Anand G. Chandavarkar
The phenomenon of urban bias in LDCs, in the sense of a general tendency for economic activity and the benefits of development to concentrate in the urban areas, has been studied mostly in real terms without due cognizance of its financial aspects which are no less important. This article attempts to analyze whether the hypothesis of urban bias in developing countries, advanced by Michael Lipton and other economists, is also valid in its financial aspects. It evaluates its nature, implications, and possible correctives in the light of the available empirical evidence for some selected Asian countries such as India and Thailand.
The term “urban” could be defined, broadly, as in national population censuses or in national banking statistics, to mean a certain minimum level of population in a continuous geographic area, with the remainder being classified as semi-urban or rural areas. For instance, the Indian banking statistics define rural centers as places with a population of up to 10,000; semi-urban centers, between 10,000 and 100,000; urban centers, between 100,000 and 10 million; and metropolitan centers as those with a population of over 10 million.
Broadly speaking, financial “urban bias” connotes a net transfer of financial resources from rural to urban areas which is not justified by social returns and which reflects failures of rural financial markets. Thus, bias would cover cases where, for instance, governments appropriate rural savings for inefficient urban or industrial investment, or where institutional or political factors drive a wedge between private and social returns resulting in excess flows to cities. The phenomenon of financial urban bias is in a sense implicit in the more familiar center-periphery economic syndrome, with the “center” being coterminous with urban (developed) areas and the “periphery” with the rural (underdeveloped) areas. Admittedly, backward areas exist in every type of economy, from the most affluent to the poorest, and from the most centrally planned to the most laissez-faire as, for instance, the Appalachian region of the United States, the Mezzogiorno in Italy, the Uzbek Republic of the USSR, parts of Southwest France, and the Montenegrin region of Yugoslavia. But the regional problem is accentuated in the LDCs by even greater interregional differentials of capital, income, and welfare.
The financial dualism of LDCs has been analyzed to date almost exclusively in terms of the divide between the organized and the unorganized sectors. It is therefore pertinent to analyze in what sense is there a financial urban bias which could be said to be detrimental to economic efficiency and welfare and how to determine the extent to which authorities should then try to redress such a bias.
Financial urban bias is in the nature of a portmanteau concept which may reflect the following elements: (1) higher effective rates of interest in rural areas because of a shortage of collateral, among other things, and the resultant higher risk premium; (2) greater credit rationing in rural areas even if there is willingness to pay higher rates (i.e., the fringe of unsatisfied borrowers is larger); (3) higher average credit/deposit ratios in urban areas relative to rural areas; and (4) concentration of net deposit bank offices (local deposits exceed local credits) in rural areas whereas net credit bank offices (local credits exceed local deposits) tend to concentrate in urban areas.
Financial dualism as reflected in the uneven distribution of credit offices of banks and the resultant transfer of rural savings into urban investment is the least discussed and documented aspect of financial urban bias, whereas the differentials between low urban loan rates and high rural loan rates are familiar ground. Credit rationing, although known to be more significant in rural than in urban areas, is hard to quantify. A typical expression of financial urban bias is the observation by Gunnar Myrdal that “in many countries . . . the banking system, if not regulated to act differently easily becomes an instrument for siphoning off the savings from the poorer regions to the richer and more progressive ones where rates on capital [i.e. the rates of profit] are high and secure.” Such statements occur frequently in the literature but are not backed by empirical evidence.
It is difficult to present systematic empirical evidence in support of the hypothesis of financial urban bias in LDCs. The conventional flow-of-funds tables classify sectors into government, business, and households. But even these, which in any case are available for only a few LDCs, do not disaggregate the flows between rural and urban sectors on the lines, for example, of the periodic regional flows-of-funds table compiled by the Banque de France. Generally, there is comparatively little information on the territorial aspects of financial intermediation even in developed economies, and most basic relationships, like regional credit multipliers, still await systematic testing in the LDCs. The only evidence of financial urban bias in the discussions of Lipton’s hypothesis is the data that suggest that net cash inflows into the rural sector are a very small proportion of rural incomes in typical villages in most LDCs. This is not sufficient to substantiate financial urban bias, since such inflows are only a very small part of total financial flows between rural and urban areas which take place through a variety of budgetary and banking channels, including credit cooperatives and post offices, as well as through informal channels such as family remittances and indigenous bankers. These are not fully recorded, which explains the difficulty of providing net aggregate figures for all such flows in LDCs.
However, a bank’s network of offices shows a clear division between the net credit and net deposit branches. This situation is also encountered in unit banking systems wherein rural banks may have surplus deposits, without adequately remunerative local credit outlets, which are therefore lent to banks short of loanable funds. Data relating to the distribution of bank branches, as between net credit and net deposit centers, and their respective credit/deposit ratios, are usually available to central banks (or can be collected fairly easily) but are seldom published. A plausible reason for this may be the possible political sensitivity to publication of data on the “drain” of savings from undeveloped rural regions for investment in developed urban areas, which might conceivably generate feelings of “internal colonialism.”
The available evidence for India and Thailand cited in this article shows that urban areas tend to be net users of commercial bank credit, and there is strong presumption that this is also characteristic of other developing countries, albeit in varying degrees. Thus, in Indonesia the highest concentration of credit is in the more urbanized regions of Jakarta Raya and Central and East Java, whereas there is a comparatively even spread of deposit centers. In Malaysia, the net credit centers tend to cluster in the more urbanized areas of Kuala Lumpur, Petaling Jaya, and Penang; and in the Philippines, in the greater Manila area. A classic example is Thailand where the Bangkok Thon Buri metropolitan area is a consistent net user of credit, whereas the rest of the economy, which is predominantly rural, is a net source of deposits (Table 1).
(Weighted mean by region)
Excluding Bangkok Thon Buri.
Excluding Bangkok Thon Buri.
In India, financial urban bias is evidenced by the pattern of total interregional financial flows, covering commercial and development banks, the Life Insurance Corporation, and the Unit Trust, as well as the regional distribution of credit/deposit ratios (Table 2). An authoritative study by the National Institute of Bank Management concluded that the interstate distribution of institutional finance in India showed a pronounced regressive bias against the low-income states, which are predominantly rural. The regressiveness was also reflected in the pattern of bank finance and term lending to industry. Other studies show that the four largest cities, Bombay, Calcutta, Delhi, and Madras, account for about 85 percent of the total transfer of credit within the country. In India, and similarly placed countries, the urban bias of the banking system is also accentuated by the nature of the post office savings banks which only collect savings, mostly from rural areas, but do not extend credit.
(In rupees crores 1 as of last Friday of each month)
|June 1969||June 1981||June 1982 2||Increase (June 1969|
to June 1982)
|Percent of total||3||2||13||11||14||12||15||14|
|Percent of total||22||11||23||17||23||17||23||18|
|Percent of total||26||20||25||23||25||23||25||23|
|Percent of total||49||67||39||49||38||48||37||45|
Rupees 1 crore = 10 million.
The data for June 1982 are provisional.
Rupees 1 crore = 10 million.
The data for June 1982 are provisional.
Such movements from deposit areas (predominantly rural) to net credit areas (largely urban) have also been noticed in the evolution of developed economies. For instance, the elaborate nineteenth century correspondent relationships between country banks and London banks in England are said to have served an efficient mode of transfer of loanable funds from surplus rural areas to deficit urban areas. Even in a highly developed country like the United States a comparable credit phenomenon (“redlining”) is reported to exist. This relates to the practice of banks who demarcate certain areas, and restrict mortgage and other credit to their residents who are regarded as high-risk customers. This led the U.S. Congress to pass the Community Reinvestment Act in November 1978, which requires financial institutions to “demonstrate that their deposit facilities serve the convenience and needs of the communities in which they are chartered to do business.”
The available evidence, supported by general observation of the banking systems, is at least supportive of the hypothesis of financial urban bias, whatever its precise magnitude. But this phenomenon has also to be viewed in proper perspective. First, deposits are only one form of savings; therefore excess deposits are not necessarily the same as excess savings. The overall movement of savings is related to the net trends in the interregional balance of payments and not necessarily to the position of excess deposits and excess loans of bank offices.
Second, to some extent interregional transfers of deposits might genuinely reflect the absence of suitable local rural lending outlets, as well as the fact that branch managers may not have power to sanction loans above specified, and usually low, limits. Moreover, the concentration of credits in the urban branches might also occur because customers often borrow through them while holding their deposits at rural offices.
Finally, and most important, a financial urban bias could be said to reflect the effects of a market-oriented financial system in which bankers tend justifiably to treat the bank as one unit and are therefore free to apply the criterion of maximum return in allocating deposits. Since the private rate of return on investment usually tends to be higher in the urban areas, the bulk of rural deposits is naturally transferred to urban areas. This, of course, does not mean that only private returns tend to be higher in urban areas. Banks do not regard branches that are “deposit” centers as losing branches so long as deposits can be transferred and used in credit centers, whose earnings would cover the cost of the “deposit” center. This means that the banks are performing an essential transfer service between the surplus and deficit areas which, in fact, is one of the important reasons for the growth of branch banking. To some extent financial urban bias is not only unavoidable but economically efficient, given the greater credit-intensity of urban economic activity and the higher rates of return on urban credit and investment.
Redressing financial urban bias
The basic policy issue is thus not the financial urban bias as such, but the extent to which it further accentuates regional economic disparities. Financial urban bias may be supply-determined, insofar as it originates from the portfolio behavior of lending institutions, or demand-determined, insofar as it reflects deficiencies of branch demand, or both. But there are no a priori criteria for the optimal regional allocation of savings and credit, which has to be viewed in the overall developmental strategy of a country. Thus one country may even aim at a net financial flow to urban areas to finance diversification of the economy, whereas others may want a reverse flow for the same reason. An uncritical earmarking of local savings for local credit would lead to an uneconomic fragmentation of the national capital market and a suboptimal allocation of aggregate savings. The earmarking of specific revenues for specific items of public expenditure is rightly frowned upon in public finance for the same reasons. Nevertheless, a good case for correction of divergent regional growth rates could be made.
The criteria for redressing financial urban bias derive from considerations of dynamic economic growth and allocative efficiency, as well as regional equity and welfare. The relevant data base would be (1) rural-urban distribution of bank offices and the related credit/deposit ratios; (2) the pattern of remittances through commercial banks, post offices, etc.; and (3) the differential between the levels of per capita income, employment, and economic activity in rural and urban areas.
The following are some of the possible institutional mechanisms, in addition to fiscal and other incentives, to redress financial urban bias in LDCs:
• The expansion of the number of bank offices, or credit agencies, is a necessary but not a sufficient condition for redressing financial urban bias inasmuch as it does not ensure conversion of net deposit centers into net credit centers, allowing for the inevitable time lag for such a process.
• The creation of regional rural banks (as in India, Ghana, and the Sudan) also offers considerable scope for mobilizing local savings for local investment. The operations of the Sudanese Savings Bank in Wad Medani, the most important center in the relatively prosperous agricultural region of the Gezira, are a good example of a concerted program of mobilization of rural savings for local investment. However, a firsthand study also noted the difficulties in the implementation of such a policy because of, among other factors, the high average cost and volatility of deposits (due to small average balances being spread over a large number of accounts and the high costs of running mobile bank units) and the need for high liquidity reserves. Nevertheless, the project showed several encouraging results, notably the growth of savings deposits through conversion of hoarded currency by customers who had no previous bank accounts, and a generally higher utilization of local deposits for local credit.
The regional rural banks in India have been notably successful in extending credit to the weaker sections like marginal farmers, landless laborers, and rural artisans. At the end of 1982 the share of total credit extended by these banks to these groups was about 91 percent, a total of Rs 360 million in 164,338 accounts under the Integrated Rural Development program, as against Rs 759 million in 47,007 accounts at the end of 1981.
• The prescription of local credit targets by the central bank may also be viewed as a possible measure. In Thailand bank branches in the undeveloped Northeast region are required to extend not less than 60 percent of local deposits for local credit and not less than 20 percent of deposits for agriculture.
The efforts made by commercial banks in India to deploy more credit in rural and semi-urban areas since nationalization (1969 onward) are reflected in the significant increase in the credit/deposit ratio of rural offices. The credit/deposit ratio increased from 37.2 percent to 58.7 percent between June 1969 and June 1982, and of semi-urban offices from 39.7 percent to 50.0 percent. However, banks were still below the target ratio of 60 percent set for rural or semi-urban areas. The ratio for urban offices was 61.0 percent (59.7 percent in 1969) and for metropolitan offices 84.3 percent (106.1 percent in 1969).
The credit/deposit ratio, however, is not always an unambiguous indicator unless further refined to take account of population density (per capita) and the man/land ratio (per hectare). Moreover, the credit absorption capacities of rural areas also differ. It might be more meaningful to formulate a range for the credit/deposit ratios—disaggregated by regions—rather than prescribe quantitative targets at a particular level for the economy as a whole. The feasibility of credit targets postulates, among others, several concomitant measures, notably a greater emphasis on project viability as against rigid collateral requirements, in extending bank credit. This should be coupled with raising loan approval limits for branch managers and the formulation of active local advisory committees for bank branches to develop credit outlets. Although targets, no matter how scientific, may not always be achieved for valid economic or organizational reasons, and could even result in widespread evasion, this should not be a deterrent to appropriate policy.
• The successful experience of the Syndicate Bank in the South Kanara District, Karnataka State, India, in catering to small savers through its Pygmy Deposit Scheme (initiated as early as 1928), and to small-scale borrowers offers yet another possible model for utilization of local deposits for local investment. Under this scheme, the Bank collects small deposits (minimum of a quarter of a rupee a day) on a daily basis from the doorsteps of its customers (street vendors, laborers, village traders, etc.) through its field agents (collectors) who receive a commission on collections. The interest rates of these deposits are lower than the rates on regular savings deposits on the grounds that for such small driblets of savings there are no alternative investment avenues, that collection costs are proportionately very high, and they are more likely to be held as idle currency. Moreover, the attractiveness of the Pygmy deposit is not the rate of interest but the concomitant credit and technical assistance facilities (up to Rs 10,000) which are mostly available to small-scale retail traders, transport operators, industrialists, and professionals. The Bank has “self-employment” and “Farm Clinics” for credit and technical assistance to small-scale borrowers who are also insured by the Credit Guarantee Corporation of the Reserve Bank of India.
The basic issue is not financial urban bias per se, which may only reflect the greater credit-intensity of urban economic activity and its larger contribution to national net value added, but whether it is excessively regressive, and if so, what is an appropriate concerted strategy to redress it. A reduction of the disparities in regional credit/deposit ratios has to be achieved not so much by reducing transfers of savings to urban areas as by more vigorous mobilization of rural savings and by improving the credit absorption capacity of rural areas. Such corrective programs, however, need to be country-specific; they must also be based on the nature and pattern of the underlying rural-urban flows of savings and credit in relation to policy objectives and, above all, on the development of efficient rural financial markets.
Adjustment in Africa
A New Occasional Paper from the IMF
Occasional Paper 34, Adjustment Programs in Africa: the Recent Experience, by Justin B. Zulu, and Saleh M. Nsouli, reviews the design and implementation of adjustment programs supported by the Fund in Africa in the early 1980s. The role of the Fund, and the objectives and design of the programs, are examined in the light of the causes of the financial imbalances they are intended to alleviate. A discussion of the experience with implementation is extended by a detailed analysis of recent programs in Somalia and Mali.
Price: US $7.50 (US $4.50 to university libraries, faculty members, and students)
Publications Unit, Box A-852 • International Monetary Fund
700 19th Street, N.W. • Washington, D.C. 20431, U.S.A.