A new way to look at the social economy of dried fish
Value chains are networks of actors whose activities enable the production and distribution of goods or services to consumers. All value chain actors use assets to transform inputs into goods or services (outputs). These become inputs when used by other actors further ‘downstream’ in the value chain. For example, fishers use boats and nets (assets) and labor, fuel, ice and credit (inputs), to ‘make’ an output (raw fish). Raw fish is then used as an input by fish processors.
Value chains can be conceived of as being made up of three segments – the upstream, the midstream, and the downstream. In the dried fish value chain, the part of the value chain specialized in producing raw materials (i.e. fishing) is the ‘upstream’ segment. The part of the chain that specializes in transforming fresh fish into dried fish and aggregating dried fish (i.e. processing and trading) is the ‘midstream’ segment. The part of the chain that specializes in distributing dried fish to consumers (i.e. retail) is the ‘downstream’ segment.
How DFM’s ‘stacked’ approach is different to conventional value chain analysis
The approach to value chain research adopted by DFM is adapted from the “stacked value chain analysis” (SVCA) approach developed by Reardon et al. (2012). This differs from most conventional value chain analysis in several ways.
First, most value chain analysis focusses on structure, but is concerned simply with ‘mapping’ the types of actor in each segment of the chain, and estimating the share of the retail value of the final product captured by the main actors. Generally, less attention is paid to other aspects of value chain structure such as the number and geographical location of actors, or the degree to which production is concentrated among different classes of actor in a single value chain segment (e.g. the share of fish produced by small- and large-scale fishers).
Second, it is common for value chain analysis to describe technical aspects of the conduct of actors in some segments of the chain (e.g. the way in which are fish processed), but pay little attention to conditions under which inputs are sourced or assets are accumulated, or the way in which production activities are organized (e.g. use of family vs hired labor, or local vs migrant workers).
Third, chain performance is often assessed in terms of perceived inefficiencies or opportunities for technical interventions to ‘upgrade’ the performance of actors in the chain (e.g. by identifying ‘bottlenecks’ that can be addressed by training, or use of new technologies such as solar driers for fish). The social dimensions of value chain performance (e.g. gender relations) usually receive much less attention. These will be of particular interest to DFM, given its focus on the wellbeing of actors in dried fish value chains.
Fourth, ‘buyer driven’ value chains in which large firms such as supermarkets govern the behavior of actors upstream in the chain, and forms of governance based on standards, are applicable mainly to value chains oriented toward exporting to Northern countries. But most dried fish production is for domestic or regional markets where there is relatively little regulation (e.g. of food safety), so these models will rarely be relevant in DFM. (One possible exception is high value dried fish products sold through domestic supermarket chains). However, other forms of value chain governance play an important role in the chains studied by DFM. For example, the most powerful actors in dried fish value chain are often traders, and their relationships with producers and smaller traders affect the way that production and trade is organized. Governance of fisheries themselves (e.g. via customary arrangements, or decentralization of fisheries management, or IUU requirements), may also have important effects on the value chain. DFM will explore these types of issue.
Fifth, most value chain analysis is ‘static’: it produces a snapshot of a value chain at a single point in time. However, in reality, most value chains are highly dynamic. The structure and the conduct of actors may change rapidly in response to factors such as new fishing techniques, infrastructure, access to new markets, new policies, availability of workers, or the changing abundance of fish stocks. These changes can have important consequences for value chain performance. SVCA emphasizes the importance of these changes, and collects information about them by asking respondents to compare the situation at present with the situation in the recent past (e.g. at intervals of 5 or 10 years ago).
Finally, conventional value chain analysis is usually based on a combination of reviews of statistics and secondary sources, key informant interviews, and limited numbers of interviews with actors in key value chain segments. Although this can produce useful insights, there are several problems with this approach. These are outlined as follows:
(1) Because value chains are often highly dynamic, reliance on secondary sources and key informant interviews often reproduces information that is out of date or incorrect. For example, there is a widely held belief among policy makers in India, Bangladesh and China that exploitative credit relationships between paddy traders and farmers are widespread. SVCA conducted by Reardon et al. (2012) on rice value chains in these countries, found that trader credit to farmers has almost completely disappeared as a result of improvements in transport and communications and the investment of earnings from non-farm work in agriculture.
(2) Key informants often seek to represent situations in ways that they think will support their interests. For example, traders who wish to access loans from government banks on favorable terms may downplay their access to capital from other sources, or processors might report hiring only local labor rather than migrant workers when the latter is done illegally.
(3) Interviewing a relatively small sample of value chain actors on an opportunistic basis (e.g. by talking to whoever one meets first, or interviewing people recommended by officials or other key informants) may bias research findings. This is because the most easily accessible actors and locations, or actors known by key informants such as local officials, often have characteristics different to those of actors in remoter areas, or those who are less well known or less well connected.
(4) In most of the countries where DFM will be implemented there are few official statistics or detailed studies or on dried fish production, and those that do exist are often inaccurate. One of the purposes of DFM is to generate new data that will address this lack of information.