1 Poverty-driven vulnerability and household behaviour in uncertainties: Evidence from Perlis paddy smallholders’ farmers By Jamalludin Sulaiman, Ph.D Azlinda Azman, Ph.D Saidatulakmal Mohd, Ph.D Senadjki Abdelhak School of Social Sciences Universiti Sains Malaysia 11800 Penang ABSTRACT Many have attempted to measure vulnerability, a multidimensional phenomenon, using different methodologies. While poverty-driven vulnerability is dependent on the availability of tangible resources and assets, many previous studies found that severe risks and shocks, coping capacities and abilities to mange risks by households also affects ones vulnerability to poverty. However, there have been no evidences available todate on how to establish the degree of vulnerabilities, on how households minimize the negative effects of risks and shock, and lastly on the factors that determine their behaviors? This study uses a three-round panel data collected using a structured socio-economic questionnaire to assess risks and to analyze households’ strategies and vulnerabilities to poverty of paddy smallholders in Perlis Malaysia. Data were collected at 6 month interval to provide the best information on risk and vulnerabilities possible covering all possible variations or seasons in the country. Using the three datasets of paddy farmers’ socio-economic information, this study will answer the following questions: (1) how will farmers’ responses be when they experience of uncertainties, (2) how will the strategies indicate the level of their vulnerabilities, and (3) to determine if and how the farmers psychological state have any effect on their vulnerabilities? Keywords: Vulnerability to Poverty, Behaviour, Psychological situation, farmers, longitudinal data JEL codes: I32, C23 1. Introduction Vulnerability is a multidimensional phenomenon. Many have attempted to construct a common definition and measurement of vulnerability, but none have succeeded in achieving the targeted objective. There are different methodologies that have been developed for the purpose of measuring vulnerability (Chaudhuri et al., 2002; Chaudhuri, 2003; Glewwe & Hall, 1998; Christiansen & Boisvert, 2000; Chaudhuri et al., 2002; Ligon & Schechter, 2003; Calvo & dercon, 2005; Gunther & Harttgen, 2009; Ersado, 2006; Pritchett et al., 2000; Dutta et al., 2010). These methods remains ineffective as none could effectively measure the