Location and Place of Research Introduction

3.2 Location and Place of Research

After initial secondary data collection, field research was undertaken during two time periods as indicated in Table 3.1. Table 3.1: Location and time of proposed research. For objective 2, the selection of coastal communities were based on the results from objective 1. Objective Data collection time Data collection location 1 – Provincial review Secondary data: April – June 2011 Offices of regional statistics office BPS, BAPPEDA, DKP Primary data: June 2011: Interviews with staff September -October 2012: Field interviews with key informants Provincial and Kabupaten offices of DKP and BAPPEDA 25 selected coastal Kelurahans based on results of fishing dependency analysis. 2 – Sustainable livelihoods Analysis Primary data: September-December 2013: Field interviews and focus groups Sungai Pinang and Carocok, Tarusan 3 – Evaluation of projects January – March 2013: Field interviews and focus groups Selected coastal communities in Agam, Padang and Pesisir Selatan.

3.3 Data Collection

For clarity this section has been separated into the three research objectives and each section begins by clarifying the objective.

3.3.1 Objective 1: Provincial-wide review of fisheries dependence and poverty.

To identify the most highly fisheries dependent communities and those that contain the highest concentration of poor fisher households in mainland West Sumatra using two indices generated from routinely collected census and fisheries data. Secondly, to examine the relationships between poverty in fisheries and poverty in other economic sectors using a correlation matrix. Thirdly, interviews with staff from government agencies and field visits to 16 fishing communities checked the validity of the secondary data and began to explore which sectors of the fishing industry the poor work in and what are their needs and constraints. The geographical extent of this review is restricted to mainland West Sumatra and does not include the Mentawai Islands because 1 the geographical isolation of the Mentawai Islands would require much greater time and resources to survey adequately and 2 the people groups and culture of the Mentawai Islands are different to mainland West Sumatra.

3.3.2 Objective 2: Livelihoods and poverty analysis

To analyse the household livelihood context of selected representative coastal communities in West Sumatra and to investigate the shape and form of poverty in those fishing communities. The methodology proceeded logically though the following steps which are described in more detail in chapters 5 and 6: Identify enabling and constraining factors. One of the most significant aspects of this research is the attempt to quantify the asset profiles of fishers. In order to do that, interviews with stakeholders in 25 fishing communities were conducted to identify which sector of the fishing industry the poor operated in and which enabling or constraining factors were crucial for their livelihoods. Create measures for each of these. 31 enabling and constraining factors were identified and 43 measures attributes were created to score these on a range of bad to good. Choose representatives sites based on a clustering analysis. Two sites were selected, a rural, isolated village where many of the population were fisherfarmers Sungai Pinang and a more industrial site where larger vessels were based Carocok, Tarusan. Conduct interviews to score these measures. Using a structured survey, 151 households were asked questions to elicit scores for each of the 43 attributes. At points of interest or where further clarity was required respondents were asked follow-up questions in addition to the survey. The literature on the sustainable livelihoods approach offered some helpful advice regarding data collection and analysis which is relevant to the Author’s proposed research. DFID 2000 refer to the common problem of researchers overestimating the data required to understand different aspects of livelihoods and underestimating the time taken to process these data and obtain results. In reference to Rapid Rural Appraisal RRA, McCracken et al. 1988 refer to understanding being found through the rapid build-up of diverse information rather than via statistical replication alone. This principle of triangulation, to confirm and complete data by using multiple sources and methods Arksey and Knight, 1999, is crucial to build up an accurate picture of the “truth” in West Sumatran coastal communities. However, the multi-dimensional analysis described below added rigour and reproducibility to the qualitative aspects of this research.

3.3.3 Objective 3: Evaluation of existing livelihoods interventions.

To evaluate the suitability of previous livelihood projects in West Sumatra with the needs and constraints of the poor. There were two elements to the data collection of this third objective: Collating past and present project data. Livelihood improvement interventions, conducted by the provincial DKP between 2005-2009, were identified from government reports and by interviewing staff members. Government interventions were included if they were intended to directly improve livelihoods. Interventions from the first year 2012 of the current multi-agency program of poverty alleviation for coastal communities, G-PEMP, were also collated. Case studies. While the past and present project data approach above was designed to quantify what had been done to help fishers, the case studies were aiming to explore the mechanisms that facilitate successful or unsuccessful implementation of these interventions. Because current government policy is to give funding and support through groups rather than individuals, three fisher groups were selected and group leaders, members of the group and local DKP officials were interviewed to ascertain the factors contributing to success and failure of these groups.

3.4 Data Analysis

3.4.1 Objective 1: Provincial Review

A fisheries dependence index was generated using the data, total numbers of fishers, the percentage of the workforce employed as fishers and the total production of fish and shellfish. The data were normalised, weighted evenly and summed to generate a composite statistic. A poverty index was created based on frequency and percentage of poverty from BPS data in 2011. These two components were normalised and combined to form a poverty index. Fishing dependency and poverty amongst fishers were compared to dependency and poverty in all other economic sectors using a Kendall-Tau correlation matrix.

3.4.2 Objective 2: Sustainable Livelihoods Approach

Identify enabling and constraining factors. Factors were identified by analysing interviews and distilling recurring themes that were repeatedly raised by different respondents in different locations. These themes were refined to identify the generic factor that needed to be scored. Choose representatives sites based on a clustering analysis. Four criteria were used to segregate locations using a cluster analysis, geographical isolation, fishing fleet, presence of an auction and natural geography. Data were analysed using a cluster analysis from the statistics package Multibase in conjunction with Excel. Clustering analyses were performed using 6 different measures namely; furthest neighbour, nearest neighbour, centroid, ward, flexible and group average methods. Multi-Dimensional Scaling MDS . With 43 attributes from 6 fields being scored for 151 households, the 6,493 data points plus additional commentary by the respondents were entered into a spreadsheet and analyzed using a multi- dimensional scaling technique based on Rapfish Pitcher and Preikshot, 2000; Pitcher et al., 2013 and using the R programming framework R Development Core Team; www.r-project.org. Qualitative data were logged in the spreadsheet, highlighted and used to explain observations from the MDS plots.

3.4.3 Objective 3: Project Evaluation

Analysing past and present project data. Descriptive statistics were used to analyse the frequency and expenditure of each intervention. An asset analysis methodology was developed based on the principles outlined in Ashley and Hussein 2000. Each intervention was scored on whether it intended to improve human, social, financial, natural or physical capital. What is being measured is the presence or absence of an intended benefit in an asset category and not whether that benefit was realized. Intended benefits were compared with the fishing poverty sectors identified in chapter 5 in order to assess the fit of livelihood interventions Ashley and Hussein, 2000. Interview analysis of the case studies. Interviews were recorded, transcribed and translated. Recurring concepts or phrases were highlighted and these were triangulated with other stakeholders in the same location and compared and contrasted with other case studies. 4 FISHING DEPENDENCY AND POVERTY IN WEST SUMATRA

4.1 Introduction

In this chapter 4 a simple methodology is described that identifies both fishing dependency and incidences of poverty amongst fishing households that will address the following questions: Which coastal areas of West Sumatra are most dependent on fishing? Where are poor fishing households concentrated in West Sumatra? Are fishing dependency and poverty related? How is poverty in the fisheries sector related to poverty in other economic sectors? After decades of haphazard and unregulated activities in the marine environment many nations are developing marine spatial planning MSP strategies. Indonesia has joined them and MSP was given a statutory basis through the National Law No. 27 Year 2007 on Coastal and Small Islands Management. This law specifically aimed to “strengthen community involvement”, “increase social, economic and culture value” and “empower communities to improve welfare” chapter 2, article 4 and Chapter 7, article 63 in Law 27 2007 UU No. 27, 2007. Despite socio-economic components forming a significant part of the scope of MSP, existing marine spatial plans have a strong bias to physical and biological data and fail to adequately capture the human dimension. The irony of this is that while collecting comprehensive biological and physical data has meant expensive field surveys, the human and social data already exist and merely need 4 As part of this post graduate research project, the material presented in this chapter has previously been published Stanford et al., 2013 to be integrated into the MSP process. The analysis described below is the implementation of this. Before fishing dependency is analysed a clear definition is needed. Meaningful measures of fisheries dependent regions need to capture the sense that “the industry provides an essential backbone to its economic or social structure” Phillipson, 2000. Some authors propose that 5 Lindkvist, 2000 or 10 Symes, 2000 of the working population should be employed in fisheries to constitute fisheries dependence. Other authors emphasize the socio-cultural com ponent of fisheries as “a way of life” that characterize the community and that contribute much more than a source of revenue alone Van Ginkel, 2001; Jacob et al., 2001; Ross, 2013. Having weighed up the different approaches Brookfield et al. 2005 define a fisheries-dependent community as “a population in a specific territorial location which relies upon the fishing industry for its continued economic, social and cultural su ccess”. The value in this looser definition is that it 1 explicitly includes cultural aspects, 2 highlights the reliance of fisheries for ‘success’, therefore hinting that a community may survive without fishing and 3 is not tied to a specific percentage of employment in the industry. Certainly, meaningful measures of fishing dependence need to couple fisheries statistics with social indicators Phillipson, 2000. Fisheries dependency analyses in developed countries typically link fisheries to the regional or national economy through some form of input-output analysis Seung and Waters, 2006; Kwak et al., 2005. In developing countries these analyses are useful at macro-scales, and have been used to demonstrate the importance of fisheries exports to the Indonesian economy Yusuf and Tajerin, 2007 and the importance of fisheries to a given region Dault et al., 2009. However the intention of this paper is to analyse fisheries dependency at multiple scales were the data requirements of input-output model are prohibitive. Even in the comparatively data-rich, highly managed fisheries of the US, the authors of one review of input- output models concluded that “published data for these variables are not sufficiently detailed to be used for regional economic analysis of fisheries ” and “to support accurate regional economic analysis of fisheries, it is critical to have a comprehensive data collection program ” Seung and Waters, 2006. Like many developing nations, data availability and validity is one of the key stumbling blocks for conducting research in a context such as Indonesia. The multi-species nature of tropical coasts combined with widely dispersed landing and sales sites and the reality that many of these transactions occur through the informal non-tax paying sector, mean that the human component of fisheries management data can be already inaccurate even before the inherent uncertainties of stock assessment are considered. Much of the fisheries statistics collected by fisheries staff through interviews also face the underlying incentive of respondents to over-exaggerate the catch, boats or numbers of fishers in order that they may have greater access to financial support. In the face of these challenges, multi- sector census data helps to reduce the incentive to over-exaggerate see methods below. This study describes a fisheries dependency index, based on routinely collected fisheries statistics, and a poverty index, based on the Indonesian Governments definition of poor households, to explore fisheries dependency and deprivation in coastal communities of mainland West Sumatra. Fisheries dependence is analysed at three different spatial levels to ensure that fisheries dependent communities are neither overlooked nor artificially highlighted Phillipson, 2000. Next this study sets poverty amongst fishers in the context of cross-sectoral poverty in order to explore broader trends in poverty and specifically to probe the following question with empirical data; to what extent does fisheries dependence go hand in hand with poverty or, in the words of Béné 2003, does fishery rhyme with poverty?

4.2 Method