Objectives, Strategy and Perspective of Development
Introduction || Objectives of Ninth Plan || Population Growth and Demographic Profile : Perspective || Growth Target for Full employment : Perspective || Growth Performance and Poverty : Perspective || Food Requirement and Agricultural growth : Perspective || Energy and Natural Resources : Perspective || Annexures

Growth Performance and Poverty

1.81 The most significant feature of the 7.4 per cent growth perspective is that over the course of the next 15 years private consumption is likely to grow at marginally above 7 per cent per annum in real terms. With a growth rate of population averaging 1.54 per cent per annum during this period, the per capita private consumption is likely to grow at an average annual rate of 5.5 per cent. In other words, per capita private consumption in India in the terminal year of the perspective period may be over 2.2 times the level that is likely to be achieved in the base year of the Ninth Plan. With such substantial increase in per capita private consumption, the incidence of poverty in the country can potentially reduce significantly.

1.82 The Planning Commission has been estimating the incidence of poverty at the national and state levels as an integral component of formulating appropriate growth strategies. For the Ninth Plan, the quantitative index of poverty has been developed on the basis of the recommendations of the Expert Group on Estimation of Proportion and Number of Poor, constituted by the Planning Commission in 1989. The Expert Group recommended computation of state-wise poverty estimates on the basis of state specific poverty lines and the National Sample Survey (NSS) consumption distribution without any adjustment for the discrepancy between the CSO and NSS estimates of private consumption. The estimates of state-wise poverty since 1973-74 have been revised on the basis of the Expert Group methodology. The latest estimates of incidence of poverty are available for the year 1993-94 on the basis of NSS consumer expenditure data of 50th Round (July, 1993 to June, 1994). The poverty ratios and the number of poor at national level for the different years are given in Table 1-9.

Table 1-9 : Percentage and Number of Poor
-----------------------------------------------------------------
Year          Poverty Ratio            No. of poor (in million)
         ------------------------     ---------------------------
         Rural   Urban   Combined     Rural     Urban    Combined
-----------------------------------------------------------------
1973-74  56.4    49.0    54.9         261.3      60.0     321.3
1977-78  53.1    45.2    51.3         264.3      64.6     328.9
1983     45.7    40.8    44.5         252.0      70.9     322.9
1987-88  39.1    38.2    38.9         231.9      75.2     307.1
1993-94  37.3    32.4    36.0         244.0      76.3     320.3

----------------------------------------------------------------

1.83 The incidence of poverty expressed as percentage of people below the poverty line is observed to have declined from 56.4 per cent in 1973-74 to 37.3 in 1993-94 in rural areas and from 49.0 per cent in 1973-74 to 32.4 per cent in 1993-94 in urban areas. For the country as a whole, the percentage of people below the poverty line declined from 54.9 per cent in 1973-74 to 36 per cent in 1993-94. However, the number of poor in the country remained more or less stable at around 320 million due to the rise in population. This fact should not over-shadow the 300 million persons who joined the non-poor group during the twenty year period between 1973-74 and 1993-94, but underlines the need for containment of the growth rate of population.

1.84 The state-specific poverty lines in rural and urban areas for 1973-74 and 1993-94 are given in Annexure-1.1. The state-wise poverty ratios and number of poor in rural and urban areas as well as for the state as a whole for these two years by Expert Group method are given in Annexures-1.2 and 1.3. It may be seen that the decline in poverty ratio during the last 20 years has been uneven among the states. The pace of poverty reduction was relatively rapid in Kerala, Andhra Pradesh, Tamil Nadu, Gujarat, Punjab and West Bengal. The decline in poverty ratio, however, was not enough to reduce the number of poor in eight major states. These are : Assam, Bihar, Haryana, Himachal Pradesh, Madhya Pradesh, Maharashtra, Orissa and Uttar Pradesh.

1.85 Poverty is closely related with hunger. The percentage of persons without two square meals a day for some months of the year is often termed as the "hunger ratio". It is observed that about 19 per cent of the rural population and 7 per cent of the urban population remained hungry in 1983. As per the latest available information these have come down to about 5 per cent in rural areas and 2 per cent in urban in 1993-94. The percentage of chronically hungry people in total population in rural areas, which was 35 to 40 per cent in Bihar, Orissa and West Bengal in 1983, reduced to 6.6 per cent in case of Bihar and 14 to 16 per cent in other two states in 1993-94. Hunger ratios less than the poverty ratio indicate that all the poor people do not remain hungry. However, any incidence of chronic hunger is still a cause of concern.

1.86 In order to examine the likely effects of the relatively faster growth of private per capita consumption on the incidence of poverty during the next 15 years, it is necessary to project the behaviour of the poverty ratio on the basis of the trends that are likely to obtain in the macro-economic variables. There are a number of ways in which such projections can be made. One method would be to relate the past trends in poverty ratios with major macro-economic variables such as the growth rate of GDP, the share of agriculture, inflation rate, and policy variables such as public expenditure on anti-poverty programmes, using econometric methods. This approach, however, has the disadvantage that it is unable to capture certain important facets of the behaviour of poverty and inequality in the Indian economy. An alternative methodology is to explicitly recognise these specific features and to extrapolate the future trends on the basis of specific assumptions being made regarding the future movements of these variables which are directly involved in the estimation of poverty. The projections on poverty made for the perspective period are based on this latter approach.

1.87 The behaviour of various measures of poverty and inequality in India are given in Table 1-10. The normal method of measuring poverty in India is by the poverty or "headcount" ratio, which is simply the proportion of the number of people below the poverty line in the total population. This ratio, however, makes no distinction within the broad category of the poor depending upon their actual levels of consumption and deprivation. As a result, the poverty ratio fails to capture the depth and severity of poverty in an adequate manner. A measure for capturing the depth of poverty is the poverty gap (PG) index, which adjusts the poverty ratio with the difference between the per capita consumption of the poor and the poverty line expressed as a percentage of the poverty line. This is therefore a measure of the magnitude of the effort that would be required to shift

Table 1-10 : Indices of Poverty and Inequality
----------------------------------------------------------------------------------
            Poverty Ratio     Poverty Gap         Squared Poverty     Lorenz Ratio
                                 Index                  Gap
Year      ----------------  -------------------  -------------------  ------------
          Rural Urban Total Rural  Urban  Total  Rural Urban  Total  Rural   Urban
-----------------------------------------------------------------------------------
1973-74   56.4  49.0  54.9  16.56  13.64  15.95   6.81  5.26  6.48  0.27581 0.30125
1977-78   53.1  45.2  51.3  15.73  13.13  15.15   6.48  5.25  6.21  0.33861 0.34481
1983-84   45.7  40.8  44.5  12.32  10.61  11.96   4.78  4.07  4.61  0.29759 0.33027
1987-88   39.1  38.2  38.9   9.11   9.94   9.32   3.15  3.60  3.26  0.29826 0.35369
1993-94   37.3  32.4  36.0   8.45   7.88   8.30   2.78  2.82  2.79  0.28190 0.33940
------------------------------------------------------------------

the consumption of all persons below the poverty line to the level of the poverty line. A more comprehensive measure of the severity of poverty is the squared poverty gap (SPG). The SPG is composed of not only the poverty ratio and the poverty gap ratio, but it also captures the consumption distribution of the poor as measured by the coefficient of variation. This measure explicitly takes into account the higher intensity of efforts that are required to address the people who are progressively further below the poverty norm in order to bring them out of poverty.

1.88 It will be seen that the annual average rate of decline of the poverty ratio during the period from 1973-74 to 1993-94 has been 2.05 per cent in both rural and urban areas, and 2.09 per cent for the country as a whole. The poverty gap ratio during this period declined by 3.31 per cent per year in rural areas, 2.71 per cent per year in urban areas and 3.21 per cent per year for the country as a whole. The SPG declined even faster by 4.38 per cent, 3.07 per cent and 4.13 per cent per year in rural, urban and the country respectively. The fact that the more distribution-sensitive measures of poverty have declined progressively faster during the last 20 years is a cause for optimism. It implies that the depth and severity of poverty has reduced faster than the incidence, and that over time there has been a trend towards a greater degree of clustering of the poor around the poverty line leading to a situation where economic growth is likely to have a considerably larger impact on poverty reduction than before. This would of course require that the measures of relative inequality do not deteriorate sufficiently to overcome the growth effect.

1.89 The most commonly used measure of relative inequality is the Lorenz ratio. The behaviour of the Lorenz ratio of household consumption expenditure as given by the NSS surveys shows a very high degree of stability, as can be seen in Table 1-10. The near constancy of this measure of inequality for 20 years gives a certain degree of justification to believe that it may continue to remain more or less stable in the future as well. However, there is reason to believe that the measures of inequality based on the NSS consumption distribution may actually understate the true level of inequality in the country. A characteristic of consumption behaviour in India, specially in the recent past, has been that the discrepancy between private consumption expenditure as measured by the CSO and household consumption expenditure as obtained from the NSS has been growing. Although a part of this discrepancy may be attributed to consumption by non-household private entities, such as voluntary agencies and non-governmental organisations, the possibility cannot be ruled out that there is a growing incidence of under-reporting of consumption expenditures by the higher income categories.

1.90 The methodology for projecting the likely behaviour of the incidence of poverty in the country is based on these two specific features of consumption behaviour in the country namely : (a) stability of the Lorenz ratio for household consumption expenditure; and (b) the increasing discrepancy between the CSO and NSS measures of total non-governmental consumption. Since the plan model yields projections of total private consumption, which is equivalent to the CSO measure, the analogue of the NSS measure of consumption has to be derived by projecting the behaviour of the discrepancy term. This is done by relating the extent of discrepancy to the behaviour of commodity structure of output in the economy over the perspective period as obtained from the inter-sectorally balanced growth path given by the plan model.

1.91 The household component of consumption as per the latest estimate of NSS for 1993-94 is 59.7 per cent of total private consumption as given by the CSO and, more importantly, this component for various reasons has steadily reduced in the past 10 to 15 years. For the projections, the household component of the total private consumption has been worked out for three different commodity groups, viz., foodgrain, other food and non-food. The over-all discrepancy between the CSO and NSS estimates of consumption can be decomposed by these three commodity groups, and it is found that in 1993-94 the NSS consumption of food-grains was virtually identical with the CSO estimate, that for other food was 65 per cent of CSO, while that for non-food was only 27.7 per cent. The over-all figure of the NSS analogue of consumption for the perspective period is worked out as a weighted average of these three groups under the assumption that the discrepancy factors remain constant at the 1993-94 values, and with the commodity structure of CSO consumption in the base year as weights. It should be clear from the above estimates of the commodity group-wise discrepancy factors that a relatively higher growth rate of agriculture will lead to a lower increase in the over-all discrepancy term. In other words, this approach explicitly incorporates the observed positive relationship between agricultural growth and poverty reduction by recognising that the trend in increasing implicit inequality has more to do with non-agricultural sectors than with agriculture.

1.92 The over-all rates of economic growth of 6.5 per cent per year in the Ninth Plan (1997-2002), 7.7 per cent per year in the Tenth Plan (2002-2007) and 8.1 per cent per year in the Eleventh Plan (2007-2012), as per the selected trajectory, are accompanied by growth rates of per capita household consumption, which is the analogue of the NSS measure of consumption, of 4.3 per cent per year, 4.8 per cent per year and 5.3 per cent per year in the successive plan periods. These are derived from the assumed structure of growth, which yields annual rates of growth of foodgrain, other food and non-food as 3.2 per cent, 5.8 per cent and 7.5 per cent respectively in the Ninth Plan. In the Tenth and Eleventh Plans, the rates of growth of foodgrain and other food remain similar at 3 per cent per year and 6 per cent per year respectively. In case of non-food, the growth rate is assumed to be 8 per cent per year in the Tenth Plan and 8.5 per cent per year in the Eleventh Plan. This structure of consumption growth in conjunction with the group specific household shares for 1993-94 determines the over-all household share of total consumption in the three successive plan periods. The changing pattern of growth in the successive plan periods results in a gradual lowering of the household component. It was 59.7 per cent in 1993-94, and, as a result of the changes in structure of growth, it is projected to go down to 55.6 per cent in Ninth Plan, 53.5 per cent in the Tenth Plan and 50.1 per cent in the Eleventh Plan. As a result, the rates of growth of per capita household consumption are assumed to be significantly lower than the corresponding growth rates of per capita private consumption.

1.93 The household component of consumption in the base year of the Ninth Plan and the terminal years of each of the subsequent three plan periods is flowed through a vector of state-wise consumption proportions to allocate the total consumption among the states. This vector of consumption proportions is obtained from the growth in total consumption of each state between 1983-84 and 1993-94 relative to the growth in total national consumption. This procedure implies that the relative position of the states in terms of total consumption witnessed during the period 1983-84 to 1993-94 is assumed to remain unchanged throughout the plan as well as in the perspective period. The underlying assumption is that the rate of growth of consumption in the states during the period are derived from their past performance, and their relative positions do not change. The state-wise growth of per capita consumption in the three future plan periods as derived by this procedure are given in Table 1-11. It should be noted that although the growth in consumption of each state should logically be derived from an assumed growth of state domestic product (SDP), this procedure by-passes this connection and assumes constancy in the consumption shares. No implication can necessarily be drawn about the growth in SDPs from these constant shares, since the state-wise savings behaviour can be quite different from each other.

Table  1-11  :  Projected Growth  in  Per  Capita  Household  
                Consumption In the States
                                            (per cent per year)
----------------------------------------------------------------
Sl.   State          Ninth Plan     Tenth Plan   Eleventh Plan
No.                 (1997-2002)    (2002-2007)    (2007-2012)
----------------------------------------------------------------
1.   Andhra Pradesh      4.75           5.26           5.83
2.   Assam               4.62           4.81           5.47
3.   Bihar               4.04           4.58           4.99
4.   Gujarat             4.43           4.94           5.52
5.   Haryana             4.38           4.68           5.32
6.   Himachal Pradesh    3.41           4.36           6.00
7.   Jammu  and  Kashmir     3.53           4.51           6.20
8.   Karnataka           4.61           5.04           5.65
9.   Kerala              5.01           5.31           6.03
10.  Madhya Pradesh      4.06           4.57           5.04
11.  Maharashtra         4.79           5.17           5.80
12.  Orissa              4.99           5.41           6.00
13.  Punjab              4.79           5.05           5.74
14.  Rajasthan           4.08           4.47           4.97
15.  Tamil Nadu          5.05           5.47           6.08
16.  Uttar Pradesh       3.62           4.18           4.47
17.  West Bengal         4.55           5.00           5.59
18.  Other states/UTs    2.39           3.27           4.87
19.  All India           4.29           4.76           5.31
----------------------------------------------------------------

1.94 The state-wise consumption (household component) derived on the basis of the above growth rates are broken up into rural and urban components on the basis of the urban-rural per capita consumption differentials as given by the per capita urban-rural expenditure data of 1993-94, which are assumed to remain unchanged throughout the perspective period, and the population growth in the state. These per capita consumption of the states, which are worked out separately for rural and urban areas, are assumed to follow a two parameter log-normal distribution, which is a standard distribution function used in such analyses and has the distinct advantage that one of the two parameters is directly determined from the Lorenz ratio. As has already been mentioned, the stability of the Lorenz ratio for household consumption expenditure is a critical component of the projection exercise. The dispersion of each of the state-wise distribution functions is computed from the Lorenz ratio of the consumption distribution of each state, using the NSS data on consumer expenditure of 1993-94, and these too are assumed to remain unchanged throughout the three plan periods comprising the perspective.

1.95 From the state-wise distribution functions of consumption expenditure, the state-specific estimates of poverty for the Ninth Plan and for the perspective periods are worked out using state specific poverty lines computed as per the methodology of the Expert Group. These poverty lines have been computed on the basis of the state-wise price indices available for the base year of the Ninth Plan (1996-97) and are assumed to increase at the same rate as the over-all rate of inflation for the projection period. Therefore, this exercise assumes that there is no differential rates of inflation between the states over the perspective period. The poverty ratios are worked out separately for each state in rural and urban areas. The estimate of national level poverty ratio in rural and urban areas have been computed as the weighted average of state-wise poverty ratios as per the Expert Group methodology. In order to check the robustness of the estimates, the incidence of poverty in rural and urban areas at the national level have also been worked out from the national level consumption distribution of the NSS and a national poverty line. The estimates of poverty following this procedure have been worked out for the base and terminal year of the Ninth Plan (1996-97 and 2001-02) and at the end of the Tenth and Eleventh Five Year Plans. The poverty ratios at the national and state levels are given in Table 1-12. As can be seen, the increase in per capita consumption projected over the perspective period is able to reduce substantially the poverty ratio in all the states, including the poorest states.

Table  1-12 : Projection of Statewise Poverty Ratios in  the
              Perspective Period
                                               (per cent)
---------------------------------------------------------------
Sl.No.  State           1996-97   2001-02   2006-07    2011-12
---------------------------------------------------------------
1.   Andhra Pradesh      17.35     11.13     5.35      2.44
2.   Assam               26.46     10.65     3.82      2.07
3.   Bihar               44.09     27.46    14.08      6.52
4.   Gujarat             17.07      9.05     3.94      1.28
5.   Haryana             18.41     10.16     5.00      2.58
6.   Himachal Pradesh    22.38     14.02     7.34      3.14
7.   Karnataka           30.72     17.86     8.68      3.45
8.   Kerala              21.22     11.35     4.76      1.38
9.   Madhya Pradesh      33.38     21.97    12.75      6.81
10.  Maharashtra         32.74     20.66    11.41      5.43
11.  Orissa              40.21     22.93    10.76      4.63
12.  Punjab               8.02      2.98     0.85      0.15
13.  Rajasthan           20.31     10.85     4.70      1.52
14.  Tamil Nadu          30.73     18.11     8.96      3.59
15.  Uttar Pradesh       32.52     21.91    12.88      6.92
16.  West Bengal         25.08     13.54     6.26      2.86
-------------------------------------------------------------
All India                29.18     17.98     9.53      4.37
All India (weighted)     29.02     17.65     9.19      4.39
-------------------------------------------------------------

1.96 These estimates of state-wise poverty, however, are critically dependent on three important assumptions : (a) unchanged inequality in the distribution of consumption; (b) unchanged urban-rural per capita consumption differentials; and (c) unchanged poverty norms for all states and the nation. The likely validity of each of these assumptions needs to be examined. The structure of consumption growth in the various states when compared with the pace of poverty reduction reveals that reduction in poverty in all the states is not uniformly dependent on the rise in per capita consumption. Some of the poor states are projected to experience a rate of poverty reduction which is higher than that at the national level despite a lower rate of growth in per capita consumption as compared to the national average. Bihar, and to some extent Orissa, are examples. The projected poverty reductions in most of the poorer states are quite significant in both rural and urban areas. It appears that the states with relatively lower initial inequality in consumption distribution have been projected to experience a faster reduction in poverty than the states with a relatively higher inequality. This phenomenon arises out of the sensitivity of the analysis to the assumed constancy of the Lorenz ratios. It is recognised, however, that as the more backward states start developing, the relative inequality measure may in fact worsen and become progressively similar to those obtaining for the more industrialised states. As a consequence, the reduction in poverty in these states may not be as sharp as projected above. On the other hand, the proportionately higher growth rate in agriculture relative to the growth rate of GDP projected for the perspective period as compared to the historical proportions suggests that the urban-rural consumption differentials may indeed narrow, which would argue for a sharper reduction in poverty than projected. In this context it is interesting to note that an agriculturally developed state like Punjab has lower inequality than the more industrialised Maharashtra. This appears to suggest that the focus that is being placed on agriculture in the Ninth Plan is more likely to be poverty reducing than an industry-based growth strategy. In so far as the constant poverty lines are concerned, it is not possible to project the inflation rates for different states and therefore no statement can be made regarding the impact of differential rates of inflation. Finally, it needs to be mentioned that the observed clustering of the poor around the poverty line and its implication for a faster reduction in the poverty rate has not been taken into account in the above exercise. If this trend continues, the pace of poverty reduction may accelerate even more than projected in this exercise.

1.97 The projected poverty ratios in the rural and urban areas at the national level are presented in Table 1-13. The growth in per capita consumption between 1993-94 and 1996-97 reduces the incidence of poverty to 30.55 per cent in rural reas, 25.58 per cent in urban areas and 29.18 per cent for the country as a whole. The incidence of poverty in the terminal year of the Ninth Plan is estimated as 18.61 per cent in rural areas, 16.46 per cent in urban areas and 17.98 per cent for the country as a whole. By the end of the perpsective period the povery ratio is projected to be less than 5 per cent. The poverty reductions are the impact of consumption growth alone as the inequality in the distribution of per capita household consumption is assumed to remain unchanged during the perspective period. Indeed, since the discrepancy between household and private consumption is projected to increase, there is an implicit inference that the extent of relative inequality may rise to a certain extent. Under the assumptions made, the reduction in poverty is higher in rural areas as compared to urban in the Ninth Plan and also in the perspective period. This arises almost entirely due to the fact that relative inequalities are more pronounced in urban areas as compared to the rural.

Table 1-13 : Projection of National Poverty Ratios
                                                  (percent)
------------------------------------------------------------
Region         	1996-97     2001-02    2006-07     2011-12
------------------------------------------------------------
Rural             30.55       18.61       9.64        4.31
Urban             25.58       16.46       9.28        4.49
Total             29.18       17.98       9.53        4.37
-------------------------------------------------------------

1.98 The projections of statewise and national ratios for the perspective period are based on the assumed stability of two key parameters: (a) the Lorenz ratio; and (b) the per capita urban-rural consumption differential. The rationale for the assumed constancy of these ratios lies in their past behaviour. However, public policy and development strategy can conceptually affect their future behaviour. In order to understand the magnitude of such effects, the sensitivity of the poverty measures to changes in the distributional parameters need to be seen. The impact of a reduction in inequality on poverty worked out through a simulation exercise shows that 5 per cent reduction in the Lorenz ratio of consumption distribution, ceteris paribus, reduces the incidence of poverty by an additional 2 percentage points in the terminal year of the Ninth Plan and reduction in the Lorenz ratio by 10 per cent reduces the incidence of poverty in the Ninth Plan by 4 percentage points. On the other hand, reduction in the urban-rural per capita consumption differential by 5 and 10 per cent reduce the poverty ratio by only 0.2 and 0.4 percentage points respectively. The principal reason for this limited effect is that while rural poverty declines, there is a partial off-set through increase in the urban poverty. For this reason, it may not be desirable to focus excessively on policies to reduce the urban-rural differential except on the basis of distribution of the incremental growth. The national poverty ratios arising from the sensitivity analysis are presented in Table 1-14.

Table  1-14: Projection of National Poverty Ratios:
             Sensitivity Analysis
                                                   (percent)
--------------------------------------------------------------
Scenario              1996-97    2001-02    2006-07   2011-12
--------------------------------------------------------------
1. Scenario I            27.17    16.00     8.01      3.51
2. Scenario II           28.99    17.84     9.45      4.35
3. Scenario III          26.98    15.86     7.93      3.48
4. Scenario IV           25.08    14.02     6.60      2.79
5. Scenario V            28.81    17.71     9.37      4.32
6. Scenario VI           24.69    13.76     6.48      2.75
Base Case                29.18    17.98     9.53      4.37
----------------------------------------------------------------

Notes:
1.Scenario I is based on a reduction of Lorenz ratio by 5%, i.e., from 0.2816 to 0.2675 in rural areas and from 0.34 to 0.323 in urban areas.
2.Scenario II is based on a reduction of per capita urban-rural consumption differential by 5%, i.e., from 62.78% to 59.64%.
3.Scenario III is based on the impact of both Scenario I and Scenario II.
4.Scenario IV is based on a reduction of Lorenz ratio by 10%, i.e., from 0.2816 to 0.2534 in rural areas and from 0.34 to 0.306 in urban areas.
5.Scenario V is based on a reduction of per capita urban-rural consumption differential by 10%, i.e., from 62.78%to 56.50%.
6 Scenario VI is based on the impact of both Scenario IV and Scenario V.

1.99 There are a number of ways by which the reduction in inequality in consumption distribution can be effected. The increased outlay on anti-poverty programmes, if directed towards states with higher poverty and high poverty regions within the states, can lead to an improvement in the inequality parameters and thereby to a faster reduction in poverty than projected. Another option lies in transferring consumption from top to the bottom deciles of the population. A 5 per cent rise in the share of consumption of bottom three deciles of the population at the expense of top 10 per cent means a reduction in the income of the top 10 per cent by 2 to 3 per cent of their consumption share. In a similar way, a 10 per cent rise in the share of consumption of bottom three deciles of the population means a reduction in the income of top 10 per cent of the population by 5 to 6 per cent of their consumption share. It needs to be realised, however, that although the percentage changes in the distributional parameters appear to be fairly small, the magnitudes involved are quite substantial and may involve considerable effort. For example, a 5 per cent and 10 per cent rise in the share of consumption of bottom 30 per cent of the population involve a transfer of 0.75 per cent and 1.5 per cent of the total consumption respectively. This translates into an annual transfer of total income of Rs. 4753 crore and Rs. 9505 crore at 1996-97 prices during the course of the Ninth Plan for the two cases. These transfers are in addition to the existing redistributive measures through anti-poverty and other programmes which are implicitly built into the Ninth Plan. The effect of these transfers would be to reduce the incidence of poverty from 18 per cent in the terminal year of the Ninth Plan in the base-line to 15.9 per cent and 13.8 per cent respectively. It needs to be understood that the income transfer from the top to bottom deciles may not be reflected entirely in reduction of the consumption of the former. A certain portion of the reduction in income would almost inevitably fall on their savings. Thus, such a transfer is likely to result in a reduction of the savings rate in the economy. The magnitude of this shortfall, in itself, may not be a serious problem, provided it is made up through increased efforts towards mobilising additional investible resources.

1.100 These projections of the possible reduction of poverty during the perspective period, though quite impressive, should not be a cause for complacency. In particular, there is no cause to bring into question the utility or size of the anti-poverty programmes at the present time. These programmes, it should be remembered, not only alleviate the immediate deprivation of the poor, but also contribute quite significantly to the creation of productive assets - private assets in the case of schemes like IRDP, and community assets through JRY and EAS. Therefore, the impact of these schemes are embedded in the above analysis both through their direct effect on the consumption of the poor and through the indirect effects arising from the income generated from the assets created. Reduction of these schemes is likely to alter the parametric basis of the analysis and there is no estimate of the magnitude of such changes. More importantly, time series analyses indicate that the existence of these schemes, along with the PDS, have prevented the distribution-sensitive measures of poverty, such as the PG and SPG indices, from worsening during natural calamities. Closure of these programmes would expose the population below or around the poverty line to vulnerabilities which continue to be present in the Indian economy. It should be recognised that starting new programmes during times of necessity is expensive in terms of both human and financial costs and also tends to be inefficient. There is, however, a strong case for better targetting and improving the effectiveness of these programmes.