerhat Mongla Chandpai 
&Mithakhali
Dry medium saline, wet rainwater low 
saline
BG 17
Khulna Batiaghata Jalma Dry medium to low saline, wet rainwater 
low saline
BG 22
Satkhira Satkhira Sadar Dhulihor Dry medium to low saline, wet rainwater 
low saline
BG 24
Bagerhat Bagerhat Sadar Kara Para Fresh water - Low saline GP 1 24
Bagerhat Chitalmari Chitalmari Fresh water - Low saline GP 1 20
Bagerhat Rampal Ujalkur Fresh water - Low saline GP 1 19
Jashore Abhaynagar Sundoli Fresh water - Low saline GP 2 46
Khulna Dumuria Rudaghara Fresh water - Low saline GP 0 22
Satkhira Assasuni Baradal Fresh water - Low saline GP 1 21
Satkhira Debhata Noapara Fresh water - Low saline GP 1 19
Bagerhat Mollarhat Chunkhola &Udaypur Very low saline GP 64
Total 23 525 203
Draft Report: Mid-term Evaluation of SAFETI Project 32 | P a g e
Draft Report: Mid-term Evaluation of SAFETI Project 33 | P a g e
Annex-3: Status on the progress of program indicators of the SAFETI project as evaluation by the mid-term evaluation team. 
# Result 
Reference
Performance Indicator Mid-term 
project target
MTE measured 
achieved target
Progress (%) MTE measurement 
source (s)
1 FFPr 1.2 Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance 
10,128 11,163 10.22 Sample field survey
Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance (Male)
9,617 10,307 7.17 Sample field survey
Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance (Female)
507 856 68.88 Sample field survey
Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance (Shrimp/Prawn genetics)
1,300 4,879 275.32 Sample field survey
Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance (Biosecurity management)
8,096 9,185 13.46 Sample field survey
Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance (Disease management) 
8,096 8,132 0.44 Sample field survey
Number of individuals who have applied new 
techniques or technologies as a result of USDA 
assistance (Other) 
2,032 1,978 -2.68 Sample field survey
2 FFPr 1.2.3 Number of individuals receiving financial services 
as a result of USDA assistance 
3,750 1,421 -62.11 SAFETI MEL & KII 
(TMSS and CODEC)
3 FFPr 1.2.3 Number of loans disbursed as a result of USDA 
assistance 
3,750 1,421 -62.11 SAFETI MEL & KII 
(TMSS and CODEC)
4 FFPr 1.2.3 Value of loans provided as a result of USDA 
assistance 
$656,250 $551,325 -15.99 SAFETI MEL
5 FFPr 1.4.4 
&2.4.4
Number of private enterprises, producer 
organisations, water users, women’s groups, 
trade & business associations, & community￾1023 951 -7.04 SAFETI MEL
Draft Report: Mid-term Evaluation of SAFETI Project 34 | P a g e
# Result 
Reference
Performance Indicator Mid-term 
project target
MTE measured 
achieved target
Progress (%) MTE measurement 
source (s)
based organizations (CBOs) that applied 
improved techniques and technologies as a result 
of USDA assistance
Number of private enterprises, producer 
organizations, water users, women’s groups, trade 
& business associations, & community-based 
organizations (CBOs) that applied improved 
techniques and technologies as a result of USDA 
assistance (Private Enterprises)
300 228 -24.00 SAFETI MEL
Number of private enterprises, producer 
organizations, water users, women’s groups, trade 
& business associations, & community-based 
organizations (CBOs) that applied improved 
techniques and technologies as a result of USDA 
assistance (Producer Organizations)
720 718 -0.28 KII (TMSS and 
CODEC)
Number of private enterprises, producer 
organizations, water users, women’s groups, trade 
& business associations, & community-based 
organizations (CBOs) that applied improved 
techniques and technologies as a result of USDA 
assistance (Trade and Business Associations)
3 5 66.67 SAFETI MEL
6 SO2 Value of sales by project beneficiaries [yearly 
value] 
$28,556,599 $28,756,403 0.70 Sample field survey 
& SAFETI farmers' 
registration data
Value of sales by project beneficiaries 
(Shrimp/Bagda)
$19,037,733 $14,178,727 -25.52 Sample field survey 
& SAFETI farmers' 
registration data
` Value of sales by project beneficiaries (Prawn/ 
Golda)
$9,518,866 $14,577,676 53.15 Sample field survey 
& SAFETI farmers' 
registration data
Draft Report: Mid-term Evaluation of SAFETI Project 35 | P a g e
# Result 
Reference
Performance Indicator Mid-term 
project target
MTE measured 
achieved target
Progress (%) MTE measurement 
source (s)
7 SO1 Volume of commodities (metric tons) sold by 
project beneficiaries (MT)
3,173 3445 8.59 Sample field survey 
& SAFETI farmers' 
registration data
Volume of commodities (metric tons) sold by 
project beneficiaries (Shrimp/Bagda) (MT)
1,903 1,754 -7.81 Sample field survey 
& SAFETI farmers' 
registration data
Volume of commodities (metric tons) sold by 
project beneficiaries (Prawn/ Golda) (MT)
1,270 1,691 33.16 Sample field survey 
& SAFETI farmers' 
registration data
8 SO1 Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
18,470 24247 31.28 SAFETI farmers' 
registration data
Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
(Male)
17,547 22416 27.75 SAFETI farmers' 
registration data
Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
(Female) 
923 1831 98.41 SAFETI farmers' 
registration data
Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
(Producers- Farmers)
18,000 23876 32.64 SAFETI farmers' 
registration data & 
SAFETI MEL
Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
(Shrimp/Bagda Producers- Farmers)
10,800 10727 -0.68 SAFETI farmers' 
registration data & 
SAFETI MEL
Number of individuals who have received short￾term agricultural sector productivity or food 
7,200 13149 82.63 SAFETI farmers' 
registration data & 
SAFETI MEL
Draft Report: Mid-term Evaluation of SAFETI Project 36 | P a g e
# Result 
Reference
Performance Indicator Mid-term 
project target
MTE measured 
achieved target
Progress (%) MTE measurement 
source (s)
security training as a result of USDA assistance 
(Prawn/ Golda Producers- Farmers)
agricultural sector productivity or food security 
training as a result of USDA assistance (People in 
Firms - Processors, Service Providers, Input 
Dealers)
300 233 -22.33 SAFETI farmers' 
registration data & 
SAFETI MEL
Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
(People in Government- extension workers, 
policymakers)
60 38 -36.67 SAFETI farmers' 
registration data & 
SAFETI MEL
Number of individuals who have received short￾term agricultural sector productivity or food 
security training as a result of USDA assistance 
(People in Civil Society - NGOs staff, Academic 
Institutions)
110 100 -9.09 SAFETI farmers' 
registration data & 
SAFETI MEL
9 SO1 + 
SO2
Number of individuals benefiting directly as a 
result of USDA assistance
18,470 17950 -2.82 KII (TMSS and 
CODEC)
Number of individuals benefiting directly as a 
result of USDA assistance (Male)
17,547 16573 -5.55 KII (TMSS and 
CODEC)
Number of individuals benefiting directly as a 
result of USDA assistance (Female)
923 1377 49.16 KII (TMSS and 
CODEC)
10 SO1 + 
SO2
Number of individuals benefiting indirectly as a 
result of USDA assistance
75,000 100,809 34.41 SAFETI MEL & KII 
with DoF and 
Hatchery Owner 
(MK & Desh Bangla)
11 FFPr 1.4.2 
& 2.4.2
Number of policies, regulations and/or 
adminstrative producedures in each of of the 
following stages of development as a result of 
USDA assistance
1 1 0.00
Draft Report: Mid-term Evaluation of SAFETI Project 37 | P a g e
# Result 
Reference
Performance Indicator Mid-term 
project target
MTE measured 
achieved target
Progress (%) MTE measurement 
source (s)
Stage 1: Underwent the first stage of the policy 
reform process i.e. analysis (review of existing 
policy/ regulation/ administrative procedure 
and/or proposal of new 
policy/regulation/administrative procedure.)
0 1 0.00
12 SO1 Average production of targeted commodities in 
kg per hectare in the targeted areas
450 696 54.73 Field survey data
Average production of targeted commodities in kg 
per hectare in the targeted areas (Shrimp/Bagda)
450 664 47.60 Field survey data
Average production of targeted commodities in kg 
per hectare in the targeted areas (Prawn/ Golda)
450 724 60.95 Field survey data
13 SO2 Number of Specific Pathogen Free (SPF) and PCR 
screened Shrimp/PLs sold by hatcheries (Million)
500 738 47.60 SAFETI MEL
Number of Specific Pathogen Free (SPF) 
Shrimp/PLs sold by hatcheries (Million)
400 703 75.75 SAFETI MEL
Number of PCR screened Shrimp/PLs sold by 
hatcheries (Million)
100 35 -65.00 SAFETI MEL
14 SO2 Number of input suppliers/dealers that sell or 
supply improved inputs/services because of USDA 
assistance.
60 179 198.33 SAFETI MEL
Aqua Input suppliers (Feed, fertilizer and other 
inputs suppliers)
50 179 258.00 SAFETI MEL
Shrimp PL nurseries 10 2 -80.00 SAFETI MEL
15 SO2 Number of events to improve linkage between 
Buyers and Sellers
4 4 0.00 SAFETI MEL
16 SO2 Number of capacity building events for trade 
associations
4 2 -50.00 SAFETI MEL
Draft Report: Mid-term Evaluation of SAFETI Project 38 | P a g e
Annex-4: Mid-term and end-project targets achieved by the SAFETI project as evaluated in the mid-term evaluation
Figure A-4.1: Mid-term and end-project targets achieved by the SAFETI program indicators – indicators 1-8.
PI-1: Number of individuals who have applied new techniques or technologies
PI-2: Number of individuals receiving financial services
PI-3: Number of loans disbursed
PI-4: Value of loans provided
PI-5: Number of private enterprises, producer organizations, water users,
women’s groups, trade & business associations, & community-based
organizations (CBOs) that applied improved techniques and technologies
PI-6: Value of sales by project beneficiaries [yearly value]
PI-7: Volume of commodities (metric tons) sold by project beneficiaries (MT)
PI-8: Number of individuals who have received short-term agricultural sector
productivity or food security training
PI-9: Number of individuals benefiting directly
PI-10: Number of individuals benefiting indirectly
PI-11: Number of policies, regulations and/or administrative procedures in
each of the following stages of development
PI-12: Average production of targeted commodities in kg per hectare in the
targeted areas
PI-13: Number of Specific Pathogen Free(SPF) and PCR screened Shrimp/PLs
sold by hatcheries (Million)
PI-14:Number of input suppliers/dealers that sell or supply improved
inputs/services
PI-15: Number of events to improve linkage between Buyers and Sellers
PI-16: Number of capacity building events for trade associations
10
-13 -13
105
-7
1
9
31
-50 -46 -46
28
-37
-67 -65
-8
10
-13 -13
105
-7
1
9
31
-50 -46 -46
28
-37
-67 -65
-8
-100
-50
0
50
100
150
200
250
PI-1 PI-2 PI-3 PI-4 PI-5 PI-6 PI-7 PI-8
Achievement as per mid-term and end- project targets 
Performance indicator
Mid-term End-project Mid-term End-project
Draft Report: Mid-term Evaluation of SAFETI Project 39 | P a g e PI-1: Number of individuals who have applied new techniques or technologies
PI-2: Number of individuals receiving financial services
PI-3: Number of loans disbursed
PI-4: Value of loans provided
PI-5: Number of private enterprises, producer organizations, water users,
women’s groups, trade & business associations, & community-based
organizations (CBOs) that applied improved techniques and technologies
PI-6: Value of sales by project beneficiaries [yearly value]
PI-7: Volume of commodities (metric tons) sold by project beneficiaries (MT)
PI-8: Number of individuals who have received short-term agricultural sector
productivity or food security training
PI-9: Number of individuals benefiting directly
PI-10: Number of individuals benefiting indirectly
PI-11: Number of policies, regulations and/or administrative procedures in each
of the following stages of development
PI-12: Average production of targeted commodities in kg per hectare in the
targeted areas
PI-13: Number of Specific Pathogen Free(SPF) and PCR screened Shrimp/PLs
sold by hatcheries (Million)
PI-14:Number of input suppliers/dealers that sell or supply improved
inputs/services
PI-15: Number of events to improve linkage between Buyers and Sellers
PI-16: Number of capacity building events for trade associations
Figure A-4.2: Mid-term and end-project targets achieved by the SAFETI program indicators – indicators 9-16.
-3
34
0
55 48
198
0
-50 -50 -46 -46
28
-37
-67 -65
-8
10
-13 -13
105
-7
1
9
31
-50 -46 -46
28
-37
-67 -65
-8
-100
-50
0
50
100
150
200
250
PI-9 PI-10 PI-11 PI-12 PI-13 PI-14 PI-15 PI-16
Achievement as per mid-term and end- project targets 
Performance indicator
Mid-term End-project Mid-term End-project
Draft Report: Mid-term Evaluation of SAFETI Project 40 | P a g e
Annex-5: Targeted policies and stages of development of SAFETI project’s program indicator 11
Sl. No. Policy works Status Follow-Up
01 Development of SPF Black Tiger 
Shrimp’s hatchery protocol. 
The assessment and recommended processes constitute guidelines and 
administrative processes that quality this as a policy instrument. We have a 
letter from BSFF dated 13 July 2018 that formally submits these 
recommended protocols it to DoF. Therefore, it has reached Stage 3 of the 
FFPr Policy Indicator
Year/Stage achieved
Baseline=2; Y1=0; Y2=0; Y3=3; Y4=4; Y5=0
DOF has initiated the 
Government approval 
process. SAFETI will monitor 
ongoing progress. 
02
Amendment of access to 
institutional finance for shrimp 
and fish farmers through 
amendment of Bangladesh Bank 
Agricultural & Rural Credit Policy 
and Program 2018-19.
Amendment done by the Bangladesh Bank as documented on 25 July 2018. 
Therefore, it has reached Stage 4 of the FPPr Policy indicator. Therefore, it 
has reached Stage 4 of the FFPr Policy Indicator
Year/Stage achieved
Baseline=2; Y1=0; Y2=0; Y3=3&4; Y4=0; Y5=0
Need to arrange 
dissemination and awareness 
programs in the project 
region. 
Need further improvement in 
Business plan for the access to 
finance from Bank.
03
Study of the present status of 
SPF Black Tiger Shrimp’s 
hatcheries in Bangladesh.
The assessment and recommended processes constitute guidelines and 
administrative processes that quality this as a policy instrument. We have a 
letter from BSFF dated 4 February 2019 that formally submits the report to 
DoF. Therefore, it has reached Stage 3 of the FFPr Policy Indicator
Year/Stage achieved
Baseline=2; Y1=0; Y2=0; Y3=3; Y4=4; Y5=0; 
Finalization activities are in
progress and will continue in
Year-4 for submission to Govt.
for amendment.
04
Develop methodologies for data 
collection and dissemination of 
the official data to facilitate the 
availability of key market 
information to producers and 
exporters.
The assessment and recommended processes constitute guidelines and 
administrative processes that quality this as a policy instrument. A SAFETI￾hired consultant presented his findings and recommendations on this 
subject to DoF at a workshop on April 11, 2019. We have a letter from BSFF 
dated 10 June 2019 that formally submits the recommended 
methodologies and procedures to DoF. Therefore, it has reached Stage 3 of 
the FFPr Policy Indicator. 
Year/Stage achieved
Baseline=2; Y1=0; Y2=0; Y3=3; Y4=4; Y5=5
Based on the
recommendations from the
study, follow-up with DoF on
further formulation of this as
a policy instrument is needed.
Draft Report: Mid-term Evaluation of SAFETI Project 41 | P a g e
Sl. No. Policy works Status Follow-Up
05 Improvement of “Fish Hatchery 
Acts 2010 and Rules 2011” 
The recommended improvements have been reviewed and discussed with 
DoF. This has reached Stage 2.
Year/Stage achieved
Baseline=2; Y1=0; Y2=0; Y3=0; Y4=3&4; Y5=5
Finalization activities are in
progress and will continue in
Year-4 for submission to Govt.
for amendment.
Draft Report: Mid-term Evaluation of SAFETI Project 42 | P a g e
Annex-6: Details Methodology adopted in the mid-term evaluation of the SAFETI project.
Evaluation Approach and Methodology
The SAFETI Mid-term evaluation followed inclusive, participatory and transparent exercises to gain 
insights and alternative perspectives. As much as possible, existing program data sources and analysis 
was used from available documentations, surveys and MEL systems. Data audit and quality checks were 
done. Additional data were collected following quantitative and qualitative methods. The following 
methodologies were adopted in the evaluation processes:
i. Document Review: This included project reports, communication and training material, 
monitoring data on quality and outputs of activities, and other study data (as available). In 
addition, research and other documentation of similar projects were reviewed as necessary.
ii. Quantitative Sample Survey: The evaluation undertook a sample survey to ascertained program 
effectiveness on the key participants i.e. farmers. Primary data were collected from them based 
on SAFETI program result framework as well as Food for Peace program indicators. They survey 
was implemented on both the treatment group (project direct beneficiaries) and control group 
(non-beneficiaries). 
iii. Qualitative Data Collection: We adopted various qualitative methods to collect qualitative 
information from a wide range of project participants and stakeholders. Specifically, Focus 
Group Discussion, In-Depth Interview and Direct Observation methods will be adopted for the 
qualitative information.
iv. Consultative and Dissemination Workshop: MTE arranged a consultation cum dissemination 
workshop with the participation of project management and important stakeholders to generate 
recommendations. Data triangulation was an important element of the workshop. 
Detail Plan on Evaluation and Impact Study
The evaluation planned and implemented in systematic and inclusive manner following fixed approach 
that include both quantitative and qualitative research methodologies, as detailed below. 
Inception and Preparation
CBSG consultants familiarized themselves with the project concept, stakeholders, activities, result 
frameworks and monitoring and control processes. Main objective was to prepare evaluation design 
including detail methodologies, tools, implementation approach and action plans. Inception and 
preparation processes included the following sub-activities. 
Consultation and document review
The evaluation work kicked off through an introductory meeting with the concerned SAFETI project 
team. In addition, we reviewed key project documents and other relevant research and project 
documents. These activities provided the following results:
i. Finalization of evaluation methodologies for both quantitative and qualitative approaches;
ii. Sampling design including the process of respondent identification and selections par research 
approaches;
iii. Finalization of the extent of quantitative sample survey as well as applicable quantitative 
research approach for the Impact Study;
iv. Coordination mechanism and progress reporting;
v. Implementation process and steps of data collection.
Draft Report: Mid-term Evaluation of SAFETI Project 43 | P a g e
Reconnaissance Mission 
MTE undertook a reconnaissance visit at the project sites and met few farmers, visited demonstration 
farms and consulted project stakeholders to have a first-hand understanding on the evaluation context 
and research participants. That helped the team to fine tune evaluation planning, prepare risk 
management, establish logistical arrangements in place, prepare contingency planning and establish a 
coordination mechanism with local partners.
Designing Impact Study Framework – Counterfactual sample survey 
In consultation with the SAFETI project team, MTE designed the counterfactual sample survey including 
sample selection both for the treatment group as well as the control group. The counterfactual is what 
would have happened – what the outcome would have been for a program participant in the absence of 
the program. At a given moment in time, a unit (here beneficiary farmers) either participated in the 
program or did not participate. The unit cannot be observed simultaneously in two different states 
(control and treatment). This is called the counterfactual problem - since we cannot directly observe the 
counterfactual, we must estimate it. Solving the counterfactual problem largely depend on finding a 
“perfect clone” for the treatment. Specifically, the treatment and control groups must be the same in at 
least three ways.
• Firstly, the average characteristics of treatment group and control group must be identical in the 
absence of the program. Although, it is not necessary that individual units in the treatment 
group have “perfect clone” in the control group, on average the characteristics of the treatment 
and control group should be the same. For instance, the average pond size of the farmers should 
remain same in treatment and control group. 
• Secondly, the treatment should not affect the comparison group either directly or indirectly. For 
instance, farmers of treatment group should not transfer resources that they get as part of 
program intervention. 
• Thirdly, outcomes of units in the treatment should change the same way as outcomes in the 
control groups, if both groups were given the program intervention.
• When these three conditions are met, the evidence/existence of program intervention will 
explain any differences in the outcome between the two groups. 
MTE worked with SAFETI on identifying best matching variable that explains better homogeneity 
between treatment and control groups.
We gave maximum attention to make sure that key background features of the treatment group and the 
control group respondents remained similar and comparable i.e. socio economic condition, aquatic 
asset, occupational and geographical context, and so on. In addition, we finalized sample size for the 
treatment group and the control group (we proposed 700 HHs – 500 treatment group and 200 for 
control group) and detail processes to select and identify the ultimate sample unit. 
Empirical Data Collection for Impact Study 
In accordance with the evaluation SOW, we conducted the sample survey with the program 
beneficiaries and the control group to collect additional data on key project outcome indictors. In 
addition, applicable standard Food for Peace program indicators value were collected through the 
survey. Details of the sample survey are provided below. 
Sample Survey Design
The sample survey was designed to gauge the impact of SAFETI project on the direct beneficiaries mainly 
farmers following a counterfactual sample design i.e. sample survey on the treatment group as well as a 
Draft Report: Mid-term Evaluation of SAFETI Project 44 | P a g e
control group in the same project area to make comparisons between what happened as a result of the 
project interventions, and what would have happened in their absence. The comparisons were to offer a 
measurement of the project’s contribution to improvement of technical and economic knowledge, 
production, income and other relevant factors and linkage/access of the direct farmers (may be 
indirect/neighbor) as a result of SAFETI’s activities, and adoption rate.
The following five standard tool box are commonly used in estimating impact:
i. Random Assignment 
ii. Random Promotion
iii. Discontinuity Design
iv. Difference in Differences (Diff-in-diff)
v. Matching (Propensity score matching)
The program intervention type as well as condition of prime target beneficiary (the farmers) showed 
that we could either use Random Assignment (after one/two stage stratification) or Matching 
(propensity score matching - PSM). Application of PSM method has specific data requirements (need to 
run logistic regression). Therefore, CBSG consulted with SAFETI regarding selection of method.
Approach for setting baseline data for control group
We collected midline data for the control group and estimated Propensity Score of each unit (i.e. 
farmer), and we estimated Propensity Score for baseline units under treatment, contingent upon 
receiving the raw dataset could allow the requirement to run the regression. Without the raw dataset, 
we could follow random assignment process.
Approach for handling baseline data
To bring adjustment in baseline data, we used Trimean as per the status of dataset SAFERI had for 
baseline. Statistically Trimean is more appealing.
Sample Size Estimation 
Our understanding about prime intervention units (farmers) dictated to use two stage group sampling. 
Other intervention groups (hatcheries, feed mills) were evaluated qualitatively through KIIs.
Sample size of farmers for treatment area
We assumed
Level of significance, α = .05
Power of the test, (1-β) = 90%
Known population proportion (number of sample farmers), p= 0.7
Effect size variability, δ= 0.05
So, the number of sample size would be= 361
To bring rigor in estimation we will factor a design effect of 1.40 say 500
Sample size for control area
We didn’t have baseline data for control area, therefore as starting point we simply considered 40% of 
the sample size we estimate for treatment which was 200, agreed by SAFETI management
Both “treatment” and “control” group was categorized in ‘Bagda’ monoculture, ‘Bagda’-‘Golda’ rotation 
and ‘Golda’-Carp polyculture production systems by the ratio of 5:6:8, respectively. During the inception 
Draft Report: Mid-term Evaluation of SAFETI Project 45 | P a g e
and preparatory phase, the exact survey extent (both for the treatment and the control group) was 
finalized. The final sample size undertaken is shown below.
Sample Survey Questionnaire
The sample survey questionnaire was prepared based on SAFETI project result framework and outcome 
indicators as well as applicable Food for Peace indicators. 
We reviewed baseline survey tools and included key variables that were useful to measure impact. 
However, after the review of project documents we will come-up with specific variables for data 
collection.
Experiences from the reconnaissance mission brought valuable inputs for the sample survey 
questionnaire. We pre-tested the questionnaire with the potential survey participants at the project site 
where at least 20 interviews were conducted. Pre-testing ascertained relevance and adequacy of the 
survey questionnaire including sequence of the questions, suitability of the language, and 
appropriateness of the skip instructions and the comprehensiveness of issues to address the objectives 
of the study. Afterward, MTE organized a workshop to analyse pre-test results where SAFETI project 
team and the implementing partners attended.
CAPI Development
The sample survey was conducted through Computer Aided Personal Interview (CAPI). We used 
SurveyCTO platform for the survey data collection with concurrence from SAFETI project team. We 
included main features of the electronic survey platform such as survey recording, on and offline, store 
image and identify GPS coordinates. 
Draft Report: Mid-term Evaluation of SAFETI Project 46 | P a g e
Hiring and Training Enumerator Team 
We deployed Field Coordinator and Survey Supervisors from CBSG staff. Enumerators were hired from 
our roster of over 300 survey enumerators. We made sure that the field staff had appropriate sectorial 
expertise preferably in shrimp and prawn sector surveys. Particularly, we hired highly experienced KII 
and FGD facilitators with relevant sectorial expertise. We set special specifications for the survey to 
assess and hire them through in-person interview. We hired 20% more enumerators and provided them 
training.
The underwent Enumerators three-day (31 August – 02 September 2019) training. The training was 
attended by the COP, DCOP and MEL Coordinator of SAFETI. 
Survey Quality Assurance
MEL developed a survey guidelines and protocol for survey management and quality control. 
Supervisors were trained to ensure validity and reliability of survey data. They kept track of all the errors 
and how they had corrected those and submit the record to CBSG Head quarter. Following provides the 
detail measures for survey quality control.
• Use survey protocol to guide and control respondent’s selection, proper replacements, tracking 
and reporting on errors and corrective measures 
• Site level daily review meetings to share problems and seek solutions in coordination with 
central office.
• At least 20% interviews will be spot-checked by supervisors and additionally 20% of them 
validated by back-checks in the field. 
• Centrally, CBSG used Computer Aided Telephonic Interview (CATI) for back checks under the 
supervision of central survey team.
SAFETI Mid-Term Evaluation
The mid-term evaluation reviewed and assessed the project results and impact with reference to the 
SAFETI project as well as Peace for Food program “Result Framework” while the counterfactual 
quantitative survey complemented project evaluation with required empirical data and evidences. 
Specifically, the evaluation will look into the following aspects of the project:
i. Relevance: Is there adequate evidence on the ground that shows the need for such a project 
persists and that the project has targeted the right beneficiaries and undertaken appropriate 
and adequate plan of actions to address the beneficiary needs. 
ii. Efficiency: How the project resources are used to implement the project activates; any gap in 
activity implementation and the reason, quality of project staff and management; M&E systems 
and its utilization, partner agencies capacity and motivation remain at par with the plan, 
contingency planning and action steps.
iii. Effectiveness: Assess activity quality and their effect across the value chains and their short 
term and intermediate effect on the direct participants i.e. productivity, production volume, 
technology adaptation, farmers and other stakeholders, quality and appropriateness of training 
and capacity building program. 
iv. Sustainability: Assess systems and processes that the project has been established like market 
systems and institutional support systems that institutionalize new and improved practices 
towards sustained productivity? Is there any missing opportunity along the project 
implementation? Assess stakeholder’s role and participation in promoting sustainability.
Draft Report: Mid-term Evaluation of SAFETI Project 47 | P a g e
v. Learning: How the project management has installed and practice knowledge management and 
record and incorporate learnings on an ongoing basis; what are learning so far in the project and 
how the learnings can be institutionalize and scale up along the value chains. 
Evaluation Methodology
We adopted mainly qualitative research methods for the project evaluation. Detail data, information, 
opinion and perspectives were collected from all relevant stakeholders and project participants 
including Farmers, Nursery and Hatchery Owners, Feed Manufacturers, Input Dealers, Depot Owners, 
Shrimp and Prawn Processors, Trade Association Leaders, and Government Officials. Specifically, 
following methodologies were used for the evaluation exercise. 
i. Focus Group Discussion: FGD was used to complement quantitative sample survey and mainly 
be used with the direct beneficiaries including farmers of the demonstration firm. FGDs also 
probed key issues including potentials, constraints and challenges of project activities and key
performance indicators. Learnings, insights and sustainability will be the key issue to explore in 
the FDGs. 
ii. In-Depth Interview: In-depth interview was used to interact with a wide range of stakeholders 
to gain their respective achievements, experiences, perspectives, as well as expectations from 
the SAFETI project. There werea variety of respondent groups for the In-Depth interview and 
thus brought in wide range of information, technical insights and most importantly, ideas for 
future course of actions. 
iii. Direct Observations: We undertook direct observation methods at the project site to assess 
adaptation of new technologies, challenges and limitations. However, direct observation was 
complemented by informal and unstructured discussion with project participants and direct 
beneficiaries.
iv. Internal Self-Assessment: We facilitated an Internal Self-Assessment workshop with all 
consortium partners (SAFETI project and partners) to critically review and analyse program 
strategies and activities and drew lessons learnt, challenges and best practices, and importantly, 
draw way forwards for potential future actions. 
Evaluation Study Extent
Evaluation data gathering were done from all related project participants and stakeholders. Additional 
qualitative data gathering were done from the quantitative survey respondents for data corroboration 
and triangulation. Following matrix provide our tentative plan and execution of qualitative evaluation 
data gathering which, however, was finalized during the preparatory period and in consultation with the 
SAFETI project management.
SAFETI endorsed qualitative study extent and respondents 
Methods Quantity Respondent
FGD 9 6 with beneficiary (treatment) farmers of 3 different farming 
methods (Bagda monoculture, Bagda-Golda rotational, Golda-Carp 
polycuture)
3 with non-beneficiary (control) farmers of the same farming 
methods 
KII- Hatchery 5 With hatchery owner (mainly by Ram Raj along with senior local 
team member)
3 Bagda hatchery in Cox’s Bazar and Khulna (2 SPF, 1 
Conventional)- “Aung”
Draft Report: Mid-term Evaluation of SAFETI Project 48 | P a g e
Methods Quantity Respondent
2 Golda in Khulna/Bagerhat, 1 BRAC, 1 private “Sukumar”
KII- Nursery 2 With nursery owner in Satkhira
KII-Feed 
manufacturer
2 With national level feed manufacturer (Agata, Quality/Mega)
1 with Bagda feed importer like either CP or UNI president (Consult 
with Anis, SAFETI)
KII-Aqua input 
trader
3 In 3 districts (Khulna, Bagerhat, Satkhira)
KII-Depot 3 With depot holders (Khulna, Bagerhat, Satkhira)
KII-Processor 2 With fish processing plant managers (Satkhira, Khulna) “Gaffar”
KII-Trade 
association 
leaders
2 1 with President or general secretary of BFFEA (Khulna or Dhaka) 
and 1 with SHAB (Cox’s Bazar)
“Ram Raj”
KII-Govt officials 13 DoF- Each with 10 Upazila Fishery officials, 1 with high officials in 
HQ, and 1 with DD (Khulna)
BFRI- 1 with DG and Senior officials
(mainly by Ram Raj along with senior local team member-Bagerhat 
BFRI)
KII- Partner NGO 2 2 with partner NGO officials (TMSS and CODEC) –PM must 
(supervisor can accompany)
Mobilizing and organizing farmers and financial services
Total: 9 FGDs and 34 KIIs
Features of Qualitative tools and processes:
• Tools was made more focused and performance indicator oriented. 
• We brought some information to triangulate with qualitative data and/or for interpretation of 
quantitative findings.
• We emphasized on SAFETI designed 4 types of support of productivity improvement in terms of 
their importance to the stakeholders. They include: 
• General Training, and training and communication materials
• Technical Assistance to the Farmers at field level
• Management, Monitoring and Quality assurance
• Backward and Forward linage (Financing, Input dealer and Output marketing)
• We included them in the quantitative part with preselected response layout.
• Linkage with Government and Private sector organizations/departments - through KII from 
supply side, through FGD from demand side
• We also included elements of six steps or by all six steps to get their priority importance from 
FGDs and difficulties as well.
Data Triangulation, Analysis and Reporting 
i. Both quantitative and qualitative data will be analyzed and triangulated adequate used various 
sources and points of data gathering and then synthesize for evaluation conclusion. Multiple 
data sources, cohesion and consistency was main checkpoints for triangulation. 
ii. Debriefing session: Evaluation team shared preliminary findings with the SAFETI project 
consortium partners and facilitate partner discussion to formulate recommendations for future
programming and sustainability. 
Draft Report: Mid-term Evaluation of SAFETI Project 49 | P a g e
iii. Feasibility testing: All prospective recommendations were assessed and tested against criteria 
including relevance, do-ability, and economic viability to be put forward for a follow-on project 
consideration.
iv. Narrative Report: An informative and analytical narrative report will be produced first as draft 
report for review and comments from the SAFETI project. Final report will incorporate and 
address issues and comments made on the draft.
Method wise sample distribution
Data collection tool / Shrimp-prawn culture 
system
Sample 
number
Field Survey 
‘Bagda’ monoculture ‘Demo’ farmer 9
‘Project’ farmer 164
‘Control’ farmer 74
Total 247
‘Bagda-Golda’ rotational culture ‘Demo’ farmer 7
‘Project’ farmer 190
‘Control’ farmer 65
Total 262
‘Golda-Carp’ Polyculture ‘Demo’ farmer 7
‘Project’ farmer 171
‘Control’ farmer 64
Total 242
Focus Group Discussion (FGD) 
Treatment Group 6
Control Group 3
Total 9
Key Informants Interview (KII)
KII- Hatchery 4
KII- Nursery 2
KII-Feed manufacturer 3
KII-Aqua input trader 3
KII-Depot 3
KII-Processor 3
KII-Trade association leaders 2
KII-Govt. officials 13
KII- Partner NGO 2
Total 35