From Ponds to Plates: The Role of Crop-Aquaculture Farming Systems in Building Household Food Security in Myanmar. Pacem Kotchofa, Marie-Charlotte Buisson, Shelly Win. International Water Management Institute (IWMI) Southeast Asia. Marrakech, Morocco, December 4th, 2025. Under the High Patronage of His Majesty King Mohammed VI 1 Myanmar: Demographics & Food Systems Population: + 50 million inhabitants (50% to 70%) relying on agriculture for their livelihoods (World Bank, 2024). Agriculture: Highly vulnerable to climate shocks and water scarcity. Monoculture systems are low-resilient, less productive, and water-intensive, jeopardizing the livelihoods of millions of smallholder farmers, leading to widespread food insecurity and hunger in both rural and urban areas ​(IFPRI Myanmar, 2024)​. XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Monocropping or the overreliance on a single crop… 2 Integrated Food Systems: Potential pathway Crop and Aquaculture Farming (CAF): A synergetic environment where crops and aquaculture can thrive. Increase agricultural productivity (yield) & Efficient use of resources (water and land) while bridging seasonal food gaps during dry seasons when crop harvests are critically low ​(Wang et al., 2023; Myo et al., 2024)​. Increase incomes (additional products, i.e., fish, seaweed, shrimps, or crustaceans) while serving as a vital source of protein and micronutrients for HH food consumption ​(Mekonnen et al., 2019; Dubey et al., 2024). Increase resilience against climate variability, ultimately strengthening the livelihoods of farming HH, especially in rural areas where poverty rates are often the highest ​(Wang et al., 2024). XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Integrated food system: An approach to combining multiple food production sustainably. Beyond increasing crop yields 3 CAF: Policy Gap & Research Questions Policymakers lack rigorous, causal evidence on which sustainable practices truly work. RQ1: Could CAF adoption contribute to lowering food security? RQ2: What are the primary factors that influence households' decision to adopt a CAF system? RQ3: To what extent does the CAF adoption, causally, improve the livelihoods and food security of farming households in Myanmar? XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Data & Methods Two data sources: XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org The Aquaculture Development Support Tool (Aqua-DST)​: Tool with information on climate hazards, aquaculture systems, biophysical conditions, and socioeconomic factors (Win S. et al., 2024)​. Upper Ayeyarwady River Basin, and limited to regions such as Kachin, Magway, Mandalay, Sagaing, and Southern and Eastern Shan states Myanmar Household Welfare Survey (MHWS) Round 6 ​: A nationally and regionally representative dataset with information on household characteristics, welfare, food security, and farming practices, including CAF, and shocks (IFPRI Myanmar, 2024). Sample size of 12,898 households. August and November 2023. Data & Methods Key variables of interest: Treatment variable CAF Adoption dummy variable (1, 0): practice both crop and aquaculture. Outcome variable FI dummy variable (1, 0) derived from the Household Hunger Scale/HFIAS tools. XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Food Secure (0) Mild (1-2) Moderate (3-4) Report no food access disruption. Worry about food. Go to bed hungry. Spend an entire day and night without food. Propensity Scores Matching (PSM): Selection bias in CAF adoption impact assessments, where participation depends on biophysical and socioeconomic factors specific to each household rather than on random assignment. Create a control group, statistically similar to the CAF adopters, using a set of observable covariates, such as HH wealth index (PCA), farm size, access to irrigation, education level, gender of the household head, climate shocks (drought, flood), and biophysical and value chain data from the Aqua-DST. Synthetic group. 6 Data & Methods XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Propensity Scores Matching (PSM): Propensity score : Probability of adoption, given a set of pre-treatments observable covariates using a logit regression model. Nearest Neighbor Matching (NNM) Average treatment effect (ATT): difference in the outcome, the likelihood of FI of the matched groups NN to match each CAF (treatment group) with the control group that has the most similar propensity score. 7 Key Findings Table 1: Summary Statistics: Treated Group vs. Untreated Group XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org   CAF Households (N = 211/ 12,898) Non-CAF HH (5,502) Household head gender (male) 63.03% 55.13% Marital Status (Married) 76.78 % 70.85% Age (mean) 42 (12.59) 40 (13.07) Education (Below average) 35.07% 35.21%   Location: (rural) 90.52%  91.94%  Sagaing 16.11%  13.67%  Ayeyarwady 12.80%  12.23%  Mon/Mandalay 12.80% 11.65%  Drought (Yes) Flooding (Yes) H. Temperature (Yes) 2.84% 14.69% 1.4% 3.24% 11.60%  1.56%  Food Insecurity (FI = 1) 18.96% 24.72%  Most of the CAF HH are male headed based in rural areas across Sagaing, …Mon The major climate shock reported is flooding 8 RQ1: Key Findings Table 2: Association between household FI status and CAF adoption: XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Variables CAF Households All Sample   Coeff. Std. Error Coeff. Std. Error Demographics and socioeconomics Gender (Female) 0.550 (0.445) 0.207*** (0.040) Age -0.006 (0.018) 0.005*** (0.002) Urban Areas 1.13* (0.64)     High Education -0.27 0.484) -0.35*** (0.044) Agricultural and climatic conditions Land tenure: No - - 0.731*** (0.046) CAF (Yes) - - -0.38** (0.182) Drought 0.078 (1.22) 0.409*** (0.149) Flood 1.235** (0.546) 0.679*** (0.063) High Temp. 2.55 (1.67) 0.608*** (0.164) LR Chi2 30.79** 712.31*** Pseudo R2 0.162 0.045 Female HH heads are associated with an increased odds of FI compared to their male counterparts. 9 RQ1: Key Findings Graph 1: Association between household FI status and CAF adoption: XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Test Statistics: Ho: No difference in Mean (FI) Scores between CAF and non-CAF Households. Ho: Mean (FI) CAF > Mean FI nonCAF Households. RHo The unconditional logit model shows a strong negative association between CAF adoption and FI Conclude that the CAF system significantly improves farmers’ resilience, i.e., lowers their food insecurity scores compared to those in non-CAF 10 RQ2: Key Findings Table 3: PSM household FI status and CAF adoption: XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org  Variables Coeff. Std. Error Gender (Female) 0.046 (0.255) Age 0.006 (0.01) Urban Areas -0.662 (0.492) High Education -0.085 (0.265) Farm Size (ha) 0.02*** (0.009) Drought (Yes) -0.423 (0.795) Flood -0.546 (0.435) PC 1 (Wealth Index) 0.11** (0.05) High Temp. -0.806*** (0.729) Rainfall – Increase (mm) -0.51** (-0.233) Rainfall – Decrease (mm) -0.18 (0.147) Flooding Area (km2) -0.01 (0.006) Wetlands (ha) 0   Irrigated Areas (ha) 0   RQ2: Key Findings Table 3: PSM household FI status and CAF adoption: XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org  Variables Coeff. Std. Error Number of fish markets 0.035 (0.043) Number of hatcheries 1.27*** (0.443) Number of fish nurseries 0.016 (0.141) N. of fingerling distributors -2.7*** (0.471) Number of fish feed suppliers -0.148 (0.125) F. consumption kg/per capita 0.06*** (0.019) N. of fishery officers (DoF) 0.16** (0.065) Number of fish farmers 0.01*** (0.002) Accessibility to roads (miles) 0.01*** (0.002) LR chi2 106.92   Prob > chi2 0   Pseudo R2 0.059   N. Observation 2,598   RQ3: Key Findings PSM household FI status and CAF adoption: XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Variables Sample Treated Controls Difference S.E. Tstat FI ATT 0.328 0.249 -0.071* 0.04 -1.79 Variables Sample Treated Controls Difference S.E. Tstat FI Unmatched 0.2247 0.2351 - 0.010 0.046 -0.23   ATT 0.221 0.226 - 0.007 0.054 -0.12 Graph 2: PS distribution by HH status Table 4: Regression summaries of the treatment effects. Rosenbaum bounds sensitivity test: Robustness of results. ATT showed a 7.1 percentage-point reduction in food insecurity (FI) scores among households practicing CAF compared to similar households that do not practice CAF. 13 Implications of the study Female household heads are still at a higher risk of food insecurity than their male counterparts. Better knowledge (extension services) to handle adverse weather can improve farmers’ ability to prevent or reduce food insecurity in Myanmar. Households engaged in integrated CAF systems experienced, on average, a statistically significant reduction in food insecurity compared to similar households that did not participate. The adoption of CAF systems may be more influenced by the value chains and biophysical factors associated with aquaculture expansion than by households’ demographic or financial means. Policy support for CAF uptake should allocate resources to developing gender sensitive aquaculture expansion infrastructures. XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org From Evidence to Action. 14 Conclusion Integrated food systems, e.g., Crop-Aquaculture Farming (CAF), can be a pathway to increase household resilience against food insecurity and climate variability. Major barriers to CAF expansion can be addressed through public and private investments in aquaculture value chains and in extension services for managing adverse climate shocks. Next Steps (From region to agroecological zone): Disaggregate the results and policy formulation per agroecological zones: Wet, Dry, and uplands/hilly as water and climate challenges are different. CAF can improve livelihoods and support the achievement of SDGs related to hunger, water, land, and poverty reduction, especially among female farmers.   XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Acknowledgement The authors express their gratitude to the IWMI Economics and Impact Assessment (ECIA) Research Group for their valuable insight and inputs. “The Myanmar Household Welfare Survey and Myanmar Agriculture Production Survey was undertaken as part of the Feed the Future Myanmar Agricultural Policy Support Activity (MAPSA) led by the International Food Policy Research Institute (IFPRI) in partnership with Michigan State University (MSU) and made possible by the support of the American people through the United States Agency of International Development (USAID). The opinions expressed here belong to the author(s) and do not necessarily reflect those of IFPRI, MSU, CGIAR, or USAID.”   XIX World Water Congress | Marrakech, Morocco | 1-5 December 2025 | www.worldwatercongress.org Thank you! Contact me: p.kotchofa@cgiar.org Under the High Patronage of His Majesty King Mohammed VI 17 image1.jpg image2.png image3.png image4.png image5.png image6.png image7.png image15.jpg image8.png image9.png image10.png image11.png image12.png image13.png image14.png image16.png image17.png image18.png image19.png image20.png image21.png image22.png image23.png image24.png image21.emf 0 .2 .4 .6 .8 1 Pr (F I) 0 .2 .4 .6 .8 1 ICAF Households Pr(FI) Pr(FI) image22.emf 0 .2 .4 .6 .8 1 Pr (F I) 0 .2 .4 .6 .8 1 Non-ICAF Agricultural Household Pr(FI) Pr(FI) image27.png image28.png image29.png image23.emf 0 1 0 2 0 3 0 4 0 D e n s i t y 0.2.4.6.81 Propensity Score Treated CAFControl non-CAF Propensity Score Distribution by Treatment Status image24.emf image31.png image25.png image26.png image30.png