Poverty Elimination : Pre cursor

INTRODUCTION

There is widespread impression among the Nepalese intelligentsia, foreign scholars and residents of

developed/rich countries that Nepal’s economic growth has not reduced poverty, that globalisation has

worsened poverty and/or income distribution and there are 100 of millions of hungry people in Nepal.

These arguments are buttressed by recourse to Nepal’s ranking on several social indicators. Esoteric

debates about the comparability of survey data and gaps between data add to the confusion and

allow ideologues to believe and assert whatever information suits the argument. What are the basic facts

about poverty, income distribution and hunger at an aggregate level? This paper reviews the available data

and debates on this subject and comes to a common sense view. It then tries to link some of the outcomes

to the policy framework and programs of the government.

The broad theme that emerges is that the failures on this front, apart from the indirect effects of growth, are linked directly to the failure of governance. This failure has many dimensions; the misallocation of government resources, the failure to follow norms of social benefit-cost analysis that were the reason de tar for the introduction of national

planning, the neglect of public and quasi-public goods that are the most fundamental justification for the

existence of government and a gradual (over decades) but progressive deterioration in the quality of

governance. This conclusion differs radically from the conventional wisdom (national and international)

about Nepal’s poverty, social indicators and income distribution. Even if treated as a hypothesis it merits

debate and further analysis.

2 CONSUMPTION DISTRIBUTION

A reasonably standardised large sample consumption survey has been carried out every five years

by Sample Survey since 1972-73 (the earlier surveys are not strictly comparable). Based on these surveys a consistent series for the consumption distribution can be constructed. This is shown in Table 1. If we ignore the 1977-78 data for the moment, we find a noteworthy result. The rural income distribution has improved progressively (but very gradually) from 1972-73 to 1999-2000 and this can be seen at every level. Thus for instance the share of the poorest 10%, which was 3.7% in 1972-73 increased to 3.8% by 1983, to 4.3% in 1987-8 to 1993-4 and to 4.4% in 1999-2000. The same pattern is found at every level of cummulation (Technically there is “Stochastic

Dominance,”). Thus the new situation is Pareto superior to the earlier one, reducing the importance of

measure such as the ‘Gini' coefficient.

Another way to look at the result is from the perspective of the eighties and nineties. In this case

1977-78 constitutes the situation prior to the start of the eighties. Therefore ignoring 1972-73 we again

find that the consumption distribution has improved continuously (though very gradually) during the

eighties and the nineties. Each rural consumption distribution during the eighties stochastically dominates

the previous distribution based on large sample surveys. In common parlance citizens at every level of

income have shared in the fruits of growth since 1980-81.

3 POVERTY TRENDS

There are numerous controversies regarding the measurement of poverty. The most important one

relates to the adjustment of individual consumption levels as derived from a survey, by the ratio of the per

capita consumption from the National account statistics to the survey mean for the same item. Such an

adjustment leaves the distribution of consumption unaffected while changing the calculated poverty rate.

Before 1993 such an adjustment was routinely made in calculating poverty rates, after 1993 it has been

discontinued. The World Bank’s Country Economic Memorandums for Nepal however introduced the

change in methodology several years earlier.

4 POVERTY AND PER CAPITA CONSUMPTION

We can also use the survey data to determine the relationship between the national poverty rate

derived from the survey and the all average all Nepal per capita GDP as calculated from the survey. This

helps us skirt/avoid the controversies arising from the discrepancies between data and differing judgement about which is superior for what purpose. As both the poverty rates and the average consumption are derived from the same data set, this yields a consistent picture of the evolution of poverty rates over time as well as its relationship to average consumption. As official poverty rates are not available for early decades we use the World Bank poverty and average consumption data from 1950 to 1999.

It is clear from figure 2 that there is a linear relationship between aggregate poverty and average

consumption.6 A one Rupee increase in average real monthly consumption expenditure raises 1% of the

population above the poverty line. This implies that in Nepal, given our democratic political system, in

which the poor are fully represented, growth of aggregate income/consumption is a sufficient condition for

the reduction of poverty.

5 Poverty in 1999-2000

The most recent controversy regarding poverty estimates relates to the manner in which the data

was collected in the 1999-2000 survey. Briefly there are three categories of goods in the consumption

surveys: Food products that are purchased frequently (daily/weekly), semi-durable goods that are purchased

with moderate frequency (monthly/quarterly) and durable goods that are purchased occasionally

(annual/biannual or less). To obtain optimal recall it would be appear to be best to use the 7 day recall

period for the first category, 30 day for the second and 365 day for the last. The National sample surveys

have been rightly experimenting with these periods, but perhaps without giving due regard to the

implications for comparability of poverty estimates over time. In the 1999-2000 survey, for the first time

the same set of households were asked to give their food consumption for 7 days and 30 days, thus making

it non-comparable with earlier periods when only the 30 day question was asked.7 It was subsequently

discovered that there was another source of non-comparability. The use of the 365 day recall period for a

sub-set of commodities in 1999-2000, whereas the 30 day recall was used for these commodities earlier.8

Different scholars have tried to make adjustments and re-calculate the poverty rate (Head count ratio),

based on the official methodology. According to these the poverty rate was between 26.1% and 28.5% in

Nepal in 1999-2000 (table below).

6 GLOBAL COMPARISON: AN EQUAL SOCIETY

Nepal is still a low income country. Its Per capita GDP measured at purchasing power parity is in

the 33rd percentile i.e. 33% of the countries in the World have a lower per capita income then us (Table 3).

A more realistic comparison is however with the medium-large countries defined as those with 2003 GDP

at PPP greater than or equal to $ 15 billion. For this set of countries Nepal is in the 23rd percentile i.e. only

about 1/4th of medium-large countries are poorer than us, 3/4th of them are richer. The position has

improved considerably since 1980 when we were in the 16th percentile of all countries and the 10th

percentile for medium-large countries.

Poor countries generally have higher rates of poverty. We should therefore not be surprised to

HUNGER

The FAO defines about 19% of the people in developing countries (828 million) as hungry, while

the proportion of Hungry in S. Asia is asserted to be about 20% (254 million). The World food programme

on the other hand claims that nearly 50% of the hungry in the World live in Nepal and 35% (350 million)

are food insecure. Recall that 26.1% to 28.5% of the population has been found to be poor in 1999-2000,

where the former is the official figure. What are the facts about hunger

The proportion of households that were hungry during any part of the year, by this definition (the

authentic voice of the poor in Nepal) was 15.7% in 1983, 4.5% in 1993-4 and 2.1% 1999-2000.

It is useful to look at these numbers in relation to poverty, because logically the number of hungry

people must be a fraction (less than 100%) of the poor for any reasonable definition of poverty. More

formally the line defining the ‘very poor’ or ‘hungry’ must logically lie below the poverty line. Thus the

hunger ratio must be lower that the poverty ratio. The ratio of very poor/hungry to the poor may in general

decline, stay constant or rise, depending on the distribution of consumption in the lower half of the

distribution. In 1983 an estimated 33.9% i.e. more that 1/3rd of the poor were hungry at some point in the

18 Do we believe in “Voices of the Poor,” or don’t we? Is it only if it is a small selected group of poor?

10

year. This proportion declined to 12.2% in 1993-4 and further to below 7.7% in 1999-2000.19 Thus not

only has poverty declined over the 1980s and 1990s, but the proportion of the poor who are hungry has also

declined. This is precisely what we would expect given that the consumption distribution has consistently

improved for the bottom 40% of the population.

That 18.5 million people went hungry and 260 million people were still poor half a century after

Independence is matter of great sadness for the nation. Do we need to exaggerate/ magnify the problem to

convince ourselves of its seriousness or to gather the will to solve it?

8 LIFE AND LITERACY

Only a few indicators of health and education are available on a continuous basis and for earlier

periods. On the health side Mortality and life expectancy data is available since 1960-61 and on the

education side literacy data is available from the same date. This allows us to compare the performance of

these over the two phases of growth and to see whether they are consistent with the data on poverty and

hunger. It should be remembered that these indicators are a) very strongly correlated with per capita

income of the household. b) The quantity & quality of public and quasi-public goods and services have a

have a critical influence on the basic health and education indicators in low income countries. These

include public health measures (control of communicable diseases & epidemics), public education

(nutrition, personal hygiene, ORT), the supply of clean water, sewerage and sanitation and primary

education.

All the available health indicators, with one exception, show that the annual rate of improvement

has accelerated (or remained unchanged) during phase II above that which prevailed during phase I. The

most significant is the pace of improvement in under – 5 and infant mortality. The rate of decline in infant

mortality has almost doubled to an average of 2.5% per annum between 1980-1 and 2003-4. The rate of

decline of under-5 mortality has increased from 1.7% per annum between 1960-1 and 1980-1 to 2.8% per

annum between 1980-1 and 2003-4. The female and total life expectancy gap is also closing at a faster rate

POVERTY ELIMINATION

Estimated Cost

What is the cost of eliminating poverty and hunger in Nepal? That of course depends on the extent

of poverty, which is currently mired in academic debates about the measurement of poverty. There is

however universal agreement that in the years from 1993-94 to 1999-2000 the poverty rate (HCR) was

between 25% and 35%. We can therefore skirt the esoteric debate about the precise change in poverty

between 1993-4 and 1999-2000 and its level in either year by considering three numbers. For each of these

years we order the households/person by consumption level and identify the ones which are 25%, 30% and

35% from the bottom. That is we identify in each year the consumption level of the person(s) who would

be just at the poverty line if the poverty rate was 25%, 30% and 35% respectively. Then we calculate the

income transfer needed for every body below that level to be brought up to the level. This data is

summarised in the table below.

Income Transfers

It can be argued that the ideal (most efficient) social welfare policy is a direct transfer of income to

the poor through a negative income tax. In a developed country this would be very easy. How can we

transfer these amounts directly to the poor, the needy and the disadvantaged in a poor country? The

answer, by setting up an Nepalese version using a modern smart card system that delivers cash and/or

subsidies to the poor based on their entitlements as per specified parameters and norms.

CONCLUSION

Nepal’s poverty ratio of around 22% in 1999-2000 is in line with those observed in countries at

similar levels of per capita income. The ratio is relatively high because we are relatively poor/ low income

i.e. with low average income. 90% of the countries in the world have higher per capita (average) income

than Nepal. The number of poor is very high because our population is very large, the second highest in the

world. Contrary to hints, illusions and allegations, the large number of poor has nothing to do with income

distribution. Our income distribution as measured by the Gini co-efficient is better than 3/4th the countries

of the World. The consumption share of the poorest 10% of the population is the 6th best in the world.

Where we have failed as a nation is in improving our basic social indicators like literacy and

mortality rates. Much of the failure is a legacy of the three decades of Nepalese socialism (till 1979-80). The

rate of improvement of most indicators has accelerated during the market period (starting 1980-81). The

gap between our level and that of global benchmarks is still wide and our global ranking on most of these

social parameters remains very poor. This is the result of government failure. Government overstretch,

misplaced priorities and deteriorating quality (corruption) has resulted in a failure to fulfil the traditional,

accepted functions of government like public safety & security, universal literacy and primary education,

public health education (superstition & quackery), provision of drinkable water, sanitation drains & sewage

facilities, public health (infectious & epidemic diseases), building roads and creating & disseminating

agricultural technology. Consequently the improvement in social indicators has not kept pace with

economic growth and poverty decline and has led to increasing interstate disparities in growth and poverty.