Growth, crisis and resilience: RN3.doc
Growth, crisis and resilience: household responses to economic change in rural Southeast Asia
Evidence from Northern Thailand
Research Note No. 3: The Mae Chaem household survey
1
Growth, crisis and resilience: RN3.doc
Background
This note describes the conduct of a set of household surveys conducted during 1998 and 1999 in Mae Chaem district, Northern Thailand. The study for which the surveys were conducted addresses two sets of issues relating to welfare among rural Thai households. One is the use of natural resources, and especially agricultural land, to produce income for consumption and investment. What kinds of resources are available and to whom, and how they are used, are all important subjects. The other set of issues concerns the stability of household income, consumption and investment in the face of shocks such as drought and unexpected changes in economic conditions. Households or communities that are vulnerable to shocks may experience short-term hardship, and addressing this raises questions about availability, efficiency and equitable access to formal and informal social safety nets. In some circumstances, moreover, a shock may have long-term effects, for example when the response to a shock includes disinvestments such as the removal of children from school or the "mining" of forest or soil nutrient resources. When a shock has long-term effects then future generations inherit a different set of resource endowments with which to generate their own consumption and investment opportunities. The two sets of issues, growth and resilience to shocks, are thus linked.
In our primary data collection we collected baseline and historical information on the agricultural economy of Mae Chaem in addition to gathering recall data on a wide range of economic variables for the crop years 2341 and 2342 (1998 and 1999). We focused on the combined effects of the 1998 drought and concurrent economic crisis on households in one district. Using survey and secondary data as well as information gleaned from focus group meetings and other informal sources, we defined measures of household wealth and its growth, relating these to the households' initial resource base and to their investments in land, education, businesses, migration and other non-farm activities. We then examined the ways in which each household was affected by the shocks, using both objective measures such as crop yields and remittance flows, and subjective measures as reported by households themselves. We use the information gathered to formulate and test hypotheses, and ultimately to draw conclusions, about the vulnerability of households to unexpected shocks, and the likely severity of their effects.
The survey
Site selection. We chose Mae Chaem, west of Chiang Mai, as our research site because it occupies an important watershed and combines settled agriculture with various forms of expansion at and intro the forest margins. Within the district we focused on upland field crop areas whose inhabitants are mainly northern Thai (Khon muang), as well as some Karen. Agriculture in the site is a mixture of subsistence crops (mainly rice) and commercial crops. We chose twelve villages in the district; all have mixed lowland and upland cropping systems, but vary by types of market access and distance to a market center.
A focus on non-irrigated production systems allows us to analyze the effects of structural changes in the economy on land use and agricultural intensification. To avoid excessive heterogeneity in the sample we excluded highland areas. Current and recent research in the district (and indeed in northern Thailand focuses on hill tribe (upland/highland) communities and in a relative sense, little is known about upland agriculture in spite of past research findings that the environmental impacts of deforestation are quite severe in such areas.
The survey was led by a team of researchers from Chiang Mai University (CMU) and the University of Wisconsin. December and January 1999 were spent in Chiang Mai obtaining relevant secondary data and recruiting research assistants at CMU and in Mae Chaem. The original sample selection was also checked and discussed with informants in the district.
Village focus group interviews From mid-January to March 1999, focus group discussions were held in each village to explain the study and obtain basic village data. Information emerging from those discussions informed the design of the first survey instrument. The instrument was then pretested in two villages outside our sample, one Thai and one Karen, and revised.
Household survey #1
The survey instrument combined tables to collect quantitative production and demographic data with open-ended questions about the crisis and its effect on each household. We hired five graduate students (three women and two men) from CMU to conduct interviews. Three of the enumerators are native speakers of northern Thai. The two male students were involved in the focus group interviews and guided continued interaction with each village. They were also involved in pre-testing the survey instrument and helped train the other enumerators in its use (an English translation of the survey instrument can be found on the project web site at
170 households were chosen for interview in a stratified random sample, using equiproportional sample sizes according to village population data (for details see Table 1). The research team conducted the survey itself in one month during the dry season, when farm labor requirements are at their annual low point. Four Karen secondary school students were hired to assist in translation in Karen-speaking households. The students were hired from non-sampled villages so that they assisted with translation outside their own villages. Many of the Karen interviewed., as well as most respondents in the one Hmong village, could speak northern Thai and so less translation was needed than expected.
The five Thai enumerators led all the survey interviews. After the completion of the survey the enumerators coded, checked and keyed the data. All data were keyed twice and the resulting files compared for accuracy.
Network Interviews
After initial analysis of the survey #1 data during the summer of 1999, we returned to Chiang Mai in late October to conduct more in-depth interviews to explore social networks of a smaller sample of households. We used a stratified sampling technique based on per capita income. The goal was to interview households that differed by wealth, livelihood and specific responses (coping mechanisms) to the crisis. These interviews, with a sample of 24 households, gathered qualitative information about the economic and social ties of each household and how these contributed to households' means to deal with the crisis. We asked questions about specific social relationships and institutions that provide assistance or access to productive resources for households. The specific areas of inquiry included consumption based assistance (borrowing rice, money or obtaining shop credit for consumer goods), credit sources (local or district based groups), production based ties (input supply and marketing), and employment ties (both local and external migration ties). These household based networks have been charted and illustrate specific patterns of network structure across villages and households.
Household survey #2
Following the network survey, in December 1999 we redesigned and repeated the large-sample survey, adding a number of specific questions about crisis effects based on the results of the earlier surveys. The second survey added a second year of data on agricultural production, land use, household incomes, migration and labor market participation and other key variables. The agricultural data from 1999, a climatically normal year, provide a comparison with the El Niño year, 1998. Interviews were conducted by members of the original survey team, providing continuity. Once again, data were coded, checked and keyed in Chiang Mai.
Ongoing data analysis
We now have two years of complete data on household incomes, agricultural practices and related variables in the 170-household sample. The addition of a second full household survey was not anticipated during the design of this project. Give the extraordinary economic and meteorological circumstances that prevailed in 1998, however, this step was both necessary and also highly desirable. As a result, including recall data on land use, crop yields and household income sources from years prior to 1998, we now have data for the household sample that span the crisis/drought year. The surveys thus afford us the opportunity to observe how households were affected by both drought and the 1998 economic shock, and to analyze their responses. The data should thus provide a rich source for understanding how rural households deal with shocks, and for developing normative statements about the ways in which the institutional and policy setting modulated impacts and responses.
In early 2000 the data are being used in Chiang Mai by two Masters' students, as well as by the Madison-based PI and graduate student. A workshop reporting the results of these analyses is planned for late 2000. It will feature presentations by the primary researchers as well as responses and discussion with a panel of experts from Thai academic and research institutions as well as representatives of Thai government agencies with policy mandates in the areas of agriculture, rural development, and social policy.
1
Growth, crisis and resilience: RN3.doc
Table 1 Mae Chaem survey: village sampling frame
Village / Tambon & Hamlet / Total H'holds / Sample Size / Percentage of Households / Ethnicity(majority)
Bon Na / Chang Koeng 14 / 108 / 17 / 16% / N. Thai
Mae Thaen / Chang Koeng 1 / 25 / 4 / 16% / N. Thai
Long Pong / Kong Kaek 4 / 84 / 14 / 17% / Hmong
Hua Doi / Kong Kaek 3 / 96 / 15 / 16% / N. Thai
Om Maeng / Kong Kaek 8 / 146 / 24 / 16% / N. Thai
Na Yang Din / Kong Kaek 8 / 54 / 9 / 17% / Karen
Mae Najon / Mae Najon 5 / 168 / 27 / 16% / N. Thai
Mae Najon Nua / Mae Najon 16 / 150 / 24 / 16% / N. Thai
Mae Mu / Mae Najon 4 / 63 / 10 / 16% / Karen
Mae Raek / Tha Pha 1 / 50 / 8 / 16% / Karen
Sam Sop / Tha Pha 1 / 30 / 5 / 17% / N. Thai
Yang Sarn / Tha Pha 8 / 79 / 13 / 16% / Karen
Total / 1053 / 170 / 16%
Authors: Jean Geran and Ian Coxhead, University of Wisconsin-Madison. This research was funded by a grant from the Ford Foundation and benefited from resources provided by the Thailand Development Research Institute Foundation. For more information visit or send email to .
1