ECONOMIC IMPORTANCE OF AG RICU LTU RE FOR POVERTY

REDU CTION

Chapter 1

INTRODU CTION

1.1 Motivation

Theodore Schultz began his acceptance speech for the 1979 Nobel Prize in

Economics observing: “Most of the people in the world are poor, so if we knew the

economics of being poor we would know much of the economics that really

matters. Most of the world's poor people earn their living from agriculture, so if we

knew the economics of agriculture we would know much of the economics of

being poor”(Shultz, 1979). Thirty years on we still find that people in developing

countries who depend on agriculture for their living are typically much poorer than

people who work in other sectors of the economy and that they represent a

significant share, often the majority, of the total number of poor people in the

countries where they live.

Achieving the Millennium Development G oal (MDG ) of halving poverty by

2015 requires finding ways to increase the incomes of those people. What can

government do to foster that kind of income growth? Specifically, how can we

improve development co- operation, trade and agricultural policy to better

promote agriculture’s contribution to poverty reduction? This paper constitutes

[1 ]
the first output from a research project seeking answers to that question. G lobally,

the percentage rate of poverty (though not in all countries the total number of poor

people)

has declined steadily during the past thirty years, an achievement credited largely

to economic growth (World Bank, 2008a). But what causes economic growth

and, more relevant for present purposes, what causes agricultural growth? There

is widespread agreement on a general list of necessary conditions, e.g.: access to

output and input markets accommodated by a good transportation, marketing

and processing infrastructure; non- discriminatory tax and trade policy; high rates

of investment in agricultural research and extension; a system of ownership rights

that encourages initiative; employment creating non- agricultural growth; well

functioning institutions; good governance and so on. However, debate abounds

on their relative importance and what government ought to do to promote them.

The approach we adopted in seeking to better inform such debate was to look for

shared characteristics of

developing countries posting exceptional success in reducing extreme poverty

over the past twenty to twenty- five years. Inspired by the World Bank’s 2008

G rowth Report (Commission on G rowth and Development), our method

is based on the premise that a limited number of pre- conditions are necessary

everywhere, if not always sufficient,

for sustained progress in reducing poverty. The G rowth Report examined

[2 ]
common features of thirteen countries that since 1950 grew their economies at an

average annual rate of 7% or more for 25 years or longer. In like fashion, we chose

a list of twenty- five countries that in recent years achieved reductions in national

poverty rates at a pace that might enable them to reach their respective MDG ‟s of

halving poverty by the 2015 target year. We start by looking at some indicators of

economy- wide economic performance but then give particular emphasis to

common features of the agricultural economies of the selected countries. We

make no attempt to establish causality or to quantify the relationships between the

various indicators and poverty. We aimed simply to see if those countries chosen

on the basis of their exceptional success in reducing poverty were similar in other

socioeconomic respects. The insights obtained from this multi4 country overview

are meant to inform the design of in- depth case studies of agriculture’s

contribution to poverty reduction in four countries (Ethiopia, G hana, Indonesia

and Vietnam) wherein we may seek to quantify poverty impacts through

econometric or simulation analysis. It is hoped that findings from the case studies

will eventually provide the basis for development of policy principles and

recommendations to foster progress in reducing poverty.

1.2 Objective of study

To carry on all the business of builders, real- estate- developers, contractors,

[3 ]
dealers and by advancing money to and enter into contracts and arrangements of

all kinds with builders, tenants, constructions of residential and commercial

premises including business centers and officers, to deal in and as agents for

lands, building and raising loans for constructions and advancing to other

organizations for similar purposes.

To provide city and region level urban infrastructure facilities, to lay- out

construct, re- model or any building works , road, highways, bridges, dams,

railways such purposes to prepare in designing, and modeling.

1.3 Methodology:

C omparative and logical analysis of the theoretical concepts, methods and

conclusions, published in scientific literature, mathematical and statistical

analysis, with the help of software package MS Word. Authors of scientific

literature suggest many definitions of infrastructure sector and its components;

they widely interpret the features and functions of infrastructure while the issue of

measurement is based mainly on the available data for different regions. Normally

[4 ]
data are taken there are two types such as; primary and secondary data. The

primary data was collected from the program/scheme beneficiaries who were

either individual beneficiaries or a member of SHG s. The secondary data /

information regarding the schemes and their implementation have been collected

from the classified data available with the G ovt. functionaries at different

institutions like Central Statistical organization, Central Statistical Organisation,

Insurance Regulatory and Development Authority, National Hydroelectric Power

C orporation, Reserve Bank of India, National Highways Development project,

Infrastructure Development Department. Newspapers, Magazines, Books,

Journals, Published research papers, published conference proceedings, also

formed the sources for the collection of secondary data for the study.

Poverty is defined as a complex of capital goods which are not consumed

directly; they provide services only in combination with labour and other inputs.

This description allows to distinguish a wide range of components and to analyses

their direct impact on development issues and emphasises the need of

specification of infrastructure sector in order to measure its impact. In this article

infrastructure is defined as the core physical structure consisting of: transportation

infrastructure, water supply and disposal infrastructure, telecommunications

Poverty and power unemployment, consisting of sub sectors that are defined by a

set of physical variables: transportation infrastructure (length of roads, rail tracks,

[5 ]
etc.), water supply and disposal infrastructure (resident population connected to

wastewater collection and treatment systems), telecommunications infrastructure

(number of telephone lines), power infrastructure (power plants, transmission and

distribution lines). Other indicators such as G DP and G DP per capita are from

World Development Indicators, World Bank.

1.4 Limitations of the study

U rban poverty problems in India are a age gold problems. The infrastructure

problems in India mostly took a back- seat in the economic development policy

drafts. The meager budgetary allocation to arrest infrastructure problems in India

has so far proved to be Too little to keep pace with other areas of business

development in India. Moreover, the tremendous growth of Indian IT,

telecommunication, manufacturing, and pharmaceutical industries has consumed

the limited world class urban infrastructure available in India. The U rban

infrastructure problems in India are such as : Business Premises, power, U rban

transport, Sewerage, Airport, railways, Seaports, Roads.

Rural infrastructure problems in India have gone from bad to worse in recent

years. However, the government of India has taken some important steps to arrest

the age old problems of rural India, such as: Connecting 66, 800 habitation with all

[6 ]
weather roads, Construction of 1, 46, 000 kms of new rural roads.

Chapter 2

Review of the Literature

2.1 Introduction :

Historically, few issues have attracted the attention of economists as has the role

of agriculture in economic development and poverty reduction, generating an

enormous literature of both theoretical and empirical studies. Much of this

literature focuses on the process of structural transformation of economies, from

the least developed in which economic activity is based largely on agriculture, to

high- income countries where industry and services sectors dominate.

[7 ]
 A declining share for agriculture in national employment and G DP is an

inevitable consequence of economic progress (Byerlee, de Janvry and Sadoulet,

2009; Timmer, 1988; Cervantes and Brooks, 2009). This is largely due to higher

income elasticities of demand for non- agricultural goods and services. As their

incomes grow, consumers increase their consumption of manufactured goods

and services faster than their consumption of food. Paradoxically, the process is

usually accompanied by rising incomes and a lower incidence of poverty among

those who depend on agriculture for a living. Lewis (1955) was one of the first of

many development economists attempting to explain the paradox.

He viewed economic development as a process of relocating factors of

production from an agricultural sector characterized by low productivity and the

use of traditional technology to a modern industrial sector with higher

productivity. Lewis‟s theory was interpreted as advocating industrialization and

used to justify government policies that favoured protection for domestic

industries and, explicitly or implicitly, taxed the agricultural sector (Kirkpatrick and

Barrientos, 2004). That theory and it implications for policy have been largely

debunked by later work and the degree to which economic policies of developing

countries discriminate against agriculture has lessened dramatically in recent

decades (Anderson and Valenzuela, 2008).

A paper produced by DFID (2004) emphasises the historically close

correlation between different rates of poverty reduction over the past 40 years and

[8 ]
differences in agricultural performance –particularly the rate of

growth of agricultural productivity. The authors see links between agriculture and

poverty reduction as being forged through four „ transmission mechanisms‟: 1)

direct impact of improved agricultural performance on rural incomes; 2) impact of

cheaper food for both urban and rural poor; 3) agriculture‟s contribution to

growth and the generation of economic opportunity in the non- farm sector; and 4)

agriculture‟s fundamental role in stimulating and sustaining economic transition,

as countries (and poor people‟s livelihoods) shift away from being primarily

agricultural towards a broader base of manufacturing and services. They go on to

note that the potential for future poverty reduction through these transmission

mechanisms depends on the extent to which agricultural productivity can be

increased where it is most needed.

Many recent studies focus specifically on quantifying the relationship between

agriculture and poverty. Bresciani and Valdes (2007) frame their analysis in terms

of three key channels they say links agricultural growth to poverty: 1) labour

market, 2) farm income, and 3) food prices. They provide a theoretical framework

for investigating the quantitative importance of those various channels and then

report findings from six country case studies. They conclude that when both the

direct and indirect effects of agricultural growth are taken into account, such

[9 ]
growth is more poverty reducing than growth in nonagricultural sectors.

2.2 Review of the Literature:

Bresciani and Valdes emphasize especially that agriculture‟s contribution to

poverty reduction is consistently greater than is agriculture‟s share of G DP. For

their case study countries, agriculture‟s contribution came mainly through the

labour market channel. They caution however that growth strategies based on

such findings may not be valid in circumstances where the agricultural output mix

does not feature labour intensive crops and livestock activity. Equally problematic

for such a strategy is that much progress in agriculture historically has come from

the introduction of labour saving technical change. In a paper done as

background for the World Bank‟s 2008 World Development report, Ligon and

Sadoulet (2008) combine time series and cross- section data to estimate

regression coefficients connecting consumer expenditures by decile to

agriculture and non-

agriculture G DP. Their findings are consistent with claims that agricultural sector

growth is substantially more

important than non- agricultural sector growth for those households in the lower

deciles of the expenditure distribution, i.e., the poorer segments of the

population. They find the opposite result for richer households, i.e. that the

expenditure elasticity non- agricultural growth is much higher than for agricultural

[ 10 ]
growth leading them to conclude that their findings are consistent with claims that

agricultural sector growth is pro- poor. Christiaensen and Demery (2007) point

out that the contribution of economic growth to poverty reduction might differ

across sectors because the benefits of growth might be easier for poor people to

obtain if growth occurs where they are located. This reasoning implicitly assumes

that transferring income generated in one economic sector or geographic

location to another sector or location is difficult because of market segmentations

or considerations of political economy. They too find that growth originating in

agriculture is on average significantly more poverty reducing than growth

originating outside agriculture.

Similarly, Montalvo and Ravallion (2009) find that the primary sector rather than

the secondary (manufacturing) or tertiary sectors was the real driving force in

C hina‟s spectacular success against

absolute poverty. They conclude that the idea of a trade- off between these sectors

in terms of overall progress against poverty in China is moot, given how little

evidence they found of any poverty impact of non- primary sector growth. While

most empirical studies show that agricultural growth is relatively more important

than growth in other sectors there are exceptions, underscoring the existence of

potentially important differences in the sectoral G DP elasticities of poverty across

countries, depending on the structure and institutional organization of their

[ 11 ]
economies (Loayza and Raddatz, 2006). A common finding is that the poverty

reducing powers of agriculture declines as countries get richer (Christiaensen and

Demery, 2007; Ligon and Sadoulet, 2008). G ardner (2000), for example, found

that gains in income from off- farm sources was the main reason rural poverty

declined in the U S from the 1960s. Econometric analysis by Warr (2002) based on

pooled data for Indonesia, Thailand, Malaysia and

the Philippines showed the services sector as having the greatest reduction on

poverty. Time- series analysis for

Taiwan reported in Warr and Wang (1999) found industrial growth to be most

poverty reducing. Similarly, Ravallion and Datt (1996 and 2002) found that the

elasticity of rural headcount poverty with respect to agricultural growth in India is

less than half that for non- agricultural sector growth. They speculate that the latter

occurs because of rapid growth in the informal sector of the Indian economy.

Interestingly, using a similar method of analysis for China Ravallion and Chen

(2007) estimate that agricultural growth had four times greater impact on poverty

reduction than growth in the secondary and tertiary sectors. Previous research

suggests that agricultural income growth is more effective in reducing poverty than

growth in other sectors because: 1) the incidence of poverty tends to be higher in

agricultural and rural populations than elsewhere, and 2) most of the poor live in

rural areas and a large share of them depend on agriculture for a living (World

Bank, 2008b; Christiaensen and Demery, 2007; Ravallion and Chen, 2007).

[ 12 ]
 However, even if the incidence of poverty is lower within the population of non-

farm people (whether rural or urban) growth in income from non- farm sources

could be proportionally more effective in reducing poverty. Moreover, it could be

that even for poor farm families, growth in income from non- farm sources is more

important than growth in farm income. We introduce another complication by

acknowledging that perhaps growth in per capita income economy- wide is itself

driven by growth in agricultural sector income, i.e. that agriculture is the engine of

economy- wide performance (Irz and Tiffin, 2006). G ardner and Tsakok (2007)

review past attempts to

draw causal connections between economy- wide growth and growth in one or

another economic sector. They conclude the task of explaining economic growth

might be better served by searching for a common set of factors simultaneously

driving growth in all sectors. Note that, purely in terms of the arithmetic of growth

accounting, agricultural sector growth will be a more important driver of overall

growth in countries where its sector share is large. Of course, the claim that

agriculture is the engine of economic growth is not based solely on the growth

accounting arithmetic. Many people believe there is more to the story because

agricultural sector growth exhibits a higher multiplier than growth in other sectors

(Bresciani and Valdes, 2007). Though few countries have achieved rapid poverty

reduction without it, a nation‟s economic growth is not absolutely essential to

[ 13 ]
progress in reducing poverty. As we measure it, poverty refers to how much

money poor people spend on goods and services. Earnings

from work are of course the most important source of spending money for most

poor people but some get money from other sources. For example Ravallion

(2009) using the U SD 1.25 per day indicator shows that sustainable poverty

reduction is theoretically possible through financial transfers from higher to lower

income people in all but the poorest of developing countries. One source of extra

money known to be especially effective in reducing poverty is remittances from

people who work abroad (Acosta, Fajnzylber and Lopez, 2007). Data measuring

remittances are conveniently available from the World Bank. Later we use this data

to illustrate the more general point that poverty reduction may be achieved

through channels other than pro- poor economic growth. Another

route by which poverty could be reduced even in the absence of economic growth

is through migration of farm workers to off- farm jobs, either in rural or urban

areas. C hristiaensen and Todo (2008) observe that as countries develop: a) their

economies restructure away from agriculture into manufacturing and services and

b) people move from rural to urban areas. They emphasize however that, while

intertwined, these structural and spatial transformation processes typically do not

fully overlap. They find that migration from farm to non- farm work in rural areas is

poverty reducing but not migration from farm to non- farm jobs in urban areas.

Byerlee, de Janvry and Sadoulet (2009) report findings from World Bank (World

[ 14 ]
Bank, 2008b) analysis showing that migration from rural to urban areas

accounted for less than 20% of the reduction in rural poverty during 1993- 2002.

The other 80% came from improvements in economic conditions in rural areas,

including in agriculture.

Measuring poverty and success in reducing it Our method requires first

choosing a list of countries that can be judged successful in reducing their national

poverty rates. To proceed we therefore need both a definition of poverty and a

way of ranking countries according to their progress in reducing it. In tracking

progress for the MDG ‟s, poverty in the developing world is measured by a

standard representing the poverty lines found among the poorest countries of the

world. That line was first set at U SD 1.00 a day in 1985 prices. Although the term

„ dollar a

day‟ still features in popular discussion, the line is now U SD 1.25 a day in 2005

prices, which is the average of the poverty lines found in the poorest 15 countries

in terms of per capita consumption (Chen and Ravallion, 2008). Of course,

depending on the purpose, other poverty thresholds are possible. A common

choice is U SD 2.00 per day - the one we used in this analysis. The U SD 2.00 per

day line corresponds to the median poverty line for all developing countries (Chen

and Ravallion, 2008). We chose the U SD 2.00 threshold after experimenting with

lower cut- off points, including the U SD 1.25 one. The problem was that too few

[ 15 ]
developing countries had both high rates of initial year poverty (first year for which

poverty survey data were available) and showed rapid progress in reducing them

when measured using lower cut- off points. For example, Chile posted

spectacular gains in reducing U SD 2.00 per day poverty during the past quarter

century, outpacing most other countries when using that standard. However,

when using the U SD 1.25 standard, initial year poverty rates in Chile were already

too low to show much gain from that exceptional performance. The procedure

used to decide whether, in any given period of time, someone falls below a chosen

poverty line requires three kinds of information: 1) the composition of the basket of

goods and services consumed by that individual, including goods produced for

self- consumption; 2) a local currency price to value each item in the basket in

2005; and 3) an exchange rate to convert from local currency to U S dollars. The

World Bank collects and harmonizes consumption estimates obtained from

household surveys done by national statistical offices –purportedly the world‟s

largest single statistical endeavour. The frequency of the surveys and the country

coverage has increased sharply in recent years. Current estimates are based on

675 surveys, spanning 1979- 2006 and 116 countries (Chen and Ravallion, 2008).

The main data source for prices and exchange rates has been the price surveys

within countries done for the International Comparison Program (ICP) managed

by the World Bank‟s Development Data G roup. Local currency expenditures are

converted to dollars using purchasing power parity (PPP) exchange rates in order

[ 16 ]
to assure international comparability of consumer expenditures, i.e. those U SD

2.00 have the same command of goods and services in one country as another

(and irrespective of whether those goods and services are tradable or not). In

2008, the PPP exchange rates were updated based on price surveys from 2005, a

year for which country coverage of the World

Bank‟s cost of living surveys is much greater than in the past. Accordingly, 2005 is

also the base year for price information. With this information in hand one then

calculates the level of an individual‟s real expenditures on

goods and services in a particular survey year by, in effect if not in reality,

multiplying each item in his/her consumption basket by its local currency price in

2005, then converting to dollars by multiplying by the dollar to local currency PPP

exchange rate. If those expenditures are less than the chosen poverty threshold –

e.g. the U SD 2.00 per day figure which we use, that individual is considered to be

in poverty. The results for individual survey respondents are then extrapolated to

the whole population to obtain estimates of the total number of people in poverty

(the poverty head- count) as well as the percentage of the population in poverty

(the poverty rate). Thus, in comparing between two time periods the poverty head-

count and the poverty rate both rise and fall as real expenditures rise and fall

around the poverty threshold. The change in real expenditures between any two

time periods will reflect changes in income or prices between those two periods.

If, per capita income rises, expenditures on goods and services will also rise. The

[ 17 ]
mathematical relationship between consumer expenditures and income, the

marginal propensity to consume, tends to be higher for poor than for rich people.

Thus, as the incomes of poor people increase some of them begin to spend more

than the threshold expenditure per day leading in turn to a lower poverty head

count and poverty rate. Likewise, a reduction in consumer prices permits

consumers to purchase more goods and services with the same budget and will

also show up as an increase in real expenditures leading to a lower number of

people judged to be in poverty. G ood agricultural performance operates to

reduce measured poverty through both the income and the price channels.

Because a high share of the poor depend on agriculture for their incomes, it is

natural to think that an increase in farm income would be poverty

reducing, perhaps as findings from previous research suggests, even more so

than a general rise in incomes. Similarly, because food constitutes such a high

share of consumer expenditures by the poor it is also tempting to think that lower

food prices, such as might accompany increased food production per capita,

would be poverty reducing. However, this relationship is not guaranteed. An

ambiguity arises precisely because so many poor people depend on farming for a

living. Thus, depending on what causes prices to fall, how much they fall and the

commodity composition, a decline in food prices might simultaneously reduce the

earnings and purchasing power of some poor farmers while increasing the

purchasing power of some poor consumers. These possibilities put a question

[ 18 ]
mark on the relationship between poverty and food production as an area

requiring further exploration.

Selection process

We turn now to the specifics of the selection procedure and results obtained in

applying it. There were four distinct steps. First, we identified a list of countries

that: a) exhibited an initial U SD 2.00 per day and a poverty rate of more than 10% ;

b) posted reductions in that rate over the entire range of years for which poverty

data are available, within the 1980- 2005 range; and c) had at least two years of

poverty survey data to calculate trends. This meant we automatically excluded

countries where poverty was already relatively low and where the poverty rate

either stayed the same or increased. In the second step we calculated the average

annual reduction in the poverty rate posted by each of those countries over the

entire range of years for which poverty estimates are available. The range of years

covered by poverty surveys (from the initial to the most recently published survey)

and the number of annual surveys conducted within that range of years varies

greatly from one country to another. Our third step in selection process was based

on the observed pace of poverty reduction. In this step we chose only those

countries

where the annual average decline in the poverty rate from the year of the first to the

[ 19 ]
year of the last observation (survey) would permit a halving of their respective

initial poverty rate in 30 years or less. Finally, we dropped

countries which for one reason or another (oil rich countries, small island states,

etc.) we judged unrepresentative for drawing general conclusions. Twenty- five

countries made the final cut. The first column of Table 1 lists them. The two

subsequent columns show the rate of poverty observed in the first and final survey

years respectively. The

Third column contains the

estimated annual average reduction in the poverty rate for the years of data

availability while the final column presents the year ranges and number of annual

surveys used in making the calculations. Although our selection procedure

guarantees that every country in the list achieved some progress in reducing

[ 20]
poverty there are large differences among them in just how much progress was

actually achieved. China represents an overwhelmingly important extreme case. In

1981, the first year of poverty data availability for that country, 98% of the

population was living below the U SD 2.00 per day standard whereas by 2005 that

percentage had fallen to only 36% . Including China, eight countries in the list

halved poverty rates in the years between their respective first and last survey year

and others are on pace to achieve similar reductions in the next few years. In other

countries though, e.g. Mali, the poverty rate was extremely high in the first year of

data availability and has been declining only very slowly since.

G eneral characteristics of selected countries

Did the countries chosen on the basis of their achievement in reducing poverty

perform well on other indicators of socioeconomic progress? Table 2 contains

estimates of a development indicator monitored by the U nited Nations

Development Program called the Human Development Index (HDI). The HDI index

is a summary composite index that measures a country's average achievements in

three basic aspects of human development: health, knowledge, and a decent

standard of living. Health is measured by life expectancy at birth; knowledge is

measured by a combination of the adult literacy rate and the combined primary,

secondary, and tertiary gross enrolment ratio; and standard of living by G DP per

capita (PPP U SD). It is expressed as a value between 0 and 1. The closer a

country‟s index is to 1 the higher its rank on the HDI. We use the index here to

[ 21]
corroborate, rather than to explain, the achievements made by our countries in

reducing national poverty rates. The rows of that table separate countries into

High, Medium and Low groups. Interestingly the only two of our countries in the

Low Human Development group are African countries while all but one of our

countries appearing in the top group are Latin American countries. The middle

group constitutes a mix of countries from different continents. With two

exceptions (Tajikistan and Kenya) all twenty- five countries chosen for their

exceptional progress in reducing poverty also posted improvements in their HDI

scores. In most cases those countries posting the fastest progress in reducing

poverty also posted the greatest improvement in their HDI scores.

[ 22]
 We now turn to a comparison of the features of economy- wide economic

performance of our twentyfive

countries that might help to explain their achievements in poverty reduction and

the corroborating improvements in their Human Development scores. Table 3

contains a short list of macroeconomic indicators that often feature in descriptions

of a country‟s economic performance. The list begins with an estimate of the

economy- wide growth in G DP/capita. Economic growth is viewed by many

economists as the only sustainable cure for poverty. U nsurprisingly then, the

majority of the countries in our list experienced positive per capita income growth

during the years when their poverty rates were falling. Some countries posted

reductions in poverty even though per capita incomes were falling. In some cases,

e.g. Tajikistan, this may be explained by differences in coverage of the poverty and

[ 23]
income data. It could well be the case that achievements in poverty reduction

occurred during sub- periods when per capita incomes were rising even if they fell

when considering the entire range used in calculating income growthrates.1

Additionally, as already noted, economic growth is not strictly necessary for a

country to achieve progress in reducing poverty.

[ 24]
 The middle columns of Table 3 compare for each of the selected countries the

evolution from 1980 to 2005 of an index of trade openness - the sum of exports

and imports expressed as a percentage of national G DP. The higher the value of

this percentage, the less restrictive trade policy is seen to be. Interpreted in this

way, almost all countries improved their performance (became more trade

friendly) during the period when their poverty scores were also improving. In the

few cases where trade openness did not improve, the declines were relatively very

small.

The final columns of Table 3 show the evolution of an indicator of macroeconomic

performance based on data from the International Country Risk G uide (PRS-

[ 25]
G roup, 2009) and used as a barometer of overall economic

health of a country. A country‟s score on this indicator is based on the average of

three measures

[ 26]
 Chapter 3

Data Analysis

This index too indicates significant improvement in economic conditions in

virtually every one of the selected countries from the mid- 1980s to present times.

The overall picture that comes into to focus when looking at the figures in Table 3,

corroborated by findings from other analyses, e.g. in the World Bank‟s

development report on agriculture (World Bank, 2008b), is that countries

achieving success in reducing poverty did so while posting impressive progress in

macroeconomic performance. The accumulated body of research findings on the

subject leaves little doubt that successful macroeconomic performance is, if not

strictly causal, a necessary pre- condition to success in combating poverty.Table 4

tabulates growth rates of real agricultural G DP/worker, non- agricultural G DP/

worker and remittances per capita. The agricultural G DP per worker series is, as

the name implies, the ratio of total G DP for the sector divided by the estimated

number of economically active workers claiming agriculture as their main source

of income. Non- agricultural G DP per worker was defined residually, i.e. as the

difference between total national and agricultural G DP divided by the difference

between total national and agricultural employment. Agricultural G DP comprises

[ 27]
the returns to land, labour and capital used in agriculture. It constitutes a good

indicator of farm income trends assuming farmers own most of the land and

capital and supply most of the labour used in the sector. There are known biases

in, and measurement problems with the data. Particularly troubling is the fact that

the annual estimates of economically active workers are too often extrapolations

from very few, sometimes only one, actual employment surveys. Moreover,

because of a high incidence of part time farming, the number of workers in

agriculture may be overestimated and thus estimates of agricultural G DP per

worker in agriculture underestimated –a measurement problem that is more

severe the less developed is the country in question (Schmitt, 1990).

[ 28]
 In many

employment surveys an individual is counted as employed in a particular sector of

the economy if he/she earns more than 50% of their income from or devotes more

than 50% their working time to that sector. Because the incidence of part- time

work is typically much higher in agriculture than in other sectors the employment

statistics thus simultaneously over- state employment in agriculture and

understate employment in other sectors. This leads, in turn, to estimates of

average labour productivity (G DP/worker) that are biased downward for

agriculture and upward for non- agriculture. These measurement problems are

[ 29]
greater for developing than developed countries because agriculture‟s share in

total employment is typically higher in developing countries.

The data in Table 4 reveal a widely varying pattern of per worker G DP growth rates

among the selected countries over the study years. Strikingly, agriculture G DP per

worker grew in 20 of the 25 countries. That proportion rises to 23 of 25 countries

if we restrict our attention to only those ranges of years covered by the poverty

data. On the other hand, average per worker G DP in non- agriculture grew in only

12 of 25, i.e. less than half of the countries studied. This pattern is consistent with

[ 30]
two characteristics typifying the normal development process. First, it is usual that

as countries develop, per worker agricultural G DP grows faster than per worker

G DP in other sectors. Second, it is also common that in developing countries most

poor people depend on agriculture for a living.

Figures 1 to 3 plot the complete dataset of time- series and cross- section

observations for the three income variables and poverty rates. Each dot in these

Figures pairs a year by country observation for the poverty rate and, respectively:

agricultural G DP per worker (Figure 1); non- agricultural G DP per worker (Figure

2); remittances per capita (Figure 3), for each year of survey data available. These

plots reveal the expected negative relationships between poverty rates the three

income categories. But, among the three, which has been the most important

source of reduction in observed poverty rates? Answering such a question

requires, first, quantitative estimates of the statistical relationship between each of

the three variables and the poverty rate. We estimated these relationships using

multiple regression analysis employing a dataset that combined all of the cross-

section and time- series data for all available years of poverty surveys.

[ 31]
[ 32]
[ 33]
 The estimating equation, estimated coefficients and their statistical

properties are reproduced in the Annex. The regression equation explains a high

percentage of variation in the time- series, cross- section poverty rate data. The

regression coefficients for agricultural G DP/worker, non- agricultural G DP/worker

and remittances per capita are all statistically significantly negative as suggested

by theory and confirmed by the data plotted in Figures 1- 3. The estimated

coefficient on agricultural G DP/worker is significantly higher than that for either of

the other two variables but this does not necessarily imply that growth in

agricultural G DP/worker was more important than growth in the other two

variables since the answer to that question also depends on actual rates of growth

in the three variables over the study period. To make judgements about the relative

historical importance of agricultural versus non- agricultural growth versus

remittances, we used the estimated regression equation to simulate historical data

and then attribute reductions in predicted poverty rates among the three variables.

There were three steps. In

the first step we generated a baseline of predicted poverty rates by plugging into

the regression equations observed values for each of the three independent

variables for each year of the entire study period 1980- 2005.

In the second step, we created three alternatives to that baseline by replacing

actual observations for one or another of the three income variables by its sample

[ 34]
mean. In the final step we compared, one by one, the predicted values obtained in

the three alternative scenarios to those from the baseline. These comparisons

allowed us to calculate how much of the predicted change in poverty could be

attributed uniquely to each income source.Table 5 shows the breakdown thus

obtained, revealing that for 12 out of the 25 countries growth in agricultural G DP

per worker was more important, followed by growth in remittances per capita (9

out of 25) with only four countries shown to have reduced poverty mainly because

of growth in non- agricultural G DP per worker. Notice however that in some

countries, e.g. Vietnam, there was little or no difference between the estimated

contributions of growth in agricultural G DP per worker as compared to that of non

agriculture G DP/worker. As another example, in Mexico, the contribution was the

same for agriculture as for remittances.

[ 35]
Another way of looking at these results is to ask, what proportion of the observed

reduction in predicted poverty rates was due to each of the variables individually.

Figure 4 shows these results, revealing basically the same

pattern as suggested by the country lists in the table. Specifically, overone- half the

reductions in poverty in the

selected countries was due to growth in

agricultural incomes, over one- third to growth in remittances and only just over

[ 36]
10% due to growth in non- farm incomes.

C haracteristics of countries where agriculture contributed positively to poverty

reduction

The above analysis is fully consistent with most prior analyses in showing that

agricultural progress contributes strongly to poverty reduction. Now we want to

see if there are common characteristics of the agricultural economies of those

countries where agriculture contributed positively to reducing poverty that might

help us better understand what features of agricultural performance government‟s

might wish toemphasize in their development efforts. Table 4 shows that

agricultural G DP/worker grew, and thus contributed positively to poverty

reduction, in twenty out of the twenty five countries.

3.1 Agricultural trade policy

[ 37]
 A frequently cited essential ingredient in the recipe for agricultural success is

access to world markets unfettered by too much interference either by home

country or trading partner governments (Anderson and Valenzuela, 2008) So,

what trading environment confronted the selected countries and how did it

change over the twenty five year study period? The data in Table 6 provide a partial

answer to this question. The numbers in the table are estimates of the Nominal

Rates of Assistance (NRA), an estimate of the percentage by which government

policies have raised/lowered gross returns to farmers above what they would be

without the government‟s intervention (Anderson and Valenzuela, 2008). Data

was available only for thirteen out of the twenty countries where agriculture

contributed positively to poverty reduction.The last row contains NRA results for

high income OEC D countries, included to show how much trade protection and

support farmers in these important trading partner countries received. Note that

these latter will substantially overstate OEC D trade protection confronting those

[ 38]
developing countries in the list who benefit from preferential access to OECD

markets under a wide variety of preferential trading agreements.

Interpreting the 2000- 05 averages as indicating the current state of affairs

we see that farmers in the selected countries now receive rates of government

price support that are generally positive. Note moreover that the NRA‟s during the

1980s were mostly negative, often significantly so, showing that on net,

government interventions taxed rather than subsidised farmers. That is to say, in

general over the entire period and for most all of the selected countries the rate of

disprotection caused by government interventions (export taxes, overvalued

exchange rates and so on) was declining. The turnarounds were especially

dramatic in Brazil, China, and Vietnam, three countries also posting exceptionally

rapid declines in poverty.

Now, looking at the final row in the table we see that the high, positive rates

of trade protection and price subsidy afforded rich country farmers were generally

declining. That is to say, the protection confronting developing countries in rich

country markets since the 1980s has progressively and significantly declined, a

development borne out when looking at more comprehensive estimates of OECD

farm support reported in the annual Monitoring and Evaluation Report (OECD,

[ 39]
2009). Taken together then, the trading environment confronting farmers in the

selected countries was one of declining disprotection in the home country and

declining positive protection in the rich country trading partners.

3.2 Agricultural research

Agricultural progress in modern times, typically measured by growth in total

factor productivity, has been driven more by technical advance than by any other

factor. Empirical analysis repeatedly confirms that the social rates of return to

public investments in agricultural research, extension and education are high

(Mundlak, 2000). Figure 5 compares annual average growth rates of spending on

agricultural research by governments of the selected countries with the OECD.

These data come from IFPRI‟s Agricultural Science and Technology Indicators

(ASTI) database, and was available for sixteen out of the twenty

countries. They show that, in general and with only three exceptions, rates of

spending on agricultural research by the success story countries increased during

the study period. Moreover, in most cases the pace of increase was much faster,

albeit from a lower base, than on average in OECD countries. The pattern of

findings reported in Figure 5 for Brazil, China and Chile is confirmed by findings

reported in in- depth studies of agricultural policies in those three countries done

by the OECD. Those country studies report data showing annual average rates of

increase in public spending on the entire package of research, extension and

education of 3% (1995- 2005), 16% (1993- 2005) and 10% (1990- 2005) for those

[ 40 ]
three countries respectively. The comparable rate for the OECD region is only

1.3% (1986- 2005).

3.3 Agricultural productivity and poverty

The payoff from investments in agricultural research, development,

extension and education comes in the form of sustained increase in agricultural

productivity. Comparisons of agricultural performance among countries and over

time are frequently made using partial productivity indicators such as output, e.g.

per unit of land, or head of livestock or agricultural worker. However these

indicate only the trends in output relative to one input and can be misleading in

cases where the input mix is changing or, especially, where there are technical

advances allowing increases in output for a given level of input use. A superior

measure, frequently used to overcome these problems is total factor productivity

(TFP). Thirtle, Lin and Piesse (2003) examine the impact of total factor

productivity growth on the incidence of poverty in the LDCs, as measured by the

percentage of the population living on less than U SD 1.00 per day. Employing

regression analysis their empirical analysis shows that agricultural productivity

growth has a substantial impact on poverty reduction, whereas productivity

growth in industry and services does not.They use their empirical findings to show

that investment in agricultural R&D has had a substantial impacton poverty

reduction in Africa and Asia, as well as paying for itself by being an extremely

profitable investment We should expect therefore that our selection of countries

[ 41 ]
where agriculture contributed to extraordinary progress in poverty reduction

might also have posted strong productivity gains. Fuglie (2008) reports findings

from a comprehensive study of trends in total factor productivity covering 173

countries from 1961 to 2006. Figuruses estimates taken from that analysis to

compare performance of our selected countries and their respective

regions.Notice that TFP growth rates were positive in all twenty of our chosen

countries, with most averaging well above 1.6% per year which was the global

average estimated by Fuglie for the range 1991- 2006. Furthermore, more

countries scored at or above their respective regional average than did not.

Moreover, consistent with

findings from Thirtle, Lin, and Piesse (2003) there is a strong correlation between

rates of progress in TFP and in poverty reduction, i.e. those countries posting the

fastest progress in TFP were generally those posting the fastest progress in

reducing poverty. On the whole then it seems safe to conclude that agricultural

TFP growth was a shared characteristic of the selected countries, undoubtedly

contributing to poverty reduction.

[ 42 ]
[ 43 ]
.

[ 44 ]
3.4 Expenditures on agriculture

What about other kinds of government expenditures on agriculture? There is a

widespread belief that agricultural success is systematically related to how high is

the share of total budgetary expenditures that goes to agriculture. Indicative of this

belief are the commitments embodied in the African U nion‟s Comprehensive

Development

Program for African Agriculture (CAADP) whereby African governments have

[ 45 ]
large number of countries. Table 7 below compares estimates of the share of total
budgetary outlays on agriculture for those countries for which data are available
(fifteen out of the twenty countries). The data is generally not available as a
continuous time- series of annual observations. Accordingly, we divided the
comparisons between the averages of those observations which are available for
two different ranges of years 1989- 97 and 1998- 2005 to give some idea of the
progression. There is wide variation amongst the countries and noconsistent
pattern of change over time

[ 46 ]
 Chapter 4

Conclusion

The countries we judged successful in achieving poverty reduction constitute a

highly diverse mix. The selection includes some of the poorest and some of the

richest developing countries in the world, representing virtually all geographic

regions. The countries also differ greatly amongst themselves in their systems of

governance and economic management. During the period when they posted

their impressive success in reducing poverty they were also experiencing

substantially positive improvements on other economic performance indicators:

1) by most measures the macroeconomic context became progressively more

favourable; 2) their own governments were reducing disprotection by lowering

export taxes, overvalued exchange rates and by dismantling inefficient state

interventions in agricultural markets; and 3) the governments of rich country

trading partners were reducing the most production and trade distorting kinds of

support offered their farmers. The accumulated body of research on this issue is

clear that successful macroeconomic performance is, if not strictly causal, a

necessary pre- condition to success in combating poverty. At the same time, we

found that while economic growth generally was an important contributor to

poverty reduction, the sector mix of growth mattered substantially. Especially

relevant to the objectives of the overall project of which this paper is part was the

[ 47 ]
great importance of agricultural sector growth for poverty reduction in a majority

of the selected countries. Looking at the question in that way permitted us to make

a preliminary partition of the importance of growth in agricultural G DP/worker

relative to that of growth in non- agricultural G DP/worker and remittances per

capita. That analysis attributes to per worker growth in agricultural G DP the

majority share of progress in reducing poverty in those countries posting the

greatest progress

in doing so.During the study period public expenditures on agricultural research in

the selected countries were increasing generally and significantly faster than in the

OEC D region. Perhaps as a reflection of that extra investment, in all of the

countries where agriculture contributed to rapid progress in poverty reduction,

total factor productivity rose, and at rates generally higher than other countries in

their respective regions and globally. Although the data are somewhat shaky, the

share of the total government budget spent on agriculture is not extraordinarily

high and has generally been declining.

Due caution is needed in interpreting these findings, and in particular it is

premature to draw policy conclusions. The purpose of this paper was simply to

obtain an overall picture of the economic characteristics of those countries

achieving the fastest progress in reducing poverty. It would be wrong to conclude

on this basis of this paper, for example, that the more investment there is in

[ 48 ]
agriculture, the more growth will follow and the more poverty will be reduced.

C areful attention needs to be paid to the specific situation in individual countries,

to the nature of investments in the sector, and in particular to the macroeconomic

environment in which the sector operates.

Chapter 5

REFERENCES

Acosta, P., Fajnzylber, P. and H. Lopez (2007), “The Impact of Remittances on

Poverty and Human Capital: Evidence from Latin American Household Surveys”,

World Bank Policy Research Working

Paper No. 4 247.

Anderson, K. and E. Valenzuela (2008), Estimates of G lobal Distortions to

Agricultural Incentives, 1955

[ 49 ]
to 2007, World Bank, Washington, DC, October 2008.

ASTI- IFPRI (2009), Database on Agricultural Science and Technology Indicators,

IFPRI.

Bresciani, F. and A. Valdés (2007), Beyond Food Production: The Role of

Agriculture in Poverty

Reduction, FAO, Rome.

Byerlee, D. de Janvry, A. and E. Sadoulet (2009), “Agriculture for Development:

Toward a New

Paradigm”, Annual Review of Resource Economics, Vol. 1: 15- 35, October 2009.

C ervantes- G odoy, D. and J. Brooks (2008), “Smallholder Adjustment in Middle-

Income C ountries: Issues

and Policy Responses”, OECD Food, Agriculture and Fisheries Working Papers,

No. 12, OECD,

Paris.

[ 50]