This column attempts to widen the ongoing growth rates-based debate on ‘Gujarat vs. rest of India’ by ranking Indian states on prices, cost of living, household expenditures and inequality, which measure how well states have truly fared in the past two decades. It highlights the spatial differences in terms of these indicators, and finds that prices vary across states at any given point in time. Inequalities have risen significantly in recent years, though there are sharp differences across states. It also shows that Gujarat has always ranked highly in terms of living standards.
In the run up to the election in India and, more so, in the middle of it, with the strong possibility of a Modi-led government in Delhi, economists have spent a good deal of time and energy examining the performance of Gujarat’s economy under Narendra Modi vis-à-vis the rest of India (Nagaraj and Pandey 2013, Ghatak and Roy 2014), and the comparative performance of the Indian economy under the United Progressive Alliance (UPA) and National Democratic Alliance (NDA) regimes (Ghatak, Ghosh and Kotwal 2014)1. Though these studies did not deal exclusively with Gujarat nor did they base their evidence entirely on growth rates, there was an obsession with either or both in these papers. Never before has the prospect of a change of regime in Delhi generated such a heated debate among economists on such a narrow range of issues. As this debate rages, it is not entirely clear to me what purpose is being served by trying to associate (or deny) the spike in Gujarat’s growth rate with the Modi regime or by comparing the growth rates under UPA and NDA regimes.
A positive externality of this debate, however, is that it points to the need to extend the evidence and discussion on spatial differences in India to involve states, population sub-groups and indicators beyond Gujarat, Bihar and growth rates. My objective in this column is to present such evidence that shows India to be a much more diverse country with regionally disparate economic experiences than is apparent from the recent growth rate-based debates on Gujarat vs. rest of India or UPA vs. NDA regimes. The evidence relates to spatial differences in prices and cost of living, real household expenditures and inequality. I report evidence based on household unit records from the National Sample Surveys (NSS) which, notwithstanding their weaknesses, are a good source of information on ‘household real expenditure’2, which is a better measure of the standard of living of households than growth rates that are based on aggregate national or state statistics that have been used in the recent debates.
In an earlier I4I column entitled, ‘Economic Growth versus Social Development: the spatial dimension’ (Ray 2013), I had provided evidence of spatial differences in the performance of regions based on social indictors and, moreover, showed that regional rankings based on money-metric indicators such as per capita expenditure are associated very weakly, if at all, with social indicators such as child health and neonatal, infant and child mortality rates. In this column, I extend that evidence on spatial differences in India to provide evidence at the level of individual states, and based on a wider set of economic indicators. The evidence reported below is based on NSS rounds 50 (1993-94), 55 (1999-2000), 61 (2004-05) and 66 (2009-10).This is a long enough period that covers both NDA and UPA regimes, besides of course the Congress regime of the early 1990s that set in motion the process of economic reforms.
Spatial differences in prices and real expenditure3
Table 1 presents estimates of spatial price indices4 in each of the four NSS rounds. An estimate of spatial price for a state that is significantly greater than one implies that the state is more expensive than the country as a whole, and vice versa if the estimate is less than one. Table 1 contains widespread evidence of spatially different prices in India in each NSS round and in each sector. One might expect affluent states such as Gujarat, Haryana and Punjab with considerable spending power to be expensive to live in, while a much poorer state such as Bihar to be inexpensive based on food prices. There is, however, no evidence of any strong link between a state’s economic affluence and its cost of living relative to the rest of India. For example, rural Kerala which is not an affluent region has been very expensive vis-à-vis the rest of rural India. This table also records considerable movement in spatial prices over this period covering the Congress, NDA and UPA regimes, and shows, moreover, that the picture is not always the same between rural and urban areas.
Table 1. Spatial price indices, state-wise (All India=1.0): Rural and Urban, based on 11 food items
|States||Rural Price Indices||Urban Price Indices|
|50th Round||55th Round||61st Round||66th Round||50th Round||55th Round||61st Round||66th Round|
Source: Majumder, Ray and Sinha (2014).
Tables 2 and 3 present indices of nominal and real spatial expenditures in rounds 50 and 66, respectively, with all India based at 1.0. The expenditure indices are moderately, but not unduly, sensitive between the nominal and spatially deflated real expenditures. The latter takes note of the fact that the purchasing power of the Rupee varies across the states because prices vary across states. Both tables contain evidence of spatial variation in expenditures across states, and the picture is weakly consistent with our expectations, namely that states that have recorded higher growth rates should record higher real expenditures (real expenditure index greater than one). Note, however, that Gujarat which has attracted much attention recently, figures as an ‘average’ state, with a spatial expenditure index quite close to one in both rural and urban areas and in both rounds 50 and 66. In fact, once the higher cost of living in Gujarat relative to most other states is taking into account (see Table 1) via the construction of the spatial price deflated real expenditure index, Gujarat ranks below the all-India average in both rural and urban areas and in both the NSS rounds (50, 66). In contrast, Kerala, Madhya Pradesh, Haryana and Punjab record real expenditure indices that are above one in NSS round 66. Of these, the last two are generally recognised as the more affluent states. Note, however, that, as reported below, the picture changes considerably, with Gujarat ranking well above most other states once we take into account the varying consumer preferences across states. A comparison of Tables 2 and 3 suggests that there has been some convergence across states on their food spending. In other words, the disparate growth performances of the various states over this period did not translate into large differences in their food spending levels. This suggests that the superior growth performances of states such as Gujarat, Haryana and Punjab did not make much of a dent on issues of nutrition and food security that continue to prevail throughout the country.
Table 2. Spatial expenditure indices, state-wise (All India=1.0): NSS 50th Round (1993-94), 11 food items
|States||Rural Income Index by State||Urban Income Index by State|
|Nominal||Real (Spatial price deflated)||Nominal||Real (Spatial price deflated)|
Source: Majumder, Ray and Sinha (2014).
Table 3. Spatial expenditure indices, state-wise (All India=1.0): NSS 66th Round (2009-10) , 11 food items
|States||RuralNominal||Rural: Real (Spatial price deflated)||Urban Nominal||Urban: Real
(Spatial price deflated)
Source: Majumder, Ray and Sinha (2014).
How do (food) spatial real expenditure indices in the four NSS rounds translate into state rankings over this period? To answer this question, one cannot simply rank the states on the basis of the spatial expenditure indices, since these do not take into account the difference in food preferences across the states. In the culturally varied context of the different states in India, such differences are quite large in case of many food items. Following the procedure outlined in Sen (1976), that takes note of differences in dietary habits, the state rankings in each of the four NSS rounds have been reported in Majumder, Ray and Sinha 2014 (Figures 3A,3B)). The figures reveal several cases of changes in state rankings over the period spanned by the four surveys. They also reveal several rural-urban differences. Overall, there are no major changes in the state rankings over the period between 1993-94 and 2009-10. Note, also, that Gujarat, on which the recent debates have focussed, has always been at or near the top throughout this period - there does not seem to have been a ‘Modi effect’. A comparison with similar rankings reported in Sen (1976; Figure 3) for a much earlier period brings out several similarities and some sharp differences. For example, Kerala was ranked near the bottom in Sen’s rankings based on NSS rounds 16 (1960-61) and 17 (1961-62), but it has moved up sharply to be at or near the top during the period covered here. The feature in Sen’s (1976) rankings that stands out particularly sharply, and reiterated in this column, is that even in the early 1960s, Gujarat was ranked almost near the top on levels of living criterion, and this is consistent with the growth rates-based evidence that has been reported recently. In fact, Gujarat occupied (almost) the same rank in Sen’s study as it does in Majumder, Ray and Sinha (2014). Once again, there does not seem to be a ‘Modi effect’.
Spatial differences in inequality and their changes5
The above discussion has documented spatial differences in prices and highlighted the need to incorporate them in cross-state welfare comparisons in the Indian context. Another such area is comparison of inequality, and their movement, across states in India during this period. While much of the attention has focussed on poverty, inequality has slipped below the radar. The evidence for rural India is presented in Table 4 which reports the Gini inequalities6 (state-wise) over this period.
Table 4. State-specific and All India Gini Coefficients (Nominal and Temporal Price Deflated): Rural sector
|State||Gini Coefficient (nominal)*||Gini Coefficient:
Temporal Price Deflated7
(with respect to 55th Round)
|Within a state all households face the same price (nominal)||Within a state all households within a quartile face the same price (real)|
|55th Round||61st Round||66th Round||61st Round||66th Round|
|ALL INDIA (Rural)||0.222||0.215||0.290||0.235||0.288|
* The ´nominal´ and ´temporal price deflated´ Gini Coefficients are the same for the 55th round.
Source: Chakravarty, Majumder and Ray (2014).
The following features are worth noting. First, there is considerable variation in the magnitude of inequalities across states. This is true of both nominal and real expenditure inequalities8. Second, while in most states, inequalities were static or even recorded a decline between 1999-2000 and 2004-05, there was a sharp increase in inequality, in both nominal and real terms, in most states during the second half (2004-05 – 2009-10). The increase in inequality was particularly large in case of Kerala and Punjab making them two of the most unequal states in India at the end of our sample period. While the sharp increase in case of Kerala is possibly due to the increased inflow of remittances from the Gulf that favoured some households over others, the inequality increase in Punjab reflects gains for large farmers that benefitted from growth-enhancing reforms and the large subsidies for diesel and fertilisers. The increase in inequality in nearly all the states during the period, 2004-05 – 2009-10, is reflected in the sharp increase in inequality recorded by the All-India figures in both nominal and real terms. Third, and most significantly, neither the magnitude nor the direction of change in inequality over the two sub-periods is identical for all the states nor is it robust between nominal and real expenditure inequality. For example, in Gujarat, while nominal inequality increased sharply during the period between NSS rounds 61 and 66, real expenditure inequality declined. In Haryana, while there was a sharp increase in nominal inequality over this sub-period, real expenditure inequality remained unchanged.
The evidence presented in this column draws attention to spatial differences between states in India in food prices, real expenditure levels, and inequalities. It attempts to widen the recent macro growth rates-based debate on ‘Gujarat versus the rest of India’ by ranking all states, not just Gujarat and Bihar, on the basis of other indicators that are better measures of how well a state has fared during this period, 1993-94 – 2009-10. A result of some significance in the context of the current debate is that Gujarat was always ranked highly on the basis of living standards throughout this period. This was also the case with the rankings reported by Sen (1976) for a much earlier period. Another result is the sharp increase in inequality in all the states during the period between the NSS rounds 61 (1999-2000) and 66 (2009-10); the latter points to the need to reduce our obsession with growth rates and poverty levels, and focus more on inequality. The evidence on the latter is much less rosy than the first two, and deserves more attention. Even here, the sharp divergence across states on inequality magnitudes in 2009-10 needs to be addressed in policy formulation.
In this column, the author draws on his published and ongoing research with Manisha Chakravarty (Indian Institute of Management (IIM), Kolkata), Amita Majumder (Indian Statistical Institute (ISI), Kolkata), and Kompal Sinha (Centre for Health Economics (CHE), Monash University).
- See, also, the critique of these studies by Dholakia (2014) and Marjit (2014).
- A household’s real expenditure indicates the ability of the household to purchase goods and services at the prevailing prices. Real expenditure is obtained by deflating (dividing) nominal household expenditure by the price index, where nominal household expenditure is the actual, money value of household expenditure.
- This section is based on Majumder, Ray and Sinha (2014). The reader is referred to that paper for technical and other details.
- The data on prices was constructed from unit values of the various food items obtained from information on expenditure and quantity purchased of these items by each household, available in the various NSS surveys.
- This section draws on Chakravarty, Majumder and Ray (2014).
- The Gini coefficient is the most widely used measure of inequality. It lies between 0 (complete equality) and 1 (complete inequality). A higher Gini signifies greater inequality, and a smaller Gini means lesser inequality.
- This implies that the Gini coefficient for a particular state takes into account price movements over time in the state.
- As explained earlier, the former does not take into account differences in the purchasing power of the Rupee across states, the latter does.
- Chakravarty, M, Majumder, A, and R Ray (2014), ‘Preferences, Spatial Prices and Inequality’, Monash University, Economics Discussion Paper No. 52-12 (revised).
- Dholakia, R. H. (2014), “Relative Growth Performance of Gujarat: A Comment on the Recent Debate”, Economic and Political Weekly, Vol. XLIX, No. 18, 75-77.
- Ghatak, M, Ghosh, P, and A Kotwal (2014), “Growth in the Time of UPA: Myths and Reality”, Economic and Political Weekly, Vol. XLIX, No. 16, 34-43.
- Ghatak, M, and S Roy (2014), “Did Gujarat’s Growth Rate Accelerate Under Modi?”, Economic and Political Weekly, Vol. XLIX, No. 15, 12-15.
- Majumder, A, Ray, R, and K Sinha (2014), “Spatial Comparisons of Prices and Expenditure in a Heterogeneous Country: Methodology with Application to India”, forthcoming in Macroeconomic Dynamics.
- Marjit, S (2014), “On UPA vs NDA; Gujarat vs Rest of India: Myth and Reality”, Economic and Political Weekly, XLIX, No. 18, 82-83.
- Nagaraj, R, and S Pandey (2013), “Have Gujarat and Bihar Outperformed the Rest of India?”, Economic and Political Weekly, XLVIII, No. 39, 39-41.
- Ray, R (2013), ‘Economic Growth versus Social Development: the Spatial Dimension’, Ideas for India, 31 July, 2013.
- Sen, A (1976), “Real National Income”, Review of Economic Studies, 43 (1), 19-39.