"1. INTRODUCTION An estimated 1.8 billion people–almost one quarter of the world’s population–lived in fragile states in 2020. This share is expected to grow to 26 percent in 2030 and 31 percent in 2050 (OECD 2020). Moreover, three-quarters of the extreme poor in the world reside in fragile states (OECD 2020). Yet, despite the enormous importance of these areas for reducing global poverty and food insecurity, there is relatively little research on livelihoods and trade because safe and reliable data collection is often a challenge. Access to food, food prices, and the functioning of food value chains are critical when states fail but, insights into the functioning of agricultural markets and food value chains are rare within such conflict-affected areas. In this paper, we examine agricultural value chains in Myanmar during a period of major political instability. Myanmar’s military seized control of the country in a coup on February 1, 2021, setting the country down a path of widespread violence and major economic collapse. In protest of the coup, Myanmar’s people organized a Civil Disobedience Movement (CDM) and worker strikes, which led to disruptions in service delivery of both public institutions and private businesses. Banks closed to in-person services and faced severe liquidity shortages, limiting businesses’ ability to pay employees and suppliers, and for residents to access their money. Access to formal and informal credit mostly dried up. Internet access and communication was restricted. Violence and insecurity rose sharply after the coup. Cumulatively, these disruptions had major economic consequences as GDP declined by 18 percent and poverty rates increased to between 40 and 50 percent of Myanmar’s population (Boughton et al. 2021, World Bank 2021, UNDP 2021). We study the rice value chain for two reasons. First, rice is the main staple in Myanmar, accounting for 51 and 62 percent of urban and rural calories consumed, respectively, making it crucial for food security in the country. It is also the predominant crop for a large number of farmers and an important export. Second, approximately 70 percent of rice that is consumed is purchased. Value chains are therefore essential in assuring sales of paddy from farmers, milling of paddy to rice, and distribution of rice to consumers. Rice is typically traded over large distances in the country, implying that coup-related disruptions in the functioning of the value chain coup will be widely felt. To study the rice value chain, we use unique primary data collected by phone before and after the coup at the midstream (rice millers) and downstream (retail food vendors) and combine these data with a number of secondary datasets. We look at three research topics. First, we assess the extent of the disruptions to the rice value chain following the military coup by investigating disruptions midstream and downstream, and rice availability in retail markets. Second, we analyze the size of processing and distribution margins within the value chain and assess how they have been affected by the coup. Third, we look at spatial price dispersion from the mills to food vendors comparing rice prices before and after the coup using a market-pair regression methodology of miller-vendor pairs with miller and vendor fixed effects. We test what explanatory variables of price dispersion used in international literature matter in this context and to what extent these factors have been affected by the instability. We examine travel costs (Minten and Kyle 1999, Minten et al. 2016), border crossings (as a proxy for export market access) (Aker et al. 2014), and differences of ethnic group composition between markets (Aker et al. 2014, Robinson 2016). We also assess the direct effects of measures of violent events in sending and receiving markets on price dispersion. Our data reveal significant business disturbances for food vendors and rice millers after the coup linked to banking and transport. Yet, despite these challenges, local trade and milling continued, though at lower activity levels than normal, assuring availability of rice in most retail markets in the country. Upstream, farm prices and milling margins were found to be mostly stable after the coup. However, we see increases in retail rice prices by 11 percent on average. Simulation results of these increased distribution margins indicate sizable welfare costs for rice consumers and producers, estimated at almost 0.5 billion USD at the national level (equivalent to about 3 percent of agricultural GDP in 2020). Regressions reveal that spatial price dispersion increased after the coup with the distance between mills and vendors widening the price differences. Violence near mills or vendors, which increased dramatically during the coup, also increased price dispersion. However, in contrast with previous findings (Aker et al. 2014, Robinson 2016), we find that differences in ethnic composition of sending and receiving townships/markets did not affect dispersion. Price dispersion further decreased with proximity of food vendors to land borders, possibly due to the price reducing effects of competition from exports in those markets. However, the border effect was weakened severely during periods of border closures. The results indicate that the most severe effect of the coup on retail prices were found in the areas furthest away from export opportunities and major production zones that were affected by violence. When the state fails to assume its normal role and private sector activities are curtailed, it is crucial that agricultural value chains continue to function well to ensure food security and to prevent increases in food retail prices to avoid further suffering of vulnerable households. Our findings indicate that easing transport restrictions, stabilizing fuel prices, and facilitating safe spatial arbitrage of food products would reduce the welfare costs of market disruptions by preventing food price inflation while assuring remunerative prices for farmers. To the extent possible, maintaining safe border trade of agricultural commodities may help stabilize prices in the domestic market. Finally, as mobile phone use is often widespread, even in a failed state, close monitoring of a number of important aspects in such environments can take place and should therefore be encouraged. The paper proceeds as follows. In the next section, we provide important background information on the political conflict and the rice value chain in Myanmar. The econometric specification used to empirically test price changes during the crisis is discussed in Section 3 along with our data and survey methods. Section 4 illustrates the disruptions noted in the value chain. Section 5 presents graphical analyses of price and margin changes, revealing retail price increases following the coup, as well as price regressions results. Section 6 discusses the implications of our results and section 7 concludes by summarizing the key results and by discussing main policy implications..."