ABSTRACT
A flow of funds in the financial sector arises from financial transactions among economic agents. These financial activities of an economy are then registered in a flow of funds account, which shows the flows of borrowing (i.e. sources) and lending (i.e. uses) of funds among disaggregated sectors for disaggregated financial instruments. Modelling the disaggregated sectors’ portfolio behaviour is termed as a flow of funds model. Flow of funds modelling is not widely used in developing countries, mainly because of an almost complete absence of sufficiently detailed data to construct even simple flow of funds accounts. Nevertheless, it is generally recognised that flow of funds analysis is potentially of great importance to developing countries because the flow of funds accounts reveal the sources and uses of a particular fund that are needed for growth and development (Klein 2000: 9). The World Bank has used a flow of funds framework as part of its country-
based RMSM-X modelling,1 but this framework has its limitations. In typical applications, the whole private sector has to be treated as a single entity because of a lack of suitably disaggregated sectoral data (Holsen 1991). Green and Murinde (2003) argued that an important priority in flow of funds modelling for developing countries is to investigate patterns of intersectoral financial flows, focusing on a more disaggregated treatment of the portfolio choices of individual sectors such as households and businesses. Disaggregation is essential in empirical work, since the balance sheet and flow of funds data of different sectors clearly show marked differences with respect to their net wealth positions and the pattern of assets and liabilities. We take up the challenge in this study by constructing data for and then estimat-
ing a complete model of disaggregated sectors’ portfolio choice with associated policy simulation for India. The methodological steps are taken as follows. First, a flow of funds is modelled for individual sectors of banks, other financial institutions, private corporate businesses and households with six financial instruments by demand functions. Second, by consolidating all the estimated models of disaggregated sectors, a system-wide flow of funds model is constructed where each financial market is solved by the market-clearing conditions. Third, using the system-wide flow of funds model, policy simulation experiments are conducted in
in India. This study makes several major contributions to the literature. First, this is, to
our knowledge, the only comprehensive, sector-specific, flow of funds model for any developing country.2 The choice of India is suggested partly by the fact that it is unique among developing countries in that detailed national flow of funds accounts have been compiled by the central bank (the Reserve Bank of India: RBI) for over 40 years (RBI, various issues). Without this base, sectoral econometric modelling would be impracticable. For modelling purposes, however, the published data are incomplete in certain key respects, and our second contribution is to build on these data to construct the necessary time series of stocks in India, covering the period 1951/52 through 1993/94. Note that Indian flow of funds data are compiled on a fiscal year basis fromApril through end-March of the following calendar year. Third, the model is more firmly grounded in theory than is common in flow of
funds analysis. We use the Almost Ideal Demand System (AIDS) of Deaton and Muellbauer (1980a) as the basis for the empirical model. This allows us to test the theoretical restrictions of homogeneity, symmetry and negativity implied by the axioms of rational choice. TheAIDS model is sufficiently flexible to enable us also to model the impact on portfolio behaviour of changes in the policy regime in India. The AIDS approach circumvents the well-known atheoretical limitations of Brainard-Tobin type models, while also being relatively parsimonious in parameters (see Chapter 2 for more detail). Fourth, we use co-integration techniques to estimate the model. We follow
Barr and Cuthbertson (1991a,b,c, 1992a,b and 1994) who conducted a flow of funds model for a single sector study in the United Kingdom, in employing the Engle-Granger method to tie down the long-run relationships (Engle and Granger 1987). But we also innovate by using the Johansen procedure (Johansen 1995) to check the reliability of our estimates of the co-integrating relationships. Fifth, it is associated with the policy analyses. The financial systems in develop-
ing economies tend to be characterised by a range of restrictions. The flow of funds framework is able to incorporate these developing country-specific features and contributes to identifying effective policies for economic growth. In this book, the policy simulation experiments are conducted with a view to analysing the delivery of loanable funds to sectors which are the most in need of poverty-reducing economic growth, and they are largely in line with the financial reforms that started in the early 1990s, such as removing the ceiling for interest rates, lowering reserve requirements and the disciplinary stance of fiscal deficit. Simulation quantifies the potency of policy instruments on the flows, hence it can give a clear picture of the channels through which policies may affect different sectors of the economy (Green et al. 2002). Clear understanding of the outcomes of policy instruments is necessary for the re-designing and restructuring of the financial reforms. In this respect, we contribute to the literature on financial liberalisation in developing countries. Sixth, the system-wide simulation designed in this bookwill permit us to analyse
awide spectrum of policy effects on such issues as the determinant of interest rates,
of and allocation of credit, as a result of interactions in the disaggregated economic sectors. A flow of funds in India was modelled and simulated by Sen et al. (1996).
They have used the RBI data to estimate a general disequilibrium model of the Brainard-Tobin type (Brainard and Tobin 1968). The study was conducted in line with the stabilisation macroeconomic policy undertaken in the early 1990s in a simple general equilibrium framework. Their study has important limitations: the sample period covers just 20 years (1970/71 to 1989/90); there are only two sectors of banking and household sectors with four financial assets, assuming financial flows in other sectors to be policy-determined. The current study extends their empirical study in a more rigorous and consistent manner in scope: we use the flow of funds time series (1951/52 to 1993/94) and we study a more comprehensively defined list of assets for four sectors. Moreover, these data are based on the total net transactions of uses and sources of funds satisfying the market-clearing conditions in a consistentmanner, and not just their transactionswith any particular sector. Finally, note that there is a considerable time lag in the flow data published in the
Reserve Bank of India Bulletin. The latest detailed flow data are those for 1993/94 published in 1998. This data period means that the analysis cannot be brought fully up-to-date, although it does cover the pre-reform and immediate post-reform eras. However, this does not limit the overall value of the study because a crucial contribution of this book is the manner in which it employs simulations (both deterministic and stochastic) on a complete flow of funds model of the whole financial sector portfolio behaviour: all the AIDS model for a single sector is consolidated to embody the financial sector for the simulation policy experiments. The empirical performance of such a complete flow of funds model is rare even for industrial economies; in many cases the study tends to stop at individual sectors. Covering a wide range of assets and considerable sectoral disaggregation, such an undertaking is huge and therefore tends to discourage researchers (Green 1982). In this respect, the study conducted in this book enables a more rigorous analysis of policy responses during financial liberalisation in India, than has heretofore been carried out in other emerging financial markets. Moreover, it is distinguished from the conventional macroeconomic policy models, in which typically only money and one homogenous non-money asset are considered in an aggregated economy. The book consists of 11 chapters, and is organised as follows. In Chapter 2 the
empirical and theoretical literature on the demand functions for a flow of funds model is reviewed by addressing the main features of the different types of demand functions. Chapter 3 is dedicated to the financial system and the compilation of the flow of
fundsmatrix for India.After independence, financial markets had been functioning in a heavily regulated framework in India, and financial liberalisation started in the early 1990s, aimed at the de-regulation of the financial system. In Section 3.1, the course of financial reforms and their effects on the financial markets are
