It is by now well documented that the trends that emerged in India in the 1990s with respect to the supply of rural credit in general, and agricultural credit in particular, were disturbing. In the 1990s, there was (a) a decline in the number of commercial bank branches in rural areas; (b) a widening of inter-State inequalities in credit provision, and a fall in the proportion of bank credit directed towards regions where banking was historically underdeveloped; (c) a sharp fall in the growth of credit flow to agriculture; (d) increased sidelining of small and marginal farmers in the supply of agricultural credit; (e) increased exclusion of the disadvantaged and dispossessed sections of the population from the formal financial system; and (f) a strengthening of the hold of moneylenders on rural debt portfolios (for details, see Shetty 1997; Ramachandran and Swaminathan 2005; Chavan 2005, 2007).

This note focuses on one specific aspect of the discussion on rural banking in the 1990s and after: the decline in the number of rural bank branches. We discuss the methodological issues related to the official data on rural bank branches, and try to address the question of whether the decline in the number of rural branches can necessarily be interpreted as closure of rural branches.

Sources of Data on the Number of Rural Bank Branches

Data on rural bank branches in India are available primarily from the Reserve Bank of India’s annual publication titled “Basic Statistical Returns of Scheduled Commercial Banks in India.” In this publication, data on bank branches are given for four groups of settlements: rural, semi-urban, urban, and metropolitan. These groups, however, are defined differently by the Reserve Bank of India (RBI) and the reports of the Census of India. The RBI definitions are as follow: (i) “rural” includes all centres with populations of less than 10,000; (ii) “semi-urban” includes centres with populations of 10,000 and above, but less than 1,00,000; (iii) “urban” includes centres with populations of 1,00,000 and above, but less than 10,00,000; and (iv) “metropolitan” includes centres with populations of 10,00,000 and more. Thus, population is the criterion that determines whether a region is rural or semi-urban. This is the reason why the RBI calls these groups of centres “population groups.” The RBI uses Census of India data to arrive at this classification of centres.

On the other hand, when a Census of India report refers to a place as “rural” or “urban,” the connotations are different. In the Census reports, an urban region is defined as follows: (a) all statutory places with a municipality, corporation, cantonment board or notified town area committee, etc.; (b) a place satisfying the following three criteria simultaneously: (i) a minimum population of 5,000, (ii) at least 75 per cent of the male working population engaged in non-agricultural pursuits, and (iii) a density of population of at least 400 per square kilometre. The Census of India does not classify any region as “semi-urban” or “metropolitan”.

In sum, the definitions of rural and urban centres in the RBI reports are different from the definitions of rural or urban areas in the Census of India. The two are not comparable, even though the terms sound similar.

There is an additional, methodological issue in using the RBI data on bank branches by population groups across time. The RBI periodically updates the population group classification using the latest available Census reports, thus creating problems of comparability across years. Until 1994, the RBI used Census of India 1981 figures to classify centres into population groups. From 1995 onwards and until 2005, Census of India 1991 figures were used to classify centres into population groups. From 2006 onwards, the RBI began to use Census of India 2001 figures to classify centres into population groups. In other words, if we need comparable data on population groups, we have to compare data between 1991 and 1994, and then between 1995 and 2005. Between 1994 and 1995, a large number of centres may have moved from “rural” to “semi-urban” classification due to population growth. As a result, the number of centres used to count the number of bank branches in 1995 may be smaller than in 1994, resulting in the return of a smaller number of bank branches. The problem is similar when we compare data for 2005 and 2006. In short, any direct comparison of the number of rural bank branches across a period of time involves comparing different numbers of centres based on different Census classifications, and is thus incorrect methodologically.

Changes in the Number of Rural Bank Branches

We now look at the data on the number of bank branches after accounting for definitional changes (Table 1). Between 1991 and 1994, the number of rural bank branches rose by 123 – from 35,206 to 35,329. Between 1995 and 2005, the number of rural bank branches fell from 33,004 to 32,082 – a reduction of 922 branches.

The question to follow up, then, is the geographical distribution of the 922 branches that were closed between 1995 and 2005. However, there is a methodological difficulty in examining the change in the number of rural branches across States. In the Basic Statistical Returns (BSR), the sum of the number of rural branches across States does not match with the total number of rural branches in the country as a whole (given in Table 1). The reason for the difference is that the sources from which the RBI compiles data on these variables are different. State-wise figures for the number of branches are compiled from the “Master Office File” of the BSR. On the other hand, the national aggregate for the number of branches is based on two different sets of returns, titled “BSR-1” and “BSR-2”. The “Master Office File” of the BSR provides information on the total number of branches that actually existed in each State at the end of March each year. On the other hand, the data from “BSR-1” and “BSR-2” pertain only to the number of branches that reported data for that year to the RBI.

State-wise data on the number of rural bank branches are provided for comparison in Table 2. Clearly, the extent of decline in the number of rural branches between 1995 and 2005 varied across regions and States (Table 2). Between 1995 and 2005, the largest decline in the number of bank branches took place in the Central region of India; 331 rural bank branches closed down in this region, and of them 228 branches were in Madhya Pradesh alone. The Central region was followed by the Southern region; 294 branches closed down in the Southern region, of which 131 branches were in Tamil Nadu. Similarly, of the 212 branches that closed down in the Western region, 113 branches were in Maharashtra.

However, between 2005 and 2008, there was a shift in the trend of closure of rural bank branches that was observed between 1995 and 2005.¹ The number of rural bank branches increased by 430 between 2005 and 2008; there were 30,646 rural bank branches in 2005 and 31,076 in 2008 (Table 1). The shift in the trend after 2005 came entirely from a sharp increase in the number of rural bank branches between 2007 and 2008. Between 2007 and 2008, the number of rural bank branches increased by 525, more than compensating for the fall between 2005 and 2007.

While the increase in the number of rural branches between 2007 and 2008 took place in all the regions of India, it is clear that the increase was largely concentrated in the Central and Southern regions. Of the additional 505 rural branches that opened in India between 2007 and 2008, about 24 per cent were located in Uttar Pradesh alone.

The addition of 430 new rural branches after 2005 is indeed a positive development. However, the fact remains that there was a total loss of 922 rural branches in the 10-year period between 1995 and 2005. In addition, there is evidence that the new bank branches were not necessarily opened in the regions where branches were closed down.

Between 2005 and 2008, the total number of “banked” rural centres (i.e. centres served by at least one bank branch) in India fell from 28,643 in 2005 to 28,281 in 2008.² This was because the majority of the rural centres were single-branch centres and a closure of the only branch essentially made these centres “unbanked”.³ At the same time, the number of banked centres increased after 2005 for other population groups – semi-urban, urban, and metropolitan. Thus, a loss of 922 rural branches between 1995 and 2005 cannot be directly equated with, or set off against, a gain of 430 new branches, since, in many centres, the loss of one branch has meant the loss of the only available branch.⁴

It is not adequate, however, to limit the discussion on rural penetration of banks only to the change in the number of rural bank branches. In order to understand the access to rural banking services, we need to normalize the change in the number of rural bank branches with respect to changes in the rural population.

As discussed earlier, given the difference in the definition of “rural” in the Census of India and the BSR, it is difficult directly to compare changes in the number of rural branches with changes in the rural population. We have attempted a small exercise to compare the rates of growth of rural branches and of rural population across regions and States (Table 3). Between 2006 and 2008 – the period of increase in the number of rural branches – the ratio of the rate of growth of rural branches to the growth in rural population was less than 1 across regions and States. Even in the Central region, where the increase in the number of rural branches was the sharpest between 2006 and 2008, the ratio of growth of rural branches to that of rural population was only 0.3.

In sum, even though there has been an increase in the number of rural branches in the country in recent years, the growth of rural branches has not been able to keep pace with the growth in rural population, thus indicating a decline in banking penetration in rural areas.