Scholarly article on topic 'High rural population density Africa – What are the growth requirements and who participates?'

High rural population density Africa – What are the growth requirements and who participates? Academic research paper on "Agriculture, forestry, and fisheries"

CC BY-NC-ND
0
0
Share paper
Academic journal
Food Policy
OECD Field of science
Keywords
{Africa / "Structural transformation" / "Agricultural intensification" / Smallholders / "Development strategy" / "Green revolution" / "Rural poverty reduction"}

Abstract of research paper on Agriculture, forestry, and fisheries, author of scientific article — John W. Mellor

Abstract A large and increasing proportion of agricultural growth in Africa must come from continuous gains in land productivity in areas of high population density and hence with already relatively high yields. What that requires is analogous to the green revolution in Asia. Several features differentiate the African situation. Those include greater diversity in cropping pattern including a historically larger and more widespread tropical commodity export sector. The physical infrastructure in rural Africa is far inferior to that of most Asian countries. While the greater diversity of agriculture calls for a larger and more diverse institutional structure the reality is that the research systems, the ancillary education systems to spread innovation and the rural financial systems are generally greatly inferior to those of Asia at the beginning of the green revolution. Ethiopia’s record of a steady six to seven percent growth for agriculture and nearly halving of rural poverty demonstrates that with the right policies and investments a very poor country starting with poor physical and institutional infrastructure can bring a major contribution from agriculture growth to increased GDP and reduced poverty. As in Asia, the bulk of accelerated agricultural growth will come from small commercial farmers. They have sufficient farm income to reach or exceed the poverty level. Those are farms with, depending on the country, as little as 0.75 hectares to a few tens of hectares of land. They comprise up to half the rural population and produce on the order of 70–80 percent of agricultural output. They are in general not poor. The poor have inadequate land to reach the poverty level, initially with much underemployment, and with substantial non-farm employment. The primary driver of poverty reduction is the small commercial farmer spending on the order of half of increased income on nontradable, employment intensive goods and services from the rural non-farm sector.

Academic research paper on topic "High rural population density Africa – What are the growth requirements and who participates?"

Food Policy xxx (2014) xxx-xxx

Contents lists available at ScienceDirect

Food Policy

journal homepage: www.elsevier.com/locate/foodpol

High rural population density Africa - What are the growth requirements and who participates?

John W. Mellor *

Charles H Dyson School of Applied Economics and Management, Cornell University, USA John Mellor Associates Inc., Washington, DC, USA

ARTICLE INFO ABSTRACT

A large and increasing proportion of agricultural growth in Africa must come from continuous gains in land productivity in areas of high population density and hence with already relatively high yields. What that requires is analogous to the green revolution in Asia. Several features differentiate the African situation. Those include greater diversity in cropping pattern including a historically larger and more widespread tropical commodity export sector. The physical infrastructure in rural Africa is far inferior to that of most Asian countries. While the greater diversity of agriculture calls for a larger and more diverse institutional structure the reality is that the research systems, the ancillary education systems to spread innovation and the rural financial systems are generally greatly inferior to those of Asia at the beginning of the green revolution. Ethiopia's record of a steady six to seven percent growth for agriculture and nearly halving of rural poverty demonstrates that with the right policies and investments a very poor country starting with poor physical and institutional infrastructure can bring a major contribution from agriculture growth to increased GDP and reduced poverty. As in Asia, the bulk of accelerated agricultural growth will come from small commercial farmers. They have sufficient farm income to reach or exceed the poverty level. Those are farms with, depending on the country, as little as 0.75 hectares to a few tens of hectares of land. They comprise up to half the rural population and produce on the order of 70-80 percent of agricultural output. They are in general not poor. The poor have inadequate land to reach the poverty level, initially with much underemployment, and with substantial non-farm employment. The primary driver of poverty reduction is the small commercial farmer spending on the order of half of increased income on nontradable, employment intensive goods and services from the rural non-farm sector.

© 2014 Elsevier Ltd. All rights reserved.

Article history: Available online xxxx

Keywords: Africa

Structural transformation Agricultural intensification Smallholders Development strategy Green revolution Rural poverty reduction

Background

The other papers in this issue document that in rural Africa, a large area, number of families, and volume of production are characterized by high population densities with increased inequality of income and assets. In such areas, growth in production and income can occur only modestly from enhanced area and must increasingly be dominated by expanded value of output per hectare. That suggests an African green revolution as the appropriate approach and from that an appeal to Asian models.

The African Union has sponsored an intensive exercise to provide an appropriate model. That model, CAADP, has been signed off on by the heads of state of all the African countries (African Union, 2010). It prescribes a rapid six percent agricultural growth rate, a minimum ten percent of government expenditure on

* Tel.: +1 2025504398; fax: +1 2023478802. E-mail address: jmellor@jmassocinc.com

http://dx.doi.org/10.1016Zj.foodpol.2014.03.002 0306-9192/© 2014 Elsevier Ltd. All rights reserved.

agriculture, and agricultural growth recommendations explicitly modelled on the green revolution in Asia. It should be kept in mind that green revolution technology, although biologically based, substantially increases labor productivity. In India the elasticity of employment with respect to output was typically about 0.3 (Rao, 1975).

However, for Africa, the national level success stories in achieving such growth are few. This text draws attention to three countries sometimes cited as success stories, but with differing levels and means of achieving accelerated agricultural growth. Each made the CAADP strategy central to its approach with national adaptation. Rwanda and Ethiopia set a higher than CAADP growth target of eight percent, while Ghana set six percent (Ghana, Government of, 2010; Rwanda, Government of, 2007; Ethiopia, Government of, 2010, 2009). Each was explicit that this was a ''Green Revolution'' strategy. Each accepted in principle the CAADP target for a ten percent share of government expenditure to agriculture. Rwanda started with seven percent and then raised that

J.W. Mellor/Food Policy xxx (2014) xxx-xxx

to ten percent with a stated goal of 15 percent. Ethiopia exceeded the ten percent target by 50 percent. Ethiopia and Rwanda have been explicit in setting a high fertilizer growth target (Ethiopia, Government of, 2010, 2009; Rwanda, Republic of, 2007, 2012). Ethiopia started earlier than the other two, pre-dating CAADP, was explicit that agriculture is central to its overall strategy.1 Ghana's success has been largely restricted to the smallholder coffee areas. It is too early to judge the success of Rwanda. The Ethiopian success of 15 years of six to seven percent growth rate for cereals production is now well documented (Stuff, 2013).

Given the low level of achievement of the CAADP targets generally, this paper states how the African situation may differ from Asia, calling for modification of the strategy, but not a departure from the core of high output high input agriculture driven by biological science based technological change. The paper then provides the theoretical basis for the increasing rural income inequalities in high population density areas and the consequent dominance of the not poor small commercial farmer in agricultural growth. It then briefly explains how that growth is converted through the rural non-farm population into the rural poverty reduction that accompanies accelerated agricultural growth. Finally the paper draws on Asian and African experience to briefly summarize the critical public goods essential to high agricultural growth rates in high population density rural areas.

Intensification - green revolution - Africa is different

The green revolution had massive impact on agricultural growth rates in much of Asia, changed the public attitude towards agricultural growth, and with a few years lag greatly reduced poverty (Mellor and Desai, 1985). Six characteristics stand out as differentiating Africa from Asia and make it more expensive and perhaps time consuming for Africa to achieve the Asian results.

Greater diversity

In general, Asian countries were dominated, in production and population concentration, by irrigated rice and wheat. When Robert Chandler became the first Director-General of the International Rice Research Institute (IRRI) he stated a single objective - to double rice yields in Asia. With such a straight forward objective IRRI very quickly created the varieties that would do that. They spread rapidly, and were followed by a series of varieties adapted to an ever widening but still modest range of conditions. Doubling rice yields had massive national and regional impact. When CIMMYT in Mexico, the sister institution of IRRI, dedicated to maize and wheat production, nearly concurrently had the break-through in wheat, another very large area was covered.

Africa does not have a similar homogeneity of production conditions. The paper by Headey and Jayne (2014) in this volume notes that cereals have a much smaller share of total agricultural output in Africa than in Asia. It follows that for Africa to have as big an impact on food production there must be a much larger expenditure on research, and extension as well, to add up to a total comparable to Asia.

Of course, just as in Asia, research breakthroughs for particular commodities will cause greater specialization and decreased diver-

1 Ethiopia's Prime Minister Meles set a vision of middle income status by 2025 and a strategy of Agriculture Development Led Industrialization (ADLI). For a set of case studies on such a strategy see Mellor (1992). The agricultural development strategy was backed by an intensive analysis and quantification of the policy and investment requirements (PIF). The PIF was widely reviewed within both the Ethiopian government and the donor community before it was accepted by the government. The members of the drafting committee were Demese Chanyalew, Chairman, Berhanu, and John W. Mellor.

sity as farmers shift to the more profitable varieties and crops. Asia used to be much more diversified than it is now. For example in India the wheat crop was typically inter-planted with an oilseed crop (Mellor et al., 1968). Similarly, pulses soon largely disappeared from the irrigated areas, and imports of pulses from Africa increased. In the vast areas suitable to irrigated wheat the research breakthroughs radically increased its profitability relative to intercropped oil seed and to pulses.

Some deplore this loss of diversity and it does have a cost, but it was a major part of the green revolution results.2 With high productivity research systems, crop specific breakthroughs and diverse production conditions African countries will certainly specialize more than now. Nevertheless, ex-anti it is difficult to judge where to concentrate the research money and consequently the research expense will be higher than for Asia.

A common criticism of the CGIAR international research system in Asia is that it focused inadequately on soils and problems of fertilizer response relative to plant breeding. That criticism is far more apt in Africa (e.g. Tittonell and Giller, 2012), with its highly variable soils, very low organic matter, and some argument of low response to inorganic fertilizer, perhaps due to low organic matter.

Perennial export crops important

This of course should be a large plus factor for Africa. Perennial export crops are high value per hectare and per worker. Having a large base and potential to expand should be a major driving force in accelerated income and production growth. That has not been the case. For one, donors who control so much of the investment in African agriculture have emphasized food crops. In addition, perennial crops require more fixed investment than annual food crops and the financing institutions are lacking. More on that, below. It is argued that both inelastic demand and wide fluctuations in prices justify neglect of the tropical export crops. However, as below, Malaysia has done very well in palm oil, Kenya in tea and Ghana in cocoa. Coffee should suffer somewhat less on the price front, since the bulk of African coffee production is from high elevations with excellent quality for which demand has been growing rapidly.

Malaysia is dominated by two agro-ecological types - the poor upland soils and the rich river basins that also characterize coastal West Africa. Malaysia saw a huge potential in smallholder oil palm. It brought an improved variety from INEAC (located in the erstwhile Belgium Congo), the most advanced oil palm research station in Africa and the world, modelled its budding oil palm research system on INEAC, trained an oil palm extension system, and subsidized planting in smallholdings thereby assisting the relatively poor rural Malay people.3 Ignoring foreign aid negativism about inelastic demand, Malaysia soon became the number one producer of oil palm, which it still dominates. Shifting of the demand function with rapid growth in low and middle income countries has been far more important than its inelastic shape. Oil palm is now inconsequential in West and Central Africa. Nigeria's share of global oil palm exports dropped from 23 percent in 1961-63 to 0.6 percent in 1971-73 (Lele, 1991).

In the rich river basin soils Malaysia quickly took up the high yielding rice varieties that could pay for increased irrigation (Bell

2 For a broader discussion of the bio-diversity aspects of a green revolution see Mellor (2002).

3 The Chairman of the Malaysian Palm Oil Organization states that forty percent of

the area is held by smallholders (Bainow, 2013). I emphasize this because of the much greater impact of growth in the smallholder sector on poverty reduction through the

income and employment multiplier effect as documented in a later section as compared to large scale estates and because of the large effort by the Malaysian government to support smallholder production. This argues for support of the smallholder sector in Africa. The problem is not that the ex-colonial powers did not foster large scale plantations - it is the lack of support for smallholder production.

et al., 1982). Thus Malaysia recognized its diverse physical resources and concentrated on two crops each suitable to a specific agro-ecological niche. Malaysia gave no attention to annual food crops on the relatively poor upland soils.

The American foreign aid program (USAID) uses the rather peculiar phrase ''feed the future'' as the over view of its development assistance to agriculture. That places the emphasis on basic food staples, a position emulated by many other donors. Lele (1991) and Mellor (1989) document this broad tendency.4 Since adoption of foreign agricultural institutional structures has been central to catch-up growth in all low income countries the influence of foreign aid has been substantial. Most notably the emphasis on research both national and international has been on the food crops. Little emphasis is on reducing the cost of production of the tropical export commodities. This did not deter Malaysia from investing its own funds in oil Palm research - similar to Ethiopia putting its funds into foreign donor ignored teff (Minton et al., 2013) - with excellent results in both cases.

The donor community has backed annual horticulture export production and it has grown rapidly in a few countries. But its total is small compared to the perennial export commodities. There are other exceptions to the foreign under-emphasis on export commodities. The Kenya Tea Development Authority (KTDA) brought immense prosperity to large areas (and booming market towns) with effective foreign assistance (Lele, 1975).

Even in coffee the African institutional structures are much less well developed than in Latin America. For example, Ethiopia still has no oversight body for the sector, while Guatemala, in Anacafe, has a powerful body with major sources of revenue to finance a high level of research, coordinated extension, and quality control. Most Latin American countries have institutions similar to Anacafe, while the African analogues are mostly lobbying organizations with limited direct development effort.

Ghana is stated as a success story in agricultural growth. As noted, it has a clear plan for accelerating growth. However, the successful part of the strategy is cocoa production, for which the growth has been rapid, largely through expansion in the forest area, and led by attention from the cocoa research institute. The rapid growth in cocoa has been accompanied by rapid reduction in rural poverty, confined largely to the cocoa producing areas that experienced the rapid growth (Breisinger et al., 2011; Ghana Government of, 2007).

Less irrigation (but better rain fed)

As stated, widespread irrigation of rice and wheat, provided much more homogenous conditions for the spread of the IRRI and CIMMYT varieties in Asia. Africa is characterized by modest levels of irrigation, even in countries with large potentials. As in the previous section this barrier to the green revolution is more an example of missed opportunity than a permanent barrier. Grasping the opportunity does call for considerable catch up investment and institutional development.5 However there are three offsetting forces to the paucity of irrigated land.

4 The Lele citations on foreign aid, here and later in this paper, take a more nuanced approach to foreign aid than this paper. As Lele points out there has been significant foreign aid to agriculture in African countries. Sometimes that input was restrained by recipient country pressures (often preferring aid to urban industry and central infrastructure). However in Asian countries, at an earlier period, foreign aid to agriculture (with the exception Taiwan) fought an uphill battle against the heavy industry approach. My less nuanced approach comes from simple comparison of the results in Asia and Africa. The striking success of foreign assistance in building Asian institutions for agricultural growth has not been matched on that scale in Africa.

5 You et al. (2011) provide detailed estimates of irrigation potential - currently six percent of area with 17 percent as the potential - thereby to account for about one-third of total production. Realizing this potential will certainly not match the huge river basins and deltas of Asia but it is nevertheless enough for substantial aggregate impact. See also Xie et al. (2013).

First Africa has large areas with assured rainfall. In Ethiopia, the high potential, assured rainfall areas account for 70 percent of agricultural output (Table 1). With major research breakthroughs in those areas that percent would climb rapidly to more than 80 percent. Much of East, South, and coastal West Africa's agricultural production is in areas of reliable rainfall. It is of course the volume of production in the assured rainfall areas not the area that counts towards production potential. The assured rainfall areas have much higher yields initially than the less reliable areas and of course respond more to research breakthroughs. In Asia the foreign aid and research concentrated on the high potential areas. In Africa, there has been pressure in the name of income distribution and poverty reduction to concentrate more on the low potential areas (Lele, 1991; Mellor et al., 1987).6

Second, there is a large unrealized potential for irrigation (You et al., 2011; Xie et al., 2013). In Ethiopia, assured rainfall areas have sufficient rainfall to provide scope for small scale schemes to irrigate most of those areas. The Policy and Investment Framework (PIF) for the current 10 year plan allocated well over half of all investment to irrigation (Ethiopia, 2010). Irrigating the assured rainfall areas doubles the cropped area by allowing cropping in the dry season (initially Ethiopia is a one crop a year country) and of course occasional droughts do occur in the assured rainfall areas and that would be mitigated by irrigation. In West Africa the large river systems provide a large underdeveloped potential for irrigation. Substantial irrigation development has occurred in Mali, Senegal, and Nigeria, but certainly not at the level to provide the proportion developed comparable to Asian countries (You et al., 2011).

In West Africa, the excellent soil river basins in the high rainfall coastal areas should have year around cropping. The drier area to the north should have large irrigated areas. One reason that irrigation potentials have not been developed is the apparent high costs of irrigation development, especially compared to Asia and perhaps a view of unlimited land area for expansion (You et al., 2011). The paucity of research and low expenditure on critical public goods also reduces returns to irrigation.

Third, the large areas suitable to perennial export crops are not lagging for lack of irrigation. Most of them have adequate rainfall.

Poorer physical infrastructure

African countries are characterized by high input prices, notably fertilizer, and low output prices because of transport systems vastly inferior to those of Asia (Rashid et al., 2013). Thus, the cost of providing Asian infrastructure conditions starting now is very high. Once again Ethiopia is demonstrating what massive investment in rural infrastructure can do (Dercon et al., 2008; Minton et al., 2012).7 IFPRI studies show that in the past ten years the cost of diesel fuel in Ethiopia has gone up by 60 percent (reflection of worldwide trends) while the average cost of transportation has

6 The more productive areas initially have more poor per square mile but a lower proportion of poor than the low productivity areas - see the later discussion of the production functions in the respective areas.

7 The references to Ethiopia are copious. That is because it is unique in having from the Prime Minister a clear vision (reaching middle income status by 2025), an explicit strategy for reaching that goal in which agricultural is central - Agricultural Development Led Industrialization (ADLI), exceeding the CAADP target for government investment in agriculture by 50 percent, massive investments in rural infrastructure, since 1995, and consistent growth in agriculture of on the order of seven percent per year (Ethiopia, 2010; Stuff, 2013). The long standing IFPRI EDRI collaboration in Ethiopia provides an unusual wealth of data. Ethiopia had unusually low income even by African standards at the start. Rwanda and Ghana share much of this and are reported liberally in this paper, but they were later starters with consequently less to show in the short run. Finally Table 1 for Ethiopia is compared to the table elsewhere in this issue that shows that Ethiopia on the issue of population density is not an outlier compared to other African countries. The primary difference in Ethiopia is the temperate climate.

Table 1

Ethiopia: total area CULTIVATED by farm size and agro-ecology. Source: Calculations from the Agriculture Sample Survey of 2007/08, Central Statistical Agency (CSA).

Farm size (hectares) Moisture reliable cereal Moisture reliable enset Humid lowland Drought prone Pastoralist Total

(percentage of national total)

0.0-0.25 0.9% 1.1% 0.1% 0.6% 0.1% 2.7%

0.25-0.52 2.9% 2.4% 0.1% 2.2% 0.2% 7.9%

0.52-0.90 7.1% 2.9% 0.3% 3.8% 0.3% 14.4%

0.90-1.52 14.0% 2.7% 0.4% 6.7% 0.6% 24.3%

1.52-25.20 33.5% 2.2% 0.8% 13.1% 1.1% 50.7%

Total 58.6% 11.2% 1.6% 26.4% 2.3% 100.0%

aEach farm size interval (quintile) contains 20 percent of Ethiopia's small farms, approximately 2.57 million farms.

dropped in half (Minton et al., 2012). That anomalous result is because of a massive trunk and rural road building campaign with very large public expenditure (and an ancillary high rate of inflation!).

Rural electrification is rare indeed in Africa. The high agricultural growth areas of India, e.g. the Punjab, have universal village electrification. In the mid 1970s USAID made major pushes for increased rural electrification in the Philippines and Indonesia. More to the point, when the high yielding varieties struck in India there was little rural electrification, but political pressures brought it quickly in much of the green revolution irrigated area. Ethiopia has been pursuing rural electrification.

When Resources for the Future in collaboration with the National Electric Cooperative Association of the US put out a book of case studies they first elucidated how valuable electrification is to growth in agriculture. Then out of eight case studies, none are for an African country (Barnes, 2007). There were none to mention.

Poorer institutional structures

When the high yielding varieties hit Asia they were preceded by a long history of development of agricultural institutions and an immediately prior period of foreign aid assistance on a large scale to the key institutions of research, extension, and finance. Concurrently large foreign assistance went to higher agricultural education to train the large numbers of personnel to staff those institutions and a rapidly growing private sector as well. Thus, the institutional structure for quickly modifying and adapting applied research, extending knowledge to farmers, and financing inputs was well-established. The new high yielding varieties walked into a very hospitable institutional environment (Mellor et al., 1968).

In Africa, there was a similar, early period of foreign assistance to institutional development in a few countries e.g. Michigan state in Nigeria, Oklahoma in Ethiopia developing excellent agricultural universities and through them research and extension (Eicher, 1984). But by the time independence had been generalized and a green revolution could be thought of foreign aid was already turning away from agriculture, away from the non-poor small commercial farmer who accounts for the bulk of agricultural growth (see below), and most importantly away from national scale development of institutions by governments and towards small scale unintegrated private sector and NGO efforts with little national impact (Mellor, 1989; Mellor et al., 1987; Lele, 1991). The abandonment of assistance to higher agricultural education was most striking and most damaging (Eicher, 1984). As a result the institutional structures, particularly the central public sector component is still way short even now of the Asian standard when the green revolution struck.

Finance systems for the small commercial farmer are another a dramatic case. When the green revolution struck, essentially all Asian countries had specialized national systems for extending credit, and often also mobilizing deposits, to the small commercial farmers. Those systems moved large sums of money to those

farmers. Typically one-quarter of all working capital was financed from these systems (Desai and Mellor, 1993).

Because of rapid growth and consequent loose management systems and to some extent political interference these Asian systems, unlike their counterparts in developed countries several decades earlier, had high rates of overdue loans. These overdues commonly ran to 45 percent. This turned the donor community against these institutions, stating that they were non-viable because of the overdues (Desai and Mellor, 1993). The result was that the late developers, including a few laggard Asian countries such as Nepal, and essentially all African countries lacked these systems for financing the high cash costs of agricultural intensification. This turning away from financial institutions for the small commercial farmer was reinforced by the concern that the loans were largely to non-poor farmers.

Along came the Grameen Bank and micro credit to reach the poor and rapid growth of such institutions in Africa. They have been successful, investing large sums in the poor, more urban than rural however, and achieving high repayment rates. They were not suitable to the small commercial farmer however and those producers of output growth were left out (Desai and Mellor, 1993). Most African countries are still left with little coverage for the small commercial farmer. The most striking difference between finance systems for the poor and those for the small commercial farmer is how repayment is achieved. For the poor, group guarantees in the context of member savings drive the lending program. For the much larger sums per loan, lesser social cohesion, less connection between a group's savings and loans, repayment for the small commercial farmer must rely more on agricultural technology competent loan officers who ensure profitable loans and regular follow-up to ensure repayment on time (Desai and Mellor, 1993).

Political instability and corruption

In the early days of independence many African governments were stable for substantial periods of time. Many of those stable governments had the same leadership attitudes as for Nehru India and Sukarno Indonesia - favoring urban large scale industry over agriculture and strongly socialist views even for agriculture, and not understanding the nature of scale in agriculture (Lele, 1975; Mellor et al., 1987) In Asia however foreign aid pushed hard for emphasis on agriculture, substantially successfully so and recognized the key role of public goods to small holder agricultural growth Mellor et al., 1968). In Africa foreign aid was often supportive of the more loony ideas, as in the Ujama movement pressed by President Nyrere in Tanzania, while it had worked strongly against such schemes (large scale socialized farms) from Nehru in India (Lele, 1975, 1991).

Many governments became less stable after the initial cases of post-independence longevity - e.g. Liberia, Cote D'Ivoire, and Zimbabwe. Corruption erupted. Between the two it is difficult to have a consistent steady growth strategy and to provide the essential public goods for rapid agricultural growth. That in turn has strengthened the private sector orientation of foreign aid. Kenya

J.W. Mellor/Food Policy xxx (2014) xxx-xxx

was for a decade or more an example of an African country that was doing well in agricultural growth, largely based on the small commercial farmer (Lele, 1975; Mellor et al., 1987; Johnston and Kilby, 1975). With deterioration in government the key public-sector institutions declined in capacity. That has also happened in India with a major impact in slowing the agricultural growth rate and hence the poverty reduction rate (Desai et al., 2011; Tamboli and Nene, 2011).

Implications of Africa's distinctive characteristics

The salient implications for future policy and investment of the green revolution related differences between Asia and Africa are five. First, agricultural research and extension and by derivation higher education training of personnel must be vastly larger and more expensive with somewhat lower rates of return. They are now grossly underfinanced relative to the diverse needs. Second, the perennial tropical export commodities, including oil palm, cocoa, coffee and tea, must realize their large aggregate potential for growth, in the geographic areas in which they have comparative advantage, with provision of large additional funding and specialized institutional structures for coordination of the interrelated research, extension, financing and quality enhancement that characterize the big success stories in these commodities. Third, government leadership must ensure filling the financing void with national, specialized agricultural finance systems that lend large sums based on the profit and loss statement not the balance sheet to the small commercial farmer. Mobilizing deposits should be an integral part of that effort. Fifth, massive additional investment in rural roads, electrification, and irrigation is required. The bulk of the financial requirement is for the rural infrastructure. Much of that is also essential to results from the social expenditure for what will for several decades be a growing rural population.

The emphasis here is on public goods, as is the case for CAADP. The private sector is of course dominant in all the successful agricultural growth cases, but its growth depends on agricultural production growing and the public goods are a necessary, and now under emphasized element. All these elements cannot be done at once at the required scale. Setting priorities and leaving some less attended may well give the targeted growth for a few years, by which time the laggard areas can be brought up to give a final set of pushes to the growth rate. For example, Ethiopia has neglected agricultural finance and the coffee sector - but both are now receiving special attention and may thus play a role in sustaining the high growth rate.

Land and income distribution in high population density areas

To understand the requirements for agricultural growth and poverty reduction in high population density areas it is necessary to first understand the land and income distribution that typifies such areas and how they differ from low population density areas. We know that agricultural growth is associated with rapid decline in poverty (Ahluwalia, 1978; Timmer, 1997; Ravallion and Datt 2002; Thirtle, 2001) but as shown below the causal relations are complex in high population density areas.

The contrast in production functions that typify high and low population density areas are discussed first, followed by analysis of the dichotomy of population groups that dominates high population density areas.

Contrasting production functions

Simplistically in low population density areas the distribution of income is more even than in high population density areas. Typ-

ically most families are poor, land operating units are of similar size and the gini coefficient relatively low. Normal, in areas of low population density, is for the land to be relatively low productivity and to have a gently upward sloping production function for output from labor input.8 That results in little surplus over subsistence and hence little room for some farms to grow with hired labor and to generate the income for the goods and services from a rural non-farm sector.9 The much higher cost of physical infrastructure per family in low population density areas reinforces these tendencies and further inhibits growth.

In areas of high population density the land is by definition productive (it supports a dense population) with a more steeply sloping labor production function and hence a surplus of production beyond that required to support the labor force. That facilitates capital accumulation by some, accretion of land, some increasing their net income by hiring an additional laborer, and the conditions for a landless or near landless class.10 It is the source of the dichotomy between the small commercial farmer and the rural non-farm family discussed below. Income disparities within such regions are substantial.11

The small commercial farmer

The small commercial farmer class includes, at the lower end of the scale those with sufficient land to generate a family income just above the poverty level. They are commercial farmers, selling sufficient output to pay for the non-food items in their consumption basket. They live in and are part of the local community. They have consumption patterns that depend heavily on locally produced non-tradable, labor intensive goods and services. The definition excludes the absentee landowner with urban oriented consumption patterns that are import and capital intensive.

Typically across much of high population density Africa with traditional farming practices and cereal crop production it requires on the order of one hectare of land to provide sufficient income to reach the poverty line. If the lower limit is decreased to include those with landholding large enough to be above the poverty line with modern high yield crop varieties, the lower limit would drop to about 0.75 hectares. The upper limit would typically be on the order of 25 hectares, but rising further in some areas. The paper by MJ provides detailed data for several countries.

If we apply those rules to Ethiopia the small commercial farmer would be the two size classes from 0.90 hectares to 25.2 hectares plus half of the size class from 0.52 to 0.90 (Table 1). That is effectively a top of five hectares since, due to an earlier land reform, there are very few farmers with more than five hectares. The large commercial farmers are excluded from that calculus. They occupy two percent of the farmed land.

8 Of course, there are exceptions. Disease such as malaria (in the terai of India and Nepal, and many parts of Africa) and those associated with Tsetse fly, or massive loss of life from wars may stop the population growth that results in areas of productive land becoming high population density areas. When those forces are removed it will be a matter of a few generations before they become high population density areas.

9 An early article (Mellor and Stevens, 1956) provides a rigorous presentation of this point, with Asian examples in mind. The distinction is more important in the African context. See also Binswanger and McIntyre (1987).

10 Masters et al. (2013) show the large difference in size of farm between the lower and upper quartiles and the inadequacy of the farm size in the lower quartile to provide even a poverty level of income. See also Jayne et al. (2010) and Headey and Jayne (2014), this issue.

11 Both Timmer and Ravallion and Datt draw attention to the poverty reduction

impact of agricultural growth reduced by highly unequal distribution of land and

income. However they are referring not to the differences between the small commercial farmer and the landless and near landless, but between large scale, often absentee land owners who are not part of the consumption patterns of the rural community e.g. the haciendas of Latin America.

J.W. Mellor/Food Policy xxx (2014) xxx-xxx

In Ethiopia the small commercial farmers have 82 percent of the land and with somewhat higher yields over 85 percent of output. There are several million such farm families. Reaching the bulk of them is not a small task. These farmers determine whether the government's targeted agricultural growth rates are reached or not. They are not poor - perhaps typically twice the average income in their community. They make their living by farming, sell a minimum of 30 percent of output, and often up to 80 percent, and are focused on farming to increase their incomes.

Jayne et al. (2010) presents data for average size of farm by farm size quartile for five countries - Kenya, Ethiopia, Rwanda, Mozambique, and Zambia. Two of the countries are usually seen as land scarce, but two are usually seen as land abundant and the third as in between. The data for Ethiopia (Table 1) are consistent with the five country data in Jayne - 81 percent compared to 82 percent of the land in the top half of the distribution. The average size in the third quartile is surprisingly similar across the countries. In the top quartile Kenya and Zambia stand out for larger size probably reflecting a higher proportion of very large holdings.

There has been a concern that some areas in Africa, e.g. parts of Kenya, have such high population densities that they cannot drive an increased yield based growth path. The test of that hypothesis are two. First, are the yields in such areas already at the level of the advanced Asian countries and the high income countries? If they are the potential for rapid catch-up growth is lost. Alternatively have efforts to raise yields with poor technology resulted in irreversible damage to soils, e.g. massive, irreversible erosion (but not low organic levels that can be corrected). Second, do these areas represent a substantial portion of the nation's agricultural production. For Ethiopia, a high population density nation, the consistent six to seven percent cereals production growth rate shows the answer to both questions is no. For Kenya the answer to the second question is clearly no.

The rural non-farm population

The rural non-farm population is comprised of those with insufficient land to reach the poverty line from farming. They depend on off-farm income for survival.12 They typically comprise about half of the rural families. Of course there may be high income families in trade or professions that do not own land, but they will be few. The bulk of the rural non-farm population is poor and in high population density areas essentially all the poor fall in this category. In the data cited from Jayne et. al. about one-quarter of the rural population in each of those countries is virtually landless as in Table 1 for Ethiopia.

The rural non-farm population derives its, earned, non-farm income from producing non-tradable labor intensive goods and services. What they produce is not salable in large urban centers. That has two implications.

First, they are dependent on local demand and hence on rising local incomes if their incomes are to rise. The source of rising local incomes is agricultural growth from the small commercial farmer.13

Second, since labor is the primary input for what they produce the supply is as elastic as the labor supply. If the demand increases the supply increases with little increase in price. Because the demand is income elastic for the non-tradables, growth in incomes increases the demand and size of this sector more than proportionately to the income increase. As growth gets underway that under-

12 In Jayne et al. (2003) it is noted that in Africa there is commonly a direct relation between size of farm and off farm income. However Ethiopia fits the Asian inverse relation. Whatever that relation, those with holdings too small to provide the poverty level of living require off-farm income in order to survive. In a slow growth agriculture there may be substantial underemployed labor and endemic poverty.

13 The nature of the expenditure patterns of the small commercial farmer are quantified by Delgado et al. (1998), Hazell and Roell (1983), Hazell and Ramasamy (1991), Bell et al. (1982).

employed labor force will be gradually taken up leading eventually to rising wage rates. That is as agricultural income increases the share of rural non-farm income increases. That seems to be roughly consistent with the data in table X of Jayne et al. this publication. Data from Egypt show that the rural non-farm population increases as middle income status is achieved (Gavian et al., 2002, also Mead and Liedholm, 1998 and Liedholm and Meade 1987).

In a closed economy, a high rate of growth of agricultural production will reduce the price of food, passing the benefit to the poor who spend on the order of 80 percent of their income on food (see the modeling on this by Lele and Mellor, 1981). With open economies becoming the rule rather than the exception this means of passing agricultural growth benefits to the poor becomes less important. In addition using prices to benefit the poor deters production growth since the income of the small commercial farmer will be reduced and there employment creating expenditure is reduced as well.

What is the meaning of a large share of the rural population as rural non-farm (insufficient own farm income for survival) and a small share of rural non-farm income? It means that those with inadequate land area for reaching the poverty line are exceedingly poor, with considerable underemployed labor. Off farm income may be more from remittances from migration and public sector employment, including work programs than from expenditures by the local small commercial farmer. The small commercial farmers may also be quite poor and so provide only modest expenditure on the rural non-farm sector. As the incomes of the small commercial farmers rise they spend increasing portions on the local non-tradables. Agricultural income growth will gradually accelerate growth of income in the rural non-farm sector.

Thus we have a dichotomy of those who produce the bulk of agricultural output and are not poor and the bulk of the poor having insufficient land to produce even the minimum poverty level of food. The link that reduces poverty is through the expenditure patters of the small commercial farmer on the rural non-farm sectors locally produced, labor intensive, non-tradable goods and services.14

Prior to the green revolution in Asia it was generally assumed that increasing agricultural production would reduce poverty. As the green revolution progressed it was soon realized that those producing the higher incomes were not poor and the poor were not directly involved in the green revolution. From that came an anti-green revolution literature that was particularly influential as attention shifted to Africa.15 That literature led to an emphasis on the poor resource areas of low population density where poverty was more evenly distributed and questioning the efficacy of the green revolution. That position was further strengthened by the environmental movement and its common emphasis on low input agriculture. The result was emphasis on geographic areas with poor potential for agricultural growth and consequently little reduction in poverty. What was missed was the lags between agricultural growth

14 There is a long history in detailing how this works. Johnston and Mellor (1961) stated increased expenditure by prospering farmers as an important role of agriculture in economic growth. That concept was spelled out in an overall growth context in Mellor and Lele (1973), and Mellor (1976,1992). Hirschman (1958) stated that the linkages of agricultural growth with other sectors were very weak, but he focused only on production linkages, whereas the preceding was based largely on consumption linkages. Johnston and Kilby (1975) however show that the production linkages are also more powerful than assumed by Hirschman. The relationship has been modeled by Dorosh and Mellor (2013) in the context of underemployed rural labor, by Mellor and Ranade (2006) for a neo-classical economy, by Haggblade et al. (1991), and by Mellor and Gavian (1999) for a middle income country, showing that as incomes rise the relative importance of the employment effect increases relative to the GDP effect.

15 Representative of this literature are Griffin (1975), and Harris (1982). They cite the full range of this point of view. Lipton and Longhurst (1989) provides a nuanced critique of the Green Revolution in which he is clear that the biological breakthroughs are essential to relieving poverty, but found much more needed as well. See also Mellor and Desai (1985)

and the impact of expenditure in the rural non-farm economy (both Timmer and Ravallion and Datt noted these lags). In time the poverty reduction occurred.

The government of Ethiopia has emphasized the high potential assured rainfall areas for their big push on agricultural growth. With that success has come dramatic reductions in rural poverty (Ethiopia, 2010).16 In effect that has been the case in the cocoa areas of Ghana as well.

In the low population density areas with the gently sloping response curve to labor input fostering agricultural growth directly reduces poverty - but in those areas that growth is very difficult and costly to achieve. The conceptual problem arises with the high population density areas. In low income countries the density of poverty is very high in those areas but the agricultural growth is in the hands of the non-poor small commercial farmer. Raising their income does not directly reduce poverty. That is why poverty oriented personnel in foreign aid agencies may push for emphasis on the low population density poor resource areas, with generally poor results.

Food security for the poor

The previous sections show how rising incomes of the small commercial farmer in high population density areas converts into increased employment and hence decreased poverty for those below the poverty line. The further conversion into increased food security follows. The small commercial farmer and urban populations on average have inelastic demand for cereals, the basic calorie source, of around 0.3, while the rural non-farm poor have an elasticity of around 0.8 (Tafere et al., 2010; Bouis, 1999) With the small commercial farmer spending on the order of half of income increments on the rural non-farm sector the increased consumption of basic food staples is large, keeping a major share of increments in production in the rural areas and consumed by the poor. As for poverty reduction, the road to food security passes through the small commercial farmer in high population density areas.

Accelerating the agricultural growth rate

Binswanger and Pingali (1988), in discussing agricultural growth in Africa, note two quite different sources of growth: (a) increased traditional inputs; and, (b) application of new technology and its associated inputs.

Growth through traditional inputs

Ethiopia, a clear example of high population density agriculture, has maintained a seven-percent growth rate since the mid 1990s.17 The growth rate up to a least 2005 cannot be explained by the growth

16 Although many factor affect poverty levels, the major reduction in rural head count poverty (using the World Bank measures and definitions from 47.5 percent in 1995/96 to30.4 percent in 2011/12 with acceleration in the more recent five year period, is broadly consistent with the cereals growth rate of seven percent and the similar growth rate for overall agricultural production (Ethiopia, 2012).

17 The source of the basic data is the National Statistical Agency, a highly

professional agency with a well-trained, technocratic, committed head that publishes its methodology which in turn has been carefully scrutinized. Despite this unusually professional operation there is a committed set of opponents of the Ethiopian strategy (Agricultural Development Led Industrialization) that finds this high growth rate in agriculture inconsistent with its push for industrialization first. The Policy and Investment Framework (PIF) (Ethiopia, government of, 2010) analyzes these data in detail, with emphasis on dealing with large year to year fluctuations and supports the conclusion of a seven percent growth rate. Some estimates place the growth rate much higher, but they are clearly biased by choice of the period and the placement of the high and low fluctuations. Stuff (2013) provides a highly sophisticated analysis of these trends using traditional linear regressions and the Sen method. That analysis also shows internal consistency of the data. Note that the poverty data for Ethiopia shows declines fully consistent with the agricultural growth rate (Ethiopia, 2012).

in modern inputs (improved seeds and fertilizer). The base and the growth rates were initially too small for major aggregate impact (Ethiopia, Government of, 2010). Half the growth is explained by increased land area. This is surprising given the initially high population densities and very low incomes. In India, noted for its high rural population density in the productive areas, in the first, second, and third decades after independence agricultural growth accelerated substantially with increased land area accounting for half the growth rate (Mellor et al., 1968) - as in the recent history in Ethiopia. In both cases it is also likely that increased labor input, perhaps on labor intensive traditional practices, played a significant role. A small role was of course also played by growth in the modern inputs even from a low base.

Why did the Ethiopian growth rate so greatly exceed any previous period and that in other somewhat similar circumstance in other African countries? The only new institutional variable was a massive expansion in the extension service. In five years, 63,000 extension agents received training and were located, six in each Kebela, in the smallest administrative unit. That put them in walking distance of most farmers. There are about 400 small commercial farmers in each Kebela. That is also the size unit seen as optimal for agricultural finance branches at that stage of development (Desai and Mellor, 1993). Those extension agents were undoubtedly poorly trained, but the IFPRI surveys consistently show, corroborated by Focus Groups in the USAID AMDe project, that the small commercial farmer has a favorable opinion of them. They were helpful.

The massive size of the extension service provided close contact with farmers. Presumably they publicized methods of increasing production without modern inputs, began the process of increasing use of modern inputs, and energized farmers to increase production. The whole thrust of the ADLI strategy and its massive publicity perhaps energized farmers. It is notable that the rapid agricultural growth in Ghana also has to be explained largely without modern technology. In the case of Ghana farmers were energized to expand the cocoa area substantially thereby increasing incomes and driving down poverty (Ghana, 2007; Breisinger et al., 2011). Of course these means of production increase must eventually run out. Then, modernization must occur.

Application of new technology

In Ethiopia in the early to mid-noughts the impact of the traditional sources of growth was declining (Ethiopia, 2010) and so modern technology was to take over. That may also to be the case for Rwanda and Ghana at present. That creates immense problems of institutional development. Asian countries demonstrate the need for public goods and the servicing institutional structures. Generally the seeds of that growth are present in Africa. Most African countries have a public sector research system and an extension system. A specialized agricultural finance system is generally lacking. The private sector is ready for a wide range of activities.

Research/extension

The point was made earlier that the greater diversity of production systems and physical resources requires a substantially larger research system than for Asian countries to achieve comparable levels of growth. In contrast the African systems are generally smaller and certainly have received less foreign assistance in sharp contrast to the large scale development and close relations with lots of money between US land grant institutions and those of Asian countries, and a large effort by the Rockefeller Foundation.18

18 Of course, African countries should compensate for poor donor priorities. Ethiopia put national resources into Teff research, by value the most important cereal, that is ignored by international support (both in research and in value chain projects) and has had a major new variety with rapid uptake (Minton et al., 2013). Malaysia did likewise for oil palm research.

J.W. Mellor/Food Policy xxx (2014) xxx-xxx

Of course there has been some history of that in Africa but it ended far too prematurely and touched only a few countries.

Finance

Finance is essential and a huge problem. In Asian countries about one-quarter of farm working capital is financed by credit (Desai and Mellor, 1993). Small commercial farmers consistently report a need for credit even though they are not poor. Commercial banks do not open the numerous branches required to reach the small commercial farmer. Micro credit operates on entirely different principles for making loans and timing repayment. Micro credit rarely reaches the small commercial farmer. Thus financial systems with the broad outlines of those in Asia are needed. It is arguable that enough is now known of the errors of the Asian systems that the African ones can manage high repayment rates. There are examples of institutional credit to small commercial farmers that have close to 100 percent repayment on time (Desai and Mellor, 1993).

Fertilizer

All Asian countries have achieved high levels of fertilizer use. The countries most explicit in following CAADP place a major emphasis on increased fertilizer use. They refer to the Africa Fertilizer summit, Abuja declaration and its central role of massive increase in fertilizer use as a means towards the ''African Green Revolution'' (African Union, 2006). Ethiopia targets a 15 percent growth rate in fertilizer use (Ethiopia, PIF, 2010) and has increased use at an eight percent rate (author's calculations from Agricultural Inputs Enterprise date). Rwanda started at a very low base of 0.3 kilograms per hectare which means a tiny proportion of farmers were using fertilizer. By 2011 it had increased the proportion of farmers using fertilizer to somewhat less than 50 percent (Rwanda, 2012, 2007). That would be a much higher percent of the small commercial farmers. Desai (2002,1982) discusses the policy needs for achieving those high rates of growth. Kelly et al. (2003) raise a major set of issues relating to the difficult response coefficient situation.

Despite these high targets for fertilizer use in Ethiopia, Ghana and Rwanda, there remains in Africa considerable debate about the soil constraints to high levels of inorganic fertilizer use and the complementarity with organic sources (e.g. see the complex discussion in Dreschel et al. (2001) and in Kelly et al. (2003)) There is no room for debate as to whether large off take of nutrients accompanying high yields require large inputs of inorganic fertilizer. But there is controversy about how or even whether the response to that large input will be profitable or even possible (Dreschel et al., 2001). Having said that Rwanda and Ethiopia are moving quickly with little evidence of those problems.

Attention was drawn above to the need for relatively greater emphasis on soil science research in Africa compared to Asia. In Ethiopia, under use and modest rates of return are easily explained by poor information transmission about optimal fertilizer practices (Ethiopia, 2010). It is not clear why that is not true for the rest of Eastern and Southern Africa. Similarly it is not clear why the Savannah areas of Africa are so different to similar temperature and rainfall regimes in central India. It is understandable that the soil science problems are difficult for the tropical uplands of coastal West Africa, but that is why they should be, as in Malaysia, largely in tropical perennial crops. But the bottom line is relative to its problems African countries are way under spending on soil science research, as are the international centers.

Several African countries now subsidize fertilizer at high rates (Rashid et al., 2013). It is explicit in the Rwandan plan. Ethiopia has abjured direct subsidy. The World Bank has alternate periods of pressing for subsidies on fertilizer and periods of being opposed - then at the same time it may be opposed (in India, where it was

in favor) and favored (in a much of contemporary Africa, where it was opposed). That sentence catches the confusion! The grand lesson from Asia is that fertilizer subsidies when use is very low do not cost the government much and do increase the incentive for use, perhaps reducing perceived risk and thereby speeding uptake. By plan that leads to a high level of use at which point the subsidies become very expensive and in fact drain funds away from roads and other high return rural investments. Failing a clear exit strategy it is best not to use fertilizer subsidies.

An exit strategy is simple in concept - tie it to high transportation costs and when those costs come down reduce the subsidy. In practice it never seems to work that way. Ethiopia does not subsidize fertilizer, and has increased use to nearly 600,000 tons per year and growing.19

Physical infrastructure

A high proportion of the huge increase in finance for agricultural growth will be spent on the physical infrastructure of roads, electrification and irrigation. The impact of improved roads on farm level price relationships has already been noted. That is also a principle means of encouraging growth of the rural non-farm sector in response to rising local demand.20

Irrigation was covered in the earlier exposition of contrasts between Asia and Africa. That exposition showed substantial gaps in knowledge even while the total potential is well documented. Thus, one is back to the need for far larger research systems than now exist.

Higher agricultural education

All the critical public institutions and many of the private ones as well are huge demanders of personnel with higher education in agriculture. Ethiopia was the beneficiary of the early US efforts to build agricultural Universities. The one the US built is excellent and has adapted to changing times while the government has radically expanded additional agricultural universities. The capacity will be strained when the current effort to build a national agricultural finance system moves into high gear with huge requirements for trained agricultural loan officers. Nigeria was also a major beneficiary of that era of foreign aid. Most other African countries have not been so fortunate. Of the many errors of American foreign aid in Africa perhaps far and away the greatest was exiting building higher education in agriculture.

It is notable that India built an excellent system of agricultural universities with US assistance - typically India wide institutions with well trained staff. That growth was associated with growth in agricultural production. In recent years those institutions along with research and extension have been deteriorating - increasingly parochial, local staff, usually of the dominant caste - and as that

19 Ethiopia has a significantly lower price of fertilizer at the farm level than the rest of East Africa. There is an argument (Rashid et al., 2013) that Ethiopia does subsidize fertilizer. However that is a convoluted argument. A significant part of the ''subsidy'' is the cost of hugely excessive carryover stocks. It is difficult to argue that the cost of a failed government policy should be placed on farmers. Similarly, the ''subsidies'' documented by IFPRI are largely under-payment to the cooperatives for their services and low interest rates to the cooperatives. Focus Group Studies for the USAID AMDe project show that private grain traders would like to enter the fertilizer trade once again and find the existing margins fully profitable. These are quite different to the direct subsidies in other African countries or in India. In any case adding those costs totaling 14 percent of the price, Ethiopia's prices to farmers are still lower than in the other East African countries.

20 Hymer and Resnick (1969) in a paper that predated the analysis of the rural nonfarm sector argued that roads resulted in rapid decline of the rural non-farm sector, but derived from a view of the rural nonfarm sector as producing goods and services that were protected from urban competition by poor infrastructure. We now know that the rural non-farm sector adapts to shifting prices and concentrates on non-tradable goods and services desired by the rising incomes of the small commercial farmer. As cited earlier, a middle income country like Egypt has a much larger rural non-farm sector than lower income Ethiopia.

Table 2

Ethiopia: hypothetical commodity composition of 6 percent growth rate.

Commodity group Base, percent Growth rate Share of growth

Cereals 32 5 26.7

Livestock 33 7 38.6

Coffee 17 8 22.7

Other 18 4 12.0

Total 100 6 100

Data from National Statistics Agency.

has happened the agriculture growth rate has declined (Tamboli and Nene, 2011; Desai et al., 2011). The exit of US foreign aid from this keystone for agricultural growth was far too premature. The argument here and previously is not that foreign aid has completely abdicated from these fields, but that in Africa where the need is far greater far less is done than in earlier decades in Asia.

Commodity priorities

It was noted earlier that greater diversity in African agriculture requires greater expenditure on a more commodity specific set of institutions. Table 2 presents for Ethiopia calculations of hypothetical sector-wise distribution of growth rates. Achieving even the CAADP target of a six percent growth rate requires substantial increase in growth rates in each sector. Because of the high, now unrealized, potential for growth in the export perennial (coffee) its share of incremental output is almost as large as cereals. But note also that it is still only one-quarter of the incremental output. To emphasize coffee does not call for neglect of cereals. Most other African countries demonstrate a similar diversity.

Conclusion

Africa must increasingly obtain its agricultural growth from increased value of output per hectare. The roughly half of rural families in the high population density areas that are small commercial farmers dominate agricultural production growth. As in Asia that growth will come from new crop varieties that greatly increase output per hectare. More output of course requires more input, the largest component of which will be plant nutrients - fertilizer. Increased access to water will become more important.

To achieve growth on national scales requires a high density of all-weather rural roads and rural electrification. These are also essential to education and health of the rural populations that will for several decades continue to grow in absolute size even in the face of rapid urbanization.

Large scale development of public institutions of research, extension, and at least initially, finance, is required along with the institutions of higher learning required for staffing these institutions and the rapidly growing private sector servicing small commercial farmers.

The paucity of all these investments and institutional structures requires a full commitment to agricultural growth, priority to the high potential geographic areas and to the small commercial farmer. Because of the greater diversity of agriculture in Africa compared to Asia these investments must be much larger relative to the size of agriculture than was the case in Asia.

Accelerated growth of production by the small commercial farmer will provide large scale expenditure on the employment intensive, non-tradable rural non-farm sector providing the large declines in poverty associated with rapid growth of small-holder agriculture. Food security for the poor will follow directly from that.

References

African Union, 2006, Africa Fertilizer Summit - Abuja Declaration on Fertiliser for the African Green Revolution.

African Union, 2010, The Comprehensive Agricultural Development Plan, Addis Ababa.

Ahluwalia, M.S., 1978. Rural poverty and agricultural performance in India. J. Dev. Stud. 14, 298-323.

Bainow, Yusuf, 2013. Malaysian Palm Oil, Y.soe@mpoo.org.my.

Barnes, Douglas, 2007. The Challenge of Rural Electrification, Strategies for Developing Countries. Resources for the Future, Washington.

Bell, C.L., Hazell, P.B., Slade, R., 1982. Project Evaluation in Regional Perspective: A Study of an Irrigation Project in Northwest Malaysia. Johns Hopkins University Press, Baltimore.

Binswanger, Hans and Prabhu Pingali, 1988, Technological Priorities for Farming in Sub-Saharan Africa, World Bank Research Observer.

Binswanger, H.P., McIntire, J., 1987. Behavioral and material determinants of production relations in land-abundant tropical agriculture. Econ. Dev. Cult. Change 36, 73-99.

Bouis, H., 1999. Patterns of Food Consumption and Nutrition in Egypt. IFPRI FSR-APRP, Ministry of Trade and Supply. Government of Egypt.

Breisinger, Clemens et al., 2011. A New Era of Transformation in Ghana. International Food Policy Research Institute, Washington.

Delgado, C., Hopkins, C., Kelly, V., 1998. Agricultural Growth Linkages in Sub-Saharan Africa. International Food Policy Research Institute, Washington, DC, Research Report No. 107.

Dercon, S., Gilligan, D.O., Hoddinott, J., Woldehanna, T., 2008. The Impact of Roads and Agricultural Extension on Crop Income, Consumption and Poverty in Fifteen Ethiopian Villages. IFPRI Discussion Paper 00840. International Food Policy Research Institute, Washington DC, USA.

Desai, G.M., 1982. Sustaining Rapid Growth in India's Fertilizer Consumption: A Perspective Based on Composition of Use. International Food Policy Research Institute (IFPRI) Research Report 031. Washington, DC.

Desai, G.M., 2002, Key Issues in Achieving Sustainable Rapid Growth in Fertilizer Use in Rwanda, Agricultural Policy Report 16, Abt Associates, Washington.

Desai, Bhupat M., Mellor, John W., Sarma, Vijay Pail, Tamboli, Prabhakar, Mellor, John W., 1993. Institutional Finance for Agricultural Development, an Analytical Survey of Critical Issues, Food Policy Review 1. International Food Policy Research Institute, Washington.

Desai, Bhupat M., D'Souza, Errol, Mellor, John W., Vijay Pail Sarma, Prabhakar Tamboli, 2011. Agricultural Policy Strategy and the Road Ahead, Economic and Political Weekly, December 2011.

Dorosh, Paul A., Mellor, John W., 2013.Why Agriculture Remains a Viable Means of Poverty Reduction in Sub-Saharan Africa: The Case of Ethiopia, Development Policy Review, summer 2013.

Dreschel, Pay, Gyiele, Lucy, Kunze, Dagmar, Cofie, Olufunke, 2001, Population Density, Soil Nutrition Depletion and Economic Growth, Ecological Economics.

Eicher, 1984. Facing up to Africa's food crisis. In: Eicher, Carl, Staatz, John M. (Eds.), Agricultural Development in the Third World. Johns Hopkins University Press, Baltimore.

Ethiopia, Government of, 2009. Comprehensive Africa Agriculture Development Plan (CAADP), Ethiopia Study, Main Report, Ministry of Agriculture, Addis Ababa.

Ethiopia, Government of, 2010, Policy and Investment Framework (PIF), Main Report, Ministry of Agriculture.

Ethiopia, Government of, 2012, Ethiopia's Progress Towards Eradicating Poverty -an Interim Report on Poverty Analysis Status, 2010/11.Addis Ababa.

Gavian, S., El-Meehy, T., Bulbul, L., Ender, G., 2002. The Importance of Agricultural Growth to SME Development and Rural Employment in Egypt. MVE Unit, APRP, Special Study No. 5, Abt Associates Bethesda, Maryland.

Ghana, Government of, Institute of Statistical, Social, and Economic Research, 2007, The Structure of the Ghanaian Economy 2006.

Ghana, Government of, Institute of Statistical, Social, and Economic Research, 2010, Medium Term Agricultural Sector Investment Plan (MATASIF).

Griffin, K., 1975. The Political Economy of Agrarian Change. Macmillan, New York.

Haggblade, S., Hammer, J., Hazell, P.B.R., 1991. Modelling agricultural growth multipliers. Am. J. Agric. Econ. 73 (2), 361-374.

Harris, John., 1982. Capitalism and Peasant Farming. Oxford University Press, Bombay.

Hazell, P.B.R., Ramasamy, C., 1991. The Green Revolution Reconsidered: The Impact of High Yielding Varieties in South India. John Hopkins University Press, Baltimore.

Hazell, P.B.R., Roell, A., 1983. Rural Growth Linkages: Household Expenditures Patterns in Malaysia and Nigeria. Research Report No. 41, International Food Policy Research Institute, Washington, DC.

Heady, Derek, Jayne, T.S., 2014, Adaptation to Land Constraints - Is Africa Different? Food Policy, this issue.

Hirschman, A.O., 1958. A Strategy of Economic Development. Yale University Press, New Haven.

Hymer, S., Resnick, S., 1969. A Model of an Agrarian Economy with Non-Agricultural Activities, American Economic Review, 59 1958, A, 493-506.

Jayne, T.S., et al., 2003. Smallholder Income and Distribution in Africa - Importance to Poverty Reduction Strategy, Food Policy.

Jayne, T.S., David, M., Edward, M., 2010. Smallholder Farming under Increasingly Difficult Circumstances: Policy and Public Investment Priorities for Africa, World Development, October.

Johnston, B.F., Kilby, P., 1975. Agriculture and Structural Transformation: Economic Strategies in Late-Developing Countries. Oxford University Press (New York).

Johnston, B.F., Mellor, J.W., 1961. The Role of Agriculture in Economic Development. Am. Econ. Rev. 51, 566-593.

Kelly, Valerie, Crawford, Eric, Jayne, T.S., 2003. Agricultural Input Use and Development in Africa. Michigan State University, Policy Synthesis.

Lele, U.J., 1975. The Design of Rural Development, Lessons from Africa. Johns Hopkins University Press, Baltimore.

Lele, U.J., 1991. Aid to African Agriculture - Lessons from Two Decades of Donors' Experience. John ' Hopkins University Press, Baltimore.

Lele, U.J., Mellor, J.W., 1981. Technological change, distribution bias and labour transfer in a two-sector economy. Oxford Econ. Papers 33, 426-441.

Liedholm, C., Meade, D., 1987. Small Scale Industries in Developing Countries: Empirical Evidence and Policy Implications. Small Scale Industries in Developing Countries: Empirical Evidence and Policy Implications, East Lansing.

Lipton, Michael, Longhurst, Richard, 1989. New Seeds and Poor People. Unwin Hyman, London.

Masters, William Agnes, Djurfeldt, Cornelius, de Haan, Peter, Hazell, Thomas Jayne, Jirstrom, Magus, Reardon, Thomas., 2013. Urbanization and Farm Size in Asia and Africa, Implications for Food Security and Agricultural Research, Global Food Security. Elsevier.

Mead, D., Liedholm, C., 1998. The dynamics of micro and small enterprises in developing countries. World Dev. 26, 61-74.

Mellor, John, 1976. The New Economics of Growth. Cornell University Press, Ithaca, NY.

Mellor, John, 1989. Agricultural Development in the Third World: The Food, Poverty, Aid, Trade Nexus. Choices, vol. 4 (No. 1, First Quarter 1989). pp. 4-8.

Mellor, John, 1992. Agriculture on the Road to Industrialization. Johns Hopkins University Press, Baltimore.

Mellor, John, 2002. Poverty Reduction and Biodiversity Conservation: The Complex Role of Intensifying Agriculture. World Wildlife Fund, Washington.

Mellor, John, Desai, G.M., 1985. Agricultural Change and Rural Poverty. Johns Hopkins University Press, Baltimore.

Mellor, John, Gavian, S., 1999. Determinants of Employment Growth in Egypt: The Dominant Role of Agriculture and the Small Scale Sector. Impact Assessment Report No. 7, Abt Associates, Bethesda, MD.

Mellor, J.W., Lele, U.J., 1973. Growth linkages of new foodgrain technologies. Indian J. Agric. Econom. 28 (1), 35-55.

Mellor, John, Ranade, C., 2006. Why Does Agricultural Growth Dominate Poverty Reduction in Low and Middle Income Countries? Pakistan Development Review, 45:2 (Summer 2006).

Mellor, John, Stevens Robert, D., 1956. The average and marginal product of farm labor in underdeveloped economies. J. Farm Econom. 38 (3), 780-791.

Mellor, John, Weaver, Thomas F., Lele, Uma J., Simon, Sheldon R., 1968. Developing Rural India: Plan and Practice. Cornell University Press, Ithaca.

Mellor, John, Delgado, Christopher, Blackie, Malcolm J. (Eds.), 1987. Accelerating Food Production Growth in Sub-Saharan Africa. The Johns Hopkins University Press, MD, Baltimore.

Minton, Bart, Stifel, David, Timiru, Senesh, 2012. Structural transformation in ethiopia: evidence from cereals markets. ESSP Working Paper. International Food Policy Research Institute.

Minton, Bart, Tamru, Senesshaw, Engide, Erinius, Kum, Tadesse, 2013. Ethiopian Value Chains on the Move: The Case of Teff, ESSP Working Paper 52.

Rao, C.H.H., 1975. Technological Change and the Distribution of Gains in Indian Agriculture. MacMillan, New Delhi.

Rashid, Sahidur, Tefere, Nigussie, Minot, Nicholas, Ayela, Gezahenga, 2013. Can Modern Input Use be Promoted Without Subsidy - An Analysis of Fertilizer in Ethiopia. International Food Policy Research Institute.

Ravallion, M., Datt, G., 2002. Why has economic growth been more pro-poor in some states of India than others? J. Dev. Econ. 68, 381-400.

Rwanda, Republic of, 2007. Strategy for Developing Fertilizer Distribution Systems in Rwanda.

Rwanda, Republic of, 2012. Report on Fertilizer Use.

Stuff, Richard, 2013. An Analysis of Ethiopian Crop Productivity Trends, available by request from jmallor@jmassocinc.com, and shortly on the CSA website.

Tafere, K., Taffesse, A.S., Tamiru, S., 2010. Food Demand Elasticities in Ethiopia: Estimates Using Household Income Consumption Expenditure (HICE) Survey Data, ESSP-2 Discussion Paper, Addis Ababa, Ethiopia: IFPRI Ethiopia Strategy Support Program (processed).

Tamboli, P.M., Nene, Y.L., 2011. Revitalizing Higher Education in India. Asian Agri-History Foundation, Hyderabad.

Thirtle, C., 2001. Relationship Between Changes in Agricultural Productivity and the Incidence of Poverty in Developing Countries. DIVD, Report No. 7946, London.

Timmer, C.P., 1997. How Well do the Poor Connect to the Growth Process? CAER Discussion Paper No. 178, Harvard Institute for International Development, Cambridge.

Tittonell, P., Giller, K.E., 2012. When yield gaps are poverty traps: the paradigm of ecological intensification in African smallholder agriculture. Field Crops Res.

Xie, Hua, You, Liangzai, Wielsgosz, Benjamin, Ringler, Claudia, 2013. Estimating the potential for expanding irrigation in sub-Saharan Africa. Agric. Water Manage..

You, Liangzhi, Ringler, Claudia, Wood-Sichra, Ulrike, Robertson, Richard, Wood, Stanley, Zhu, Tingju, Nelson, Gerald, Guo, Zhe, Sun, Yan, 2011. What is the irrigation potential for Africa? A combined biophysical and socioeconomic approach. Food Policy.