Socioeconomic Convergence
4.3. Previous evidence of socioeconomic progress and convergence among Indian states
4.5.2. Trend analyses
The assessment of trends in net state domestic product reveals that until 1990-91, the NSDP among the states are more or less same. However, there was a huge divergent trend in post-1991 period. States like Maharashtra, Haryana, Punjab, Gujarat and Tamil Nadu could achieve greater progress while states like Bihar, Uttar Pradesh, Madhya Pradesh and Orissa achieved very less progress. Thus, over the period 1981-2010, the gap between states in terms of NSDP has increased enormously. Though, the decline of poverty ratios across the states was universal during 1974-2005, but the interstate gap in-terms of absolute levels of poverty ratios remain same. However, a tremendous progress has been achieved in terms of literacy rates. The trend line plots show a clear evidence of closing the literacy gap across the states. Even among the states like Uttar Pradesh, Bihar and Madhya Pradesh, the progress was remarkable.
The trends in Gini index show an increase in gap across the major states. This index was calculated based on National Sample Survey Consumption expenditure data. The majority of the states showed an increase in Gini index values. The less developed
76 states show lower levels of Gini value compared to developed counterparts. The state of Maharashtra showed highest inequality in consumption expenditure. Human Development Index, a composite measure of overall development also showed tremendous improvement over the period for all the major states. However, a comparison of the interstate gap between trend plots showed that though the gap was closing, but at a very slow pace (Figure 4.1).
Figure 4.1. Trends in selected socioeconomic indicators during 1971-2011 NSDP (1981-2010)
0 10000 20000 30000 40000 50000 60000
1980-81 1985-86 1990-91 1995-96 2000-01 2005-06 2010
NSDP (in Rupees)
Period
AP Assam Bihar Gujarat HR Karnataka Kerala MP Maharashtra Orissa Punjab Rajasthan Tamil Nadu UP
West Bengal
77
78
79 4.5.3. Catching-up process
In the Solow’s growth framework or the neo-classical growth model, the catching-up mechanism is necessary for convergence. The catching-up process was identified by plotting a scatter diagram for change in an indicator in two points of time against values in the initial period. According to the Solow growth framework, the advanced states experience less change with better values in the initial period. In contrast the laggard states experience a greater change in poorer values in the initial period. In this chapter, I have examined the catching-up process in socioeconomic indicators for the 15 major states of India.
Figure 4.2. Scatter plots of change in selected socioeconomic indicator by values in the initial period of 15 major states of India.
Figure 4.2 presents the scatter plots of change in selected socioeconomic indicators against the mean values in the initial period for 15 major states of India. In case of per capita NSDP, the results showed a positive association between change and initial values of per capita NSDP. This indicates that the catching-up process was not yet evident from the laggard states in case of the per capita NSDP. The economically advanced states (e.g. Punjab, Haryana, Maharashtra, Gujarat and Karnataka) in India
9.51010.5 11
6.8 7 7.2 7.4 7.6 7.8 Log SNDP in 1981
1020304050
30 40 50 60
Poverty ratios in 1973-74
152025303540
Change in literacy rate (2011-1981)
30 40 50 60 70 80
Literacy rates in 1981
-.1-.05 0
.05
.25 .3 .35
Gini Index values in 1983
.05 .1.15 .2.25
Change in HDI (2006-1981)
.25 .3 .35 .4 .45 .5
HDI in 1981 Log NSDP in 1981
Change in log NSDP (2010-1981)
80 have continued to experience higher growth rate in per capita NSDP compared to the laggard states (e.g. Bihar, Uttar Pradesh, Odisha, Rajasthan and Madhya Pradesh). A similar pattern was also evident for other two economic indicators: poverty ratios and Gini index. A change in both these indicators was positively associated with values in the initial period which implies that the reduction in poverty and economic inequality (in terms of Gini index) in some states was much greater in comparison to other states.
The catching-up process in terms of progress in reduction of poverty and economic inequality from laggard states of India was not evident as yet.
However, the results in case of literacy rate indicate a contrasting picture. The states (e.g. Rajasthan, Bihar, Uttar Pradesh, Andhra Pradesh, Madhya Pradesh and Odisha) with low literacy rates in the initial period are experiencing greater change in literacy rates compared to their counterparts (e.g. Kerala, Tamil Nadu, Karnataka, Gujarat, Maharashtra) with higher literacy rates in the initial period. This is an indication of the catching-up process in terms of progress in literacy rates in India by laggard states to leading states. Similarly, there was also clear evidence of the catching-up process in a composite measure of human development. The results revealed a greater progress in human development index in laggard states (e.g. Bihar, Uttar Pradesh, Madhya Pradesh Rajasthan, and Odisha) compared to advanced states (e.g. Kerala, Maharashtra, Gujarat, Punjab, Haryana, Tamil Nadu, Karnataka) during 1981 to 2006.
4.5.4. β-convergences
Though, the catching-up process provides clues about convergence, the real convergence mechanism can be identified only by appropriate convergence models.
Therefore, this chapter used β-convergence models to test the convergence hypothesis for the progress of socioeconomic indicators.
Absolute β-convergence
Table 4.2 presents the results of absolute β-convergence model for log per capita net state domestic product among the major states during 1981-2010. Though, the results revealed that during the period of 30 years, the per capita NSDP showed absolute β -Convergence (β = -0.0138) across the 15 major states, but the estimate was statistically not significant and adjusted R2 was very low. Moreover, the results of piece-wise
81 convergence models (estimated for sub-periods in a long period) were evidently a mixed pattern. During the initial period, 1981-91, the results showed a clear divergence phase (β = .07587) in the NSDP. In contrast, in 1991-01, the earlier phase of divergence was replaced with convergence (β = -.25090). Moreover, the volume of convergence was high. For the recent period, convergence phase in the previous period was replaced by divergence (β = .12591; p<0.032) and the estimate was statistically significant.
Table 4.2. Absolute β − convergence model estimates for per capita NSDP of the major Indian states, 1981-2010.
Period 𝛽 coefficient P value Adjusted R2 n df
1981-2010 -.0138 0.877 0.0749 15 14
1981-1991 .07587 0.576 0.0503 15 14
1991-2001 -.25090 0.120 0.1123 15 14
2001-2010 .12591 .322 0.0041 15 14
Constanta 1.278 0.070
Note- n=sample, df=degree of freedom, a= constant value refers to a convergence model of period, 1981-2010.
Table 4.3. Absolute β − convergence model estimates for Poverty Ratios of the major Indian states, 1974-2005.
Period 𝛽 coefficient P value Adjusted R2 n df
1974-2010 .06291 0.036 0.2424 15 14
1974-1984 .10874 0.008 0.3809 15 14
1984-1994 -.00388 0.908 0.0758 15 14
1994-2005 .12538 0.019 0.3058 15 14
Constanta -5.82597 0.001
Note- n=sample, df=degree of freedom, a= constant value refers to the convergence model of period, 1974-2005.
Absolute β-Convergence model estimates for poverty ratios for 15 major states during 1974-2010 are presented in table 4.3. The results revealed that in the long run, the reduction of poverty ratios across the 15 major states were diverging (β =0.06291, p<0.036). However, the results of convergence model estimates for sub-periods during 1974-2010 again indicated a mixed pattern. The reduction of poverty rates across the states during 1974-84 showed divergent progress (β =0.10874; p<0.008), but followed by a convergence phase during 1984-94 (β = -0.00388; p<0.908) and again divergence (β = 0.12538) during 1994-2005. Though, the results showed progress of convergence
82 in the reduction of poverty ratios during 1984-94, but the model was not statistically significant and adjusted R2 value showed the poor goodness of fit. However, the absolute β-convergence estimates for the entire period, 1974-2005 and other sub-periods, 1974-84 and 1994-2005 showed statistically significant divergence in the reduction of poverty across 15 major states.
Table 4.4. Absolute β − convergence model estimates for literacy rates of the major Indian states, 1981-2011.
Period LR
𝛽 coefficient
P value Adjusted R2 n df
1981-2011 -.04468 0.000 0.9408 15 14
1981-1991 -.03158 0.000 0.7334 15 14
1991-2001 -.07056 0.000 0.7079 15 14
2001-2011 -.05189 0.000 0.6777 15 14
Constanta 3.8670 0.000
Note- n=sample, df=degree of freedom, a= constant value refers to a convergence model of period, 1981-2011.
Table 4.5. Absolute β − convergence model estimates for human development index values of the major Indian states, 1981-2006.
Period 𝛽 coefficient P value Adjusted R2 n df
1981-2006 -6.3717 0.000 0.7627 15 14
1981-1991 -4.4534 0.020 0.3020 15 14
1991-2001 -3.3580 0.006 0.4042 15 14
2001-2006 -9.5074 0.000 0.5951 15 14
Constanta 3.6714 0.000
Note- n=sample, df=degree of freedom a= constant value refers to the convergence model of period, 1981-2010.
Table 4.4 presents absolute β-convergence model estimates for literacy rates, for major states of India. The results suggest a strong convergence phase (β = -0.04468, p<0.00) in literacy rates during 1981-2011. The results of piecewise convergence model estimates showed that the estimates for all the sub periods support convergence.
However, the convergence process was stronger in recent two decades (β = -.07056, p<0.00 during 1991-2001 and β = -.05189, p<0.00 during 2001-2011) compared to the earlier initial period, 1981-1991 (β = -.07056, p<0.00). The absolute β-convergence model estimates for literacy rates showed high goodness fit with the adjusted R2 value greater than 0.90. The absolute β-convergence model estimates for human development
83 index values of the major states of India are presented in table 4.5. The results for the entire period, 1981-2006 (β = -6.3717, p<0.00) and for all the sub periods indicated a strong absolute β-Convergence in human development across the major states of India.
However, the volume of convergence was greater in the recent period, 2001-06 (β = -9.5074, p<0.00) compared to the initial period, 1981-1991 (β = -4.4534, p<0.00). The β-Convergence model estimates for human development index values were statistically significant and goodness fit for the model was also high.