![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
Abstract
This paper studies the flow of funds and portfolio behaviour of Indian banks from 1951 to 1994. In this period, financial controls such as variable reserve ratios were important constraints on bank behaviour, especially before liberalization took place in the early 1990s. We estimate a system of demand functions which uses as framework the Almost Ideal Demand System and which incorporates the reserve ratio regulations. Attention is paid to cointegration and to structural breaks. The estimated model provides coherent and plausible parameter estimates for prices and other variables. We find that a standard portfolio model can usefully be applied to the study of financial behaviour in a developing economy such as India, and some interesting policy implications can be drawn.
Acknowledgements
This paper is a revised version of Chapter 5 of Tomoe Moore's PhD thesis at Loughborough University. An earlier draft paper was presented at the International Conference on Finance and Development: Evidence and Policy Issues, Nairobi, Kenya, July 10–11, and funded by The Kenya Institute for Public Policy Research (KIPPRA) and the Department for International Development (DFID). We thank participants in this conference for their helpful comments. We thank without implicating Kunal Sen for kindly providing his data at an early stage of the project, and Kunal Sen and Rajendra Vajda for several helpful conversations on the material in this paper. We also thank the referees and editor of the European Journal of Finance for their very helpful comments on an earlier draft of this paper. The research underlying this paper forms part of a general research programme on finance and development funded by DFID. We thank DFID for their financial support. The interpretations and conclusions expressed in this paper are entirely our own and should not be attributed in any manner to DFID.
Notes
One of the authors has proceeded further to estimate and simulate a complete flow of funds model for India, based on a four-sector disaggregation (Moore Citation2003).
Indian flow of funds data are compiled on a fiscal year basis from April through end-March of the following calendar year. The main practical problem with the data is that there is typically a long lag in publication. At the time this research was carried out, the latest available detailed data were for 1994, although very summary data were available through 1996. Work on data for other sectors of the economy is reported by Moore Citation(2003).
Public sector banks, consisting of the 20 nationalized commercial banks and the eight state banks, were liable for around 92% of total deposits in the banking sector during the 1980s.
During the fiscal crisis of the early 1990s the CRR and SLR were raised to the statutory maxima of 15% and 38.5%, respectively. With 40% of the remainder allocated as loans to priority sectors, the banks were left with under 28% of their liabilities to be invested at their own discretion.
Current wealth (W t ) is the sum of end of last period wealth, current saving and capital gains or losses.
Netting-off required reserves from deposits is standard in the literature.
Separability of deposits and assets and exogeneity of deposits jointly imply that we take bank size as given. The composition of bank assets is not separable; otherwise, there would be no portfolio choice problem!
Alternatively, capital could have been consolidated with deposits. This treatment makes no difference to the results we report.
The financial year begins on 1 April and ends on 31 March.
See Reserve Bank of India Monthly Bulletin, various issues, and Reserve Bank of India Citation(2000a).
The stock data on 31 March, 1951 for financial assets were obtained from the balance sheets of the aggregated commercial banks. Valuation changes were assumed to be negligible.
In fact, it appears that the main portfolio impact of lending instructions fell directly on bank profits (see Narasimham Committee 1991).
The relatively short time series precluded a fuller investigation of structural shifts than that described in Sections 5 and 6.3.
The VAR was of order 2.
I(0) variables may contribute to a sensible long-run relationship among I
Equation(1) variables although they would affect the estimated number of cointegrating vectors. This does not appear to be an issue in our results as we explain below (Harris Citation1995).
For PLA, although lending rates were regulated for most of the sample period, banks could carry out some lending to the corporate sector at unregulated rates. We interpret these to be the marginal market-clearing rates.
It should also be borne in mind that if dummy variables enter the deterministic part of the multivariate model or the sample size is small, the power of these tests is correspondingly reduced (Harris Citation1995).
Bai Citation(1996) argues that the most likely breakpoint is that at which the Chow test is rejected at the highest significance among other potential breakpoints.
There are insufficient observations to permit separate slope coefficients across this break.
To economize on space, these are not shown, but they are available from the authors on request.
The RBI now uses its repo rate as a principal policy instrument in addition to bank rate itself. However, bank rate is a more consistent measure of the cost of bank reserves over the sample period as a whole.
The imposed value of γ11 of −0.38 was chosen using a search in steps of −0.02 starting from zero. For values of , the restriction was rejected by the likelihood ratio test. For values of
, the restriction was accepted, but the other price parameters were less well determined.