1 Introduction

In the late twentieth century, the world experienced a dramatic increase in agricultural productivity. The Green Revolution occurred with the introduction of high-yielding varieties (HYVs), mainly in the major grains of rice, wheat, and maize. According to a recent estimate, the introduction of HYVs increased yields of food crops by 44% between 1965 and 2010 (Gollin et al. 2021). The revolution indeed played a pivotal role in avoiding the Malthusian trap, much-touted at the time.

The increase in agricultural productivity and the associated decline in real food prices led to poverty reduction at the micro-level, especially in Asian countries (e.g., Datt and Ravallion 1998). The impact of the Green Revolution is indeed different between favorable and less favorable environments. However, these unfavorable environments also benefited through technology spillovers and migration (David and Otsuka 1994). Furthermore, poverty reduction led to higher human capital investment, expediting the development of the nonfarm sector (Estudillo et al. 2006).

The effect of the Green Revolution on economic development and structural change can also be found at the macro level. For example, Gollin et al. (2021) show that the increase in agricultural productivity and the gross domestic product due to the Green Revolution is associated with the decrease in the share of agricultural employment. Admittedly, some Asian countries experiencing the Green Revolution are now considered emerging economies, and the role of agriculture as a driver of economic growth is declining.

Traditionally, economic development has been associated with industrialization, and agricultural and industrial sectors have often been viewed as dichotomous (e.g., Lewis 1954; Harris and Todaro 1970; Matsuyama 1992). However, at the micro-level, the change can be more gradual. For example, historical evidence shows that the Dutch cultivation system in colonial Java led to the development of the downstream industries that used the produced crop (e.g., Dell and Olken 2020). Thus, the development of an agriculture-related industry can be a crucial missing link in the dichotomy between agricultural and industrial development.

This chapter aims to discuss the issues of integrating agricultural development into broader economic development and poverty reduction, employing the perspective of global value chains (GVCs). In the modern economy, the production process is fragmented worldwide (e.g., Feenstra 1998; Timmer et al. 2014). However, the level of value-added is heterogeneous across production stages, and firms in developing countries tend to be trapped in the low value-added middle stage of the production process due to their comparative advantages in labor-intensive industries. Similar arguments hold for agriculture and how linking the development of the upstream and downstream industry to farm production is a key to integrating agricultural development and industrial development. In this regard, we aim to redefine agricultural development as a part of long-term economic development.

2 Conceptual Framework

2.1 Upstream and Downstream Industries in Food Value Chain

Our benchmark concept in analyzing agricultural development in terms of GVCs is ‘smile curves’ (e.g., Mudambi 2008). It is a graphical depiction of the difference in value-added across the production stages: the value-added is high for firms located upstream and downstream of the production process. However, it is low for firms in the middle stages of production. Because of their comparative advantages in labor-intensive industries, firms in developing countries tend to be trapped in this middle stage. Thus, they need to involve upstream and downstream industries via functional upgrading (e.g., Murakami and Otsuka 2020). The concept was originally proposed to illustrate the manufacturing value chains, though it also helps analyze the agricultural value chains.

To define the upstream and downstream industries of agriculture, we need to consider the complex structure of the value chain. Reardon et al. (2019) define the food system as a ‘dendritic cluster’ of six value chains: (1) ‘core’ supply chain of agricultural products; (2) ‘feeder’ supply chain of farm input; (3) ‘feeder’ supply chain of post-farmgate segments; (4) ‘pan-system feeder’ supply chain of financial credit; (5) ‘feeder’ supply chain as a broad set of public assets; and (6) ‘feeders’ supply chain of technology and product innovations via research and development (R&D). However, for simplicity, this chapter focuses only on industries directly connected to crop production as an input/output relationship. Specifically, we set farm production as a midstream and define the agricultural R&D sector as upstream and the food processing and the retailing sector as downstream.

2.2 Upstream: Agricultural R&D Sector

Scientific knowledge is considered a public good due to its non-excludable accessibility and non-rival use. Therefore, without strict patent control, agricultural R&D faces underinvestment. Furthermore, the development of new varieties, especially those specific to the environment of developing countries, is not a profitable market for leading biochemical companies. For this reason, the public sector, especially international research institutes, has played an essential role in agricultural R&D.

The development of HYVs that enabled the Green Revolution was led by the international research institutes currently known as the Consultative Group for International Agricultural Research (CGIAR). The International Rice Research Institute (IRRI) and the International Center for Maize and Wheat Improvement (CIMMYT) played central roles in the Green Revolution. However, CGIAR spending peaked in 2000 and has since gone downward, accounting for only 1.1% of global agricultural R&D as of 2015 (Alston and Pardey 2021). Instead, the role of the private sector has been increasing over time, and the discrepancy across countries can be accelerated.

Figure 22.1 shows the food and agricultural R&D per capita trend by income levels. The spending on R&D is increasing over time, except in low-income countries. The shares of upper- and lower-middle-income countries in R&D spending have been increasing since 2000, mainly due to the rise of China and India. However, most agricultural R&D is still conducted in high-income (and upper-middle-income) countries. Since the sector is highly knowledge- and capital-intensive, the comparative advantage still exists in developed countries.

Fig. 22.1
A bar graph on food and agricultural R and D per capita depicts dollars per person versus the year. The legend reads, high income, lower middle income, upper middle income and low income. 2010 depicts the highest for all categories except low income. Low income remains almost the same.

Food and agricultural R&D per capita by income levels (Pardey et al. 2016)

Full size image

It is also important to note that the research agenda in the agricultural R&D sector has been changing from productivity to non-productivity. For example, the investment shares in productivity-oriented activities decrease over time within CGIAR (Pingali and Kelley 2007). Similarly, in the United States, the research agenda has been shifting from productivity-oriented to quality-oriented, focusing on food safety, human nutrition, food security, natural resources, and environmental concerns (Alston and Pardey 2021).

However, the lack of comparative advantages do not necessarily justify underinvestment in agricultural R&D. For example, in Sub-Saharan African countries, agricultural productivity growth continues to be an important issue (e.g., Otsuka and Larson 2016; Otsuka and Muraoka 2017). Furthermore, recent plot-level data evidence shows that most of the actual yield differences would disappear if countries produced according to their potential yield (Adamopoulos and Restuccia 2021). Thus, identifying the factors that hinder efficient resource allocation is still essential in agricultural development. In this respect, agricultural extension services and technology localization are important issues. Even though the role of the public sector is declining, it is still important for developing-country governments to continue to work with the CGIAR on localizing advanced technologies, as the social benefits outweigh the individual benefits in this sector.

2.3 Downstream: Food Processing and Retail Sector

In the downstream of the modern food supply chain, the role of supermarkets is becoming essential, especially since the 1990s. While the development is driven by income growth and urbanization on the demand side, the supply side is driven by foreign direct investment (FDI), logistics technology, and inventory management (Reardon et al. 2003). The rapid development of supermarket chains also changes the food procurement system, leading to the development of fresh products and food-processing industries (e.g., Gereffi et al. 2005). It is also important to note that FDI leads to such value chain development while domestic investment is increasing in some Asian countries (Reardon et al. 2012).

In this context, Reardon et al. (2012) advocate development strategies that “bring the markets to the farmers” (p. 12,336). The concept is to create a cluster to link farmers to downstream channels, providing them with missing services and products by utilizing economies of agglomeration. By doing so, small-scale farmers can participate in the food value chain, and vertical integration is expected to lead to industrial development.

Such a concept is parallel to the model of cluster-based industrial development (Sonobe and Otsuka 2006, 2011). It describes industrial development as a process from an imitation phase to quantity expansion and quality expansion phases. A key factor that drives the phase shift is human capital investment: as the shift continues, entrepreneurs are required to innovate in various ways, including product differentiation and management techniques, which require a higher level of human capital.

Another critical factor in facilitating industrial development is the existence of associations. Otsuka and Sonobe (2018) summarize the role of associations in (1) marketing research, (2) dissemination of new production methods and management practices, and (3) ensuring the quality of products. The role of associations is indeed highlighted in the studies of industrial development (e.g., Hashino and Otsuka 2013). Thus, utilizing the role of associations can accelerate the economic merit of agglomeration.

In light of the cluster-based development model, it is thus essential to take advantage of industrial clusters by attracting FDI to acquire advanced technologies and investing in human capital and associations to enhance absorptive capacity. Furthermore, the development of the downstream industries can increase the demand for crops, integrating the development of the agricultural and industrial sectors.

3 Plantation Agriculture and Contract Farming

The vertical integration of the production and post-production stages is not new. Historically, we can find an example in plantation agriculture. Plantation agriculture is known as cash crop production by a large number of unskilled laborers under the supervision of a management body. In general, family farms have an advantage in agricultural production because they have less incentive to shirk work than hired labor (e.g., Hayami and Otsuka 1993). However, plantation agriculture occurs when there is large scale merit in the post-production agricultural processing stages (Hayami 2010). For example, tea production has economies of scale in the fermentation process, which requires strong coordination between farm production and processing. To meet this end, tea companies have an incentive to internalize tea farming as long as the benefit from scale outweighs the inefficiency of crop production. However, as Hayami (2010) argues, the role of plantations has been declining due to the increasing drawbacks of the system.

A modern counterpart of the system can be found in contract farming. Contract farming is an agreement between a grower and a processor regarding the production of an agricultural commodity (Bellemare and Bloem 2018). In practice, it refers to a pre-harvesting contract between a farmer and a processor/firm to deliver an agreed quantity and quality of product, usually at a predetermined price. In relation to GVCs, an important issue is the contract farming between firms in developed countries and farmers in developing countries.

Otsuka et al. (2016) classify contract farming into production and marketing. In a production contract, the contractor provides essential credit, technical assistance, and inputs in return for the delivery of harvest or produce. The contractor strictly controls production and farm management decisions under the production contract. On the other hand, the marketing contract primarily leaves production autonomy to the growers. Such differences can be interpreted as ‘captive’ and ‘relational’ in the typology of GVCs (Gereffi et al. 2005). The captive firm refers to the case where small suppliers are transactionally dependent on large buyers, facing high switching costs. In contrast, the relational firm refers to the case where complex interactions between buyers and sellers create mutual dependence and high levels of asset specificity. One of the significant differences between these two types is the supplier’s ability. Thus, captive firms must achieve functional upgrading to be relational firms (Murakami and Otsuka 2020). Similarly, farmers in production contracts need to enhance their managerial ability to enter marketing contracts and obtain larger residual claims.

4 Case Study: Tapioca Industry in Thailand

An interesting example illustrating the relationship between agricultural development and industrial development is Thailand’s cassava/tapioca industry. Thailand is the largest exporter of cassava products and the third-largest producer of cassava roots after Nigeria and DR Congo.

Figure 22.2 shows the long-term trend of the area harvested and production amount of cassava since 1960. Cassava production in Thailand experienced a remarkable increase in the last 60 years. Production and the harvested area have expanded by about 16 times. Until the year 2000, the trends of both indices were almost parallel: the harvested area and production increased until around 1990 and decreased in the next ten years. After that, however, the increase in production amount outweighed that of the harvested area, suggesting a clear increasing trend in yield, especially after 2000.

Fig. 22.2
A line graph on Cassava Production in Thailand depicts area harvested against the year. The area harvested peaks above 1600 hectares in 1990, and the production peaks to 32,000 in 2018. All values are approximated.

Trend of cassava production and harvested area in Thailand (FAOSTAT 2022)

Full size image

Part of the increase in productivity can be attributable to the HYVs of cassava. In particular, Kasetsart 50 is an important breeding achievement. It was released from 1992 to 1995 as a result of a public breeding program involving Kasetsart University and the International Center for Tropical Agriculture (CIAT). It had a higher yield and starch content than the previous varieties and is considered a successful cassava breed in Thailand (Malik et al. 2020). As a result, it still occupies a large share of the planting area in Thailand and Southeast Asian countries (Labarta et al. 2017).

Another important factor that drives cassava production development is the demand side. Figure 22.3 shows the export trend of dried cassava and cassava starch. Interestingly, until around the mid-1980s, most cassava products were exported as a form of dried cassava. However, cassava starch exports began increasing in the 1990s, whereas dried cassava exports temporarily dropped from the 1990s to 2000. As a result, more than half of current cassava products are exported as starch, with higher value-added. Importantly, in tandem with yield growth, the export of cassava products has been increasing since 2000. Such a trend is mainly driven by expanding trade volume to China. In 2020, 99% of cassava roots and 63% of cassava starch were exported to China; in 1989, shares for both were less than 1%.

Fig. 22.3
A line graph on Cassava export in Thailand. The export of dried cassava peaks above 1,500,000 in 2018, and the export of cassava starch peaks to 1,200,000 in 2018. All values are approximate.

Export of cassava products in Thailand (FAOSTAT 2022)

Full size image

Associations also appear to have played an essential role in the development of the industry. There are several cassava/tapioca industry-related associations. One of the most influential is the Thai Tapioca Starch Association (TTSA). TTSA is an association of firms producing tapioca starch and was awarded the license under the trade association act in 1976. Its primary mission is price stabilization. Since market price fluctuates as per the seasonal fluctuation of cassava supply, TTSA provides the venue to address the issue and identify possible solutions. Frequent meetings can also serve as a place for information sharing among producers. Although membership is voluntary, TTSA claims that almost all tapioca starch manufacturers in Thailand are association members or affiliated companies (TTSA [n.d.]).

The geographic distribution of TTSA member firms is also informative in discussing patterns of industrial clusters. Panel A of Fig. 22.4 shows the map of cassava yield in 2010, calculated from FAO-Global Agro-Ecological Zoning (GAEZ) data. The yield is high in the northern and northeastern regions. In contrast, the central and the eastern regions show low yields as they are more urban. Panel B shows the plot of TTSA membership firms. The firms are concentrated in the central and the northeastern regions. Two important centers of the industry are Bangkok and Nakhon Ratchasima province. The firms in Bangkok are typically the main offices, and their factories are in other cassava-producing provinces. In contrast, firms in Nakhon Ratchasima, known as a processing center of tapioca, utilize the advantage in procuring cassava roots and saving on transportation costs.

Fig. 22.4
2 maps illustrate a. the cassava yield in 2010, and b. T T S A membership firms are concentrated in the central and northeastern regions.

Cassava yield and TTSA membership (FAO and IIASA 2021; TTSA [n.d.])

Full size image

According to an interview with factory managers in Nakhon Ratchasima, many of them procure cassava roots on the spot rather than through contract farming. One possible reason is that cassava roots are relatively storable. Although a constant supply of raw materials is necessary for operations, this is not the case in reality if storage is possible (Hayami 2010). Indeed, starch factories have large warehouses for storing raw cassava roots. Other reasons can be that many cassava farmers sell roots to processors, and differences in quality, other than starch content, are not very important.

Cassava products in Thailand can be classified mainly into four categories: chips, pellet, native starch, and modified starch (Fig. 22.5). The simplest product is chips, often processed by local small-scale entrepreneurs. Chips are further processed into pellets for feedstuff, mainly for the foreign market. Another important production process is the extraction and modification of cassava starch. Cassava roots are first processed into native starch for domestic consumption and export. The native starches produced are further processed physically, enzymatically, or chemically into modified starch for multiple applications, including food products, pharmaceuticals, and binders, mainly for export. Modified starch production is highly capital- and knowledge-intensive, making higher value addition than native starch.

Fig. 22.5
The cassava products are classified into chips and native starch. The chip further has pellets, and the native starch has modified starch.

Classification of cassava products

Full size image

Table 22.1 shows the number of firms producing native and modified starch among the members of TTSA. Thirty out of 95 firms produce modified starch, while 83 firms produce native starch. This is because the production of modified starch is capital-intensive, requiring a higher level of production technology. Interestingly, 18 firms produce both native and modified starch, indicating the existence of economies of scope between these products. In terms of the cluster-based development model, the development of chips and native starch production, which require relatively low entry costs, can be viewed as the quantity expansion stage. Thus, to shift to the quality expansion stage of modified starch production, further human capital investment and foreign collaboration to absorb new technologies would be essential.

Table 22.1 Production of native and modified starch. (TTSA website, n.d.)
Full size table

To summarize, the development of the cassava-related industry in Thailand is driven by the introduction of HYVs on the supply side and integration into the global value chain, especially in China, on the demand side. The industry’s development process can be explained by the model of cluster-based development, where associations play an important role. However, further investment in human capital and attracting FDI is essential for the firms to enter the production of a higher value-added starch product.

5 Conclusion

In this chapter, we discussed the integration of agricultural and industrial development. To overcome the dichotomy of the agricultural and industrial sectors, we need to pay close attention to the development of the upstream and downstream industries of the agricultural value chain. Especially in terms of GVCs, it is important to integrate crop production and downstream food processing and retail industries. As an illustrative case study, we provided an overview of the development of the tapioca industry in Thailand, which has utilized the merit of cluster-based industrial development. On the other hand, due to the high data constraints of such a topic, elucidating the long-term industrial development process based on micro-level data remains a significant future challenge.

Recollections of Professor Keijiro Otsuka

I first met Otsuka sensei at a small conference as a graduate student. I still vividly remember the advice he gave me at the reception. After finishing my Ph.D., I joined GRIPS as a postdoctoral research fellow, though I did not have the chance to interact with him very much. However, after I joined IDE-JETRO, I was fortunate enough to start a research project with him focusing on the role of FDI in industrial development in Thailand and India. I have been learning his research style closely and intensively since then. I am greatly indebted to him for his mentorship and honored to be part of the Festschrift.