1. Introduction

The world food situation has been the focus of a flurry of recent publications aimed at providing greater insights into the evolution of global food supply, demand, and trade over the next few decades (Alexandratos 1995, 1996, 1997; Alexandratos and Bruinsma 1998; Delgado et al. 1999; Pinstrup-Andersen, Pandya-Lorch, and Rosegrant 1999; Rosegrant, Agcaoili-Sombilla, and Perez 1995; TAC 1996, 1997a). Most of this analysis, however, has focused on the past performance and future prospects for cereals and livestock. This paper analyzes recent trends in and alternative projections of the supply, demand, and trade for roots and tubers (R&T). In doing so, it seeks to provide a clearer vision of the contribution that these crops can make to the food systems of developing countries through the year 2020. A key objective of this paper is to clarify and, as much as possible, to quantify the complexity and magnitude of that contribution.

In 1995-97, the major R&T - cassava, potato, sweetpotato, and yam - occupied about 50 million hectares worldwide. Farmers produced 639 million metric tons (mt) of these crops annually, 70 percent of which were harvested in developing countries.¹ (See Box 1 for an overview of the variety of R&T.) Around 250 million mt of R&T were eaten in Asia, Africa, and Latin America, and nearly 100 million mt, almost all of it potatoes, in developed countries. The remainder was used as animal feed, planting material, processed products (for example starch), and other purposes. Production of the major R&T in developing countries alone had an estimated annual value of more than US$41 billion in 1995-97, nearly one-fourth the value of the major cereals (Table 1).

¹ Unless cited otherwise, the source data on historic supply and demand of agricultural commodities is FAO 1999a (updated April 1999, accessed July).

Individually, cassava, potato, sweetpotato, and yam rank among the most important food crops worldwide and, in terms of annual volume of production, cassava, potato, and sweetpotato rank among the top 10 food crops produced in developing countries.

The Roles of R&T in Developing-Country Food Systems

Many of the developing world's poorest producers and most undernourished households depend on R&T as a contributing, if not principal, source of food and nutrition (see, for example, Alexandratos 1995, 100-102). In part, these farm households value R&T because R&T produce large quantities of dietary energy and have stable yields under conditions in which other crops may fail (Alexandratos 1995, 189). R&T produce remarkable quantities of energy per day, even in comparison to cereals. Potatoes lead the way in energy production, followed by yam (Figure 1). In addition, some R&T are an important source of vitamins, minerals, and essential amino acids such as lysine (Low et al. 1997; Spencer and Associates 1997; Woolfe 1987, 1992).

In many parts of Sub-Saharan Africa (SSA), R&T are a major source of sustenance. They account for 20 percent of calories consumed in the region (Table 2). In 31 African countries with annual cassava production of more than 10,000 mt each, annual per capita consumption averaged 140 kilograms (kg) during the last four decades (Phil-lips 1998). Consumption in production centers and among the rural poor in many parts of the region greatly exceed this figure. Per capita consumption levels for cassava and the importance of R&T in the diet of many Africans, particularly less-well-off consumers, have remained remarkably constant despite drought, famine, wars, political and economic instability, regional population growth rates that averaged nearly three percent per year during the last 30 years, and growing urbanization. In addition, cassava leaf is an important source of protein in many parts of West and Central Africa (Spencer and Associates 1997).

Box 1: The Variety of Roots and Tubers R&T are frequently grouped together because they are bulky, perishable, and vegetatively propagated. At the same time these crops are highly differentiated in terms of origin, production and nutritional traits, and use. More than 30 edible and nonedible species of R&T are grown today. Foremost among them in terms of aggregate output and estimated value of production are cassava, potato, sweetpotato, and yam. Potato, cassava, and sweetpotato originated in Latin America (Horton 1988). Yam includes some species that have moved from Africa to North and South America, and others that have traveled from Asia to Africa (Hahn et al. 1987). Other prominent R&T include cocoyam, ginger, taro, and yam bean, as well as Andean R&T such as arracacha, mashua, oca, and ulluco. The latter group of plants is grown in the Andean region, other parts of South America, and East Asia. They are of minor importance globally in terms of total production and commercial value. Nevertheless, for particular countries, regions, or agroecologies, one or more of these other R&T can and do play an important role in food systems (Hermann and Heller 1997; Horton 1988). The variation in R&T growth patterns and production requirements helps to explain how particular commodities wedged their way into distinct production systems and varied consumption uses. For example, while a potato crop grown under irrigation in the lowland subtropics can mature in 100-120 days, cassava can take 9 to 24 months (see Appendix, Table 24). However, potato production requires adequate and timely availability of water during the crop's vegetative cycle, whereas cassava can be cultivated under near drought-like conditions. Conversely, cassava has been used more often for processed products because, among other things, it has a higher starch content on average (27-36 percent) than potato (13-16 percent) or sweet-potato (18-28 percent) (Appendix, Table 25). Moreover, in spite of their bulkiness and perishability, most R&T have proven remarkably mobile over millennia. Other differences among R&T include such things as their enormously dissimilar genetics; the diverse strategies required for genetic improvement to take account of their variable production systems and end-uses; the distinctions between their pest and pathogen complexes; and the differences in their policy environment (see TAC 1997a, 20-24 for further details).

Table 1 - Production of, edible energy and protein in, and value of major roots and tubers and cereals in developing countries, 1995-97

		Price	Production	Edible energy	Edible protein	Value
Commodity		(US$/mt)	(million mt)	(trillion kilocalories)	(million mt)	(billion US$)
Cassava		53	165.3	142	0.7	8.8
Potato		157	105.3	65	1.8	16.5
Sweetpotato		88	137.0	127	1.9	12.1
Yam		130	31.5	28	0.5	4.1
	Major R&T		439.1	362	4.9	41.4
Maize		126	257.6	786	20.1	32.5
Milled ricea		284	350.0	851	15.7	99.4
Wheat		146	272.2	687	27.4	39.7
	Major cereals		879.8	2,324	63.2	171.6

Source:	Basic data from FAO 1998a (FAOSTAT June 1998, accessed July 1998).
Note:	Coefficients for calculating edible energy and protein are based on Horton 1988. Prices are based on estimates for 1993 and 2020 IMPACT baseline scenario (see Chapter 4) interpolated for 1995-97.
a Milled rice is more readily comparable to the other commodities for the purposes of comparing international prices.

Figure 1 - Edible energy produced by major roots and tubers and cereals

Source: Horton and Fano 1985.

In much of Asia and Latin America, R&T provide an important, supplemental source of carbohydrates, vitamins, and amino acids in food systems that are dominated by other commodities. India, for example, is now among the world's largest potato producers, having achieved a phenomenal growth rate in potato production of 6 percent per year during 1962-96. India produced 25 million mt in 1997 - a level surpassed only by China and the Russian Federation.² Nearly all of India's production is harvested in the cool, dry, winter months, when cereals are in seasonally short supply in many parts of the country, and often in water-scarce areas where irrigated rice cannot be cultivated. Similar trends in production growth (4.4 percent per year in 1962-96) have prevailed in Bangladesh, where results of a national rural nutrition survey carried out during 1981-82 showed that 15 percent of the vitamin C intake came from potatoes (Scott 1988a).

² On a per capita basis, 7 out of the world's 10 largest potato producers in 1997 were located in Eastern Europe and the Former Soviet Union.

Table 2 - Percentage of calories and protein from consumption of roots and tubers as food, 1983 and 1996

	Calories		Proteins
Country/regiona	1983	1996	1983	1996
	(percent)
China	8.5	5.6	4.2	2.7
Other East Asia	1.9	1.8	1.1	1.4
India	1.9	1.3	1.2	0.9
Other South Asia	1.7	1.3	1.0	1.0
Southeast Asia	5.7	4.3	1.8	1.4
Latin America	4.8	4.3	2.8	2.6
WANA	1.9	2.3	1.4	1.6
Sub-Saharan Africa	18.7	20.1	6.6	8.0
Developing	6.3	5.4	2.9	2.6
Developed	4.3	4.3	3.4	3.3
World	5.7	5.1	3.1	2.8

Source:	FAO 1999b.
Note:	1983 is average for 1982-84 and 1996 is average for 1995-97 for all tables unless indicated otherwise; WANA is West Asia and North Africa.
a Other East Asia includes Hong Kong, Macau, Mongolia, North Korea, and South Korea. Other South Asia includes Afghanistan, Bangladesh, Bhutan, Maldives, Nepal, Pakistan, and Sri Lanka. Southeast Asia includes Brunei, Cambodia, East Timor, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand, and Vietnam. Latin America covers Central and South America and the Caribbean. WANA includes Algeria, Bahrain, Cyprus, Egypt, Gaza Strip, Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Qatar, Saudi Arabia, Syria, Tunisia, Turkey, United Arab Emirates, Western Sahara, and Yemen. Sub-Saharan Africa includes Central West, Eastern, Northern, and Southern Sub-Saharan Africa.

Production and use of R&T tend to be concentrated in countries with lower per capita incomes (Scott and Maldonado 1999). Within low-income countries, R&T frequently play a relatively greater role in the food systems in remote, often marginal, areas with particularly low income levels and limited access to farm inputs. Production and use of sweetpotato is thus more prominent in Sichuan province, China (Gitomer 1996), in eastern India (Dayal et al. 1995), or northern Uganda (Scott et al. 1999). Cassava is more prominent in northeast Brazil (Ostertag and Herrera 1992) and northeast Thailand (Titapiwatanakun 1998), and potato in the highlands of Guatemala (El Cid 1992) or Peru (Scott 1985). In addition to their role as local staples or complementary sources of energy, R&T serve as food security crops. They alleviate seasonal shortages and fill food gaps caused by natural or man-made disasters (see, for example, Tanganik et al. 1999).

More than simply food crops for the rural poor, R&T can also serve as sources of cash income for low-income farm households and raw material for processed products for both rural and urban consumption. Growth in the latter uses of R&T is relatively new and reflects the underlying dynamism of the R&T sector in many developing countries. In Africa, cassava, in addition to being a cheap, starchy staple, graduated from on-farm consumption to cash crop for sale to both urban and rural consumers (Nweke 1992; Nweke et al. 1994).

R&T have also increased in importance in Asia. Potato production in Asia now accounts for nearly 80 percent of total production in developing countries. Asia's share of global potato output soared from 7.5 percent in 1961-63 to 28.2 percent in 1995-97. With rapid economic growth in many parts of Asia, consumers have increasingly diversified their food intake from strictly cereal-based diets to greater consumption of potatoes, milk, meat, and other commodities. Empirical evidence shows that the overwhelming bulk of potatoes produced in Asia are sold for cash by small farmers (Scott 1997). High yields mean that on-farm food needs can be met by only a fraction of the harvest, with strong off-farm demand using the surplus.

Feed, processed food, and other, nonfood uses for cassava and sweetpotato have also expanded considerably in Asia over the last three decades (Pham et al. 1996; Scott 1992; Titapiwatanakun 1998). Rapid growth in demand for meat has created growth opportunities for producers in more remote areas to use R&T as animal feed. Such is the case for cassava in northern Vietnam (Nguyen 1996) and sweetpotato in Sichuan province, China (Huang 1999). Small farmers in China who have long cultivated sweetpotato as a food security crop, now process roughly half of their annual harvest of 118 million mt (1995-97 value) into animal feed.

It is estimated that these farmers convert another 20 to 30 percent of annual sweetpotato output into starch for noodles and other processed products (Huang 1999; Timmins et al. 1992). In Vietnam during 1995-97, roughly 50 percent of the annual cassava harvest of 2 million mt was processed into feed and an additional 25 percent was used to make starch (Goletti, Rich, and Wheatley 1999).

Roots and tubers also help alleviate poverty by providing employment opportunities in production, processing, and marketing. Farm surveys in Bangladesh (Scott 1988a), Egypt (Crissman et al. 1991), Colombia (Rodriguez 1996), and Rwanda (Braun, de Haen, and Blanken 1991), for example, found that potato production requires between 120 and 450 labor days per hectare per crop, work totals far greater than those for many other crops. In South Asia, jobs in potato production particularly help the landless, who make up a substantial percentage of the rural population (Scott 1988a). A recent study of the potato sector in Bolivia estimated that the crop generated over 12 million labor days per year nationwide (Zevallos 1997). Surveys on sweetpotato and cassava production have found that labor is the most important cost of production (see, for example, Achata et al. 1990; Pham et al. 1996; Cabanilla 1996). Root-crop processing for feed or starch also is often highly labor-intensive, thereby providing off-season employment and income to the rural underemployed (see, for example, Goletti, Rich, and Wheatley 1999; Nave and Scott 1992; Nweke 1992; Simpson, Cheng, and Miyazaki 1994; Timmins et al. 1992). Moreover, particularly in Sub-Saharan Africa, root and tuber production, processing, and marketing provides important income-earning opportunities for women (Gatumbi and Hagenimana 1998; Low 1998; Nweke 1992; Owori and Hagenimana 1998).

The role of R&T in developing-country food systems also raises interest about the impact of these crops on the environment. Recent case studies show instances of pesticide toxicity associated with potato production (Crissman, Antle, and Capalbo 1998); water pollution from cassava processing (Goletti, Rich, and Wheatley 1999); soil erosion linked to cassava cultivation (Howeler 1996); and loss of biodiversity for potato (Brush, Taylor, and Bellon 1992) and Andean roots and tubers (Hermann and Heller 1997) as a result of increased commercialization of production. But R&T also hold the promise of helping to alleviate environmental problems. Sweetpotato, for example, can serve as a quick cover crop to reduce soil erosion (Orno 1991). Production of potato using botanical or true potato seed can increase genetic diversity because each seed constitutes a distinct genetic entity (Upadhya et al. 1995).³

³ True potato seeds are the tiny seeds - smaller than tomato seeds - found in potato fruits. Potato plants and tubers grow from these seeds. The tubers can then be consumed or replanted as potato seed (Upadhya et al. 1995).

Production Performance of R&T

During the past four decades, developing-country food production policy has focused on achieving growth in wheat, rice, and, more recently, maize. With technological innovations resulting in high-yielding varieties of these basic staple foods, growth rates for cereal production in developing countries, and particularly in Asia, rose rapidly. Similar growth rates in production were achieved for potato and yam, particularly during the last two decades. Growth rates for cassava and sweetpotato were much lower (see Figure 2). The tendency to treat R&T as undifferentiated commodities has obscured their variable performance and clouded understanding of their future prospects (McCalla 1998). Furthermore, in contrast to cereals, growth rates for cassava, potato, and yam in developing countries were driven by an expansion in area planted rather than yields. Average yields for potato and sweetpotato in developed countries (cassava and yam are chiefly developing-country crops) remain well above those in developing countries, where yields fall far short of technically feasible levels. As the recent review of R&T by the Technical Advisory Committee (TAC) to the Consultative Group on International Agricultural Research (CGIAR) noted, “... one of the greatest similarities among root and tuber crops is unrealized yield potential that could be attained through yet-to-be-developed technologies... All too frequently this is because the needed technology is not available to deal with yield-limiting factors (water, nutrients) and yield-reducing factors (disease, pests).” (TAC 1997b, 22). The review went on to note that prospects for increasing yields of R&T appear to be much greater than the “attempts to increase the physiological yield potential of crops already trapped on a yield plateau...” This observation suggests that achieving R&T yield increases appears to be less formidable and costly a task than achieving similar gains with other crops.

Figure 2 - Production growth rates of major roots and tubers and cereals, developing countries, 1961-63 to 1995-97

Source: FAO 1999a.

Given the important contribution R&T can make to the diets and livelihoods of over 2 billion people in the tropics and subtropics, and the potential that exists for expanding production and use, R&T have recently become the subject of increasing attention (see, for example, dTp Studies 1998; Horton 1981,1988; Horton, Lynam, and Knipscher 1984; Plucknett, Phillips, and Kagbo 1998; Sarma 1989; Scott 1994a, 1997; TAC 1997b, 15; Woolfe 1987, 1992).⁴ And yet, there is a growing sense that the role and importance of R&T in the global food system are often poorly understood. This situation makes the present study on trends and future prospects for R&T particularly timely and relevant. A better understanding of the contribution R&T can make to poverty alleviation, food security, economic growth, and environmental sustainability in developing countries can improve the livelihoods and well-being of more than a third of the world's population. At the international and, perhaps even more importantly, national level, results from the analysis presented here can help guide investment decisions in agricultural research, extension, and capacity development to make R&T even more productive, marketable, and accessible for developing-country populations.

⁴ De Bruijn and Fresco (1989) estimated cassava alone as an important food crop for 500 million people in developing countries. The 1993-94 national survey of consumer expenditures in India interviewed over 115,000 households as a representative sample nationwide and found that over 85 percent - both rural and urban - reported consuming potato and that consumption was remarkably widespread across states and throughout the year (GOI 1997). Potato consumption elsewhere, combined with sweetpotato and yam in Asia, Africa, and Latin America, according to our estimates, put the aggregate figure between two and three billion (see Gitomer 1996; Lev and Shriver 1998; Woolfe 1987, 1992).

Objective and Scope of This Study

This paper attempts to provide a better understanding of the contribution roots and tubers will make to global food systems in the decades ahead. It analyzes trends and projections for the production and use of the major R&T (potato, sweetpotato, yam, and cassava) in developing countries and discusses the factors that have influenced and will influence the supply of and demand for these commodities.

Given the important differences in the patterns of production, use, and trade among R&T commodities, as well as differences at the regional and national levels, the analysis will disaggregate R&T by crop. Limited time, resources, and available data prevent further specificity. It is hoped that this study will serve as the basis for detailed reviews of other R&T, both at the global and regional levels, particularly in cases where such crops are important for local food security.

The rest of this paper is divided into six chapters. Chapter 2 discusses recent trends in the demand and use of R&T. Chapter 3 examines production trends for R&T, emphasizing the sharp regional and commodity-specific differences among these crops. Chapter 4 describes the IMPACT global food projections model and then presents baseline projections of demand and supply for R&T to the year 2020. Chapter 5 offers projections based on an alternative, high demand and production growth scenario. Chapter 6 considers the environmental implications of R&T production and use. The final chapter discusses the principal technological, institutional, and policy implications of this analysis.

The paper argues for the sustained and possibly increased importance of R&T in developing countries in the decades ahead. It also contends that achieving the potential for R&T is not a given. Rather this will require continued investments in agricultural research and institutional development, and a policy framework conducive to R&T.

Improved technology that raises productivity will be critical for increasing the availability of R&T in developing countries in the next two decades, as will post-production innovations. The historical record suggests that it is feasible to expect such innovations to occur during the time period in question. The paper points out that for research planning and resource allocation purposes, as well as for developing appropriate policy measures, it is useful to distinguish between supply-side constraints and demand-side constraints and to determine their relative importance for particular R&T in specific developing-country contexts. Ample scope exists for overcoming both types of constraints.

The paper argues that policy issues for developing countries include (1) removing policy distortions that bias market signals in favor of other agricultural commodities, and (2) giving farmers and entrepreneurs nondistorting incentives to invest in production and post-production innovations for R&T. Policy measures in industrialized countries include (1) removing subsidies on competing crops, and (2) lifting trade restrictions on imports from developing countries.