OECD UNITED NATIONS ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT ECONOMIC AND SOCIAL COMMISSION FOR ASIA AND THE PACIFIC JOINT OECD/ESCAP MEETING ON NATIONAL ACCOUNTS 1993 System of National Accounts: Five Years On Bangkok, 4-8 May 1998 Title: Estimation and Allocation of FISIM in Singapore s National Accounts Agenda item: 4 Author: Singapore
ESTIMATION AND ALLOCATION OF FISIM IN SINGAPORE S NATIONAL ACCOUNTS Aim 1. This paper outlines the exploratory studies done in Singapore on the estimation and allocation of financial intermediation services indirectly measured (FISIM) as recommended in the System of National Accounts, 1993 (SNA 93). Limitations and data gaps encountered are discussed. Present treatment 2. Current estimates of imputed bank service charges (IBSC) in Singapore s national accounts are based on the recommendations in SNA 68. IBSC is computed as interest received minus interest paid and allocated as the intermediate consumption of a nominal industry and deducted from overall GDP. Discussion of FISIM: SNA 93 3. Two methods of allocating FISIM were proposed: a) Relevant Financial Indicators The total value of FISIM could be allocated using different indicators. For example, it could be allocated in proportion to the total financial assets and liabilities that exist between financial intermediaries and various groups of users, or in proportion to other relevant financial variables. (SNA 93 Para 6.129) b) Reference Rate Method When the output is allocated among different users, a possible approach is to base the allocation on the difference between the actual rates of interest payable and receivable, and a reference rate of interest. The reference rate to be used represents the pure cost of borrowing funds. The types of rates chosen as reference rate may differ from country to country but the inter-bank lending rate or alternatively, the central bank lending rate could be used. (SNA 93 Para 6.127-128) Peter Hill s Proposal 4. In the National Account Working party held in June 1996, Peter Hill proposed evaluating the total value of indirect charges receivable by a financial institution from its customers as the sum of the total of indirect service charge receivable from borrowers and the total indirect service charges receivable from depositors. FISIM is valued on the basis of the difference between the actual rates of interest payable and receivable and a reference rate of interest. This implies FISIM on loans and deposit should be calculated for each sector as follows: FISIM L - L * (1) FISIM D = D * - R D (2) 2
Where FISIM L : FISIM on loans granted to institutional sector FISIM D : FISIM on deposits of the institutional sector R L : interest flows receivable on loans R D : interest flows payable on deposits *: internal reference rate L: stock of loans D: stock of deposits 5. Total FISIM by an institutional sector j is obtained as the sum of FISIM on loans granted to the institutional sector (FISIM L j ) and FISIM on deposits of the institutional sector (FISIM D j ). 6. Rearranging equation (3), we get: FISIM j = FISIM L j + FISIM D j - L j * + D j * - R D (3) FISIM j = FISIM L j + FISIM D j - R D - (L j D j ) * (4) 7. Total FISIM is obtained by summing the FISIM for each institutional sector. Studies conducted in Singapore using: a) Relevant Financial Indicators 8. The first study was conducted in 1994, soon after the release of SNA 93. Given the large amount of data required and the difficulty of establishing a reference interest rate, the method of allocation by total assets and liabilities was used to allocate FISIM among the various users household, government, corporate and exports. Net interest receipts were used to estimate the output of financial intermediaries. Allocation of IBSC to Rest of the World (ROW) 9. Allocation to ROW was based on the total assets and liabilities as at end period obtained from the banking schedules of financial institutions submitted to the Monetary Authority of Singapore (MAS). The proportion of non-residents share of assets/liabilities is then used to allocate IBSC of the respective financial institutions to ROW. Allocation of IBSC among resident sectors 10. The allocation of IBSC of financial institutions among resident sectors was based on the total deposits and loans and advances of commercial banks, merchant banks, Asian Currency Units (ACUs) and finance companies to resident non-bank customers. The total deposits of resident non-bank customers from MAS banking schedules were analysed by type of deposits and by the source (i.e., whether it is placed by government, statutory boards or non-bank financial institutions. Loans and advances to resident non-bank customers were also analysed by sectors. The amount allocated to household, government and exports were regarded as final demand while that to the corporate sector was then allocated to the different industries. Allocation of IBSC by industry 11. The allocation of IBSC by industry was based on the proportions of loans and advances by industry from the banking schedules. Information on deposits by industry was not available. Reasons why this method was not implemented i) Reference Rate Method preferred in ternationally 3
12. Although two methods of allocating FISIM were presented in the SNA 93, subsequent discussion in international meetings and papers presented by various statistical agencies indicated that the use of the reference rate method was preferred over the use of financial indicators to distribute FISIM to its various users. This is because the reference rate method is able to take into account different interest rate on loans and deposits faced by different sector of the economy on different instruments. ii) High per cent of allocation to exports 13. As Singapore is a financial centre, non residents account for a substantial share in the total assets and liabilities of financial intermediaries. This meant that the method would imply an allocation of about 30-40 per cent of FISIM to exports, and a relatively smaller proportion allocated to intermediate consumption to be deducted from GDP. This would have the effect of raising GDP by a significant amount. Given the limitations of this method, it was uncertain if the large proportion of FISIM allocated to exports was reflective of the underlying economic reality. iii) Constant Price FISIM 14. In the absence of an appropriate price deflator for FISIM, there were concerns whether the use of GDP deflator to deflate the current price estimates of FISIM to obtain constant price FISIM was appropriate. 15. We discussed these concerns with Professor Peter Hill during his consultancy visit to our Department in Nov-Dec 97. He agreed that in the absence of the required information to use the reference rate method, especially for historical data, the second best method, for example, allocation by total assets and liabilities might be acceptable. b) Reference Rate Method 16. Although this method is conceptually logical and easy to understand, the amount of information required is massive and detailed. Preliminary work was carried out using the domestic 3-month interbank rate. Negative output was obtained for certain financial institutions, for example, merchant banks for some years. This is likely to be attributed to the non-availability of specific reference rates for the different types of loans and deposits of specific sectors taken from/placed with different financial intermediaries. Issues of Concerns - Limitations and Data Gaps 17. Estimation of FISIM according to the reference rate method would require information on interest received on loans, interest paid on deposits by sectors, stock of loans and deposits by sectors as well as reference interest rate for specific sectors and financial institutions. Data on interest received and paid by the broad sectors are available from the Survey of Financial Institutions conducted by the Monetary Authority of Singapore (MAS). However, some data gaps remained. These include: i) Stock of loans and deposits by sector 18. Stock of loans and deposits are available for some sectors such as the corporate sector but are weak for households and NPISHs. Additional data sources or new methodology have to be developed to obtain improved estimates. ii) Reference interest rate for specific sectors and financial intermediaries 19. Ways to establish the appropriate surveys/administrative procedures to derive the reference interest rate for each sector by type of financial institutions and the different type of financial instruments have to be examined further. Data availability and other countries experience in developing their data sources will be helpful. 4
20. One possibility discussed in the note by the OECD secretariat to be presented at the May 98 OECD/ESCAP meeting on national accounts was to use the interest received/paid divided by the stock of loans/deposits to derive average deposit rate and the average borrowing rate. The reference rate is the mid-point between these two rates. As noted in the same paper, although this is a good practical solution, the original theoretical justification for using a reference rate namely that it represents the pure cost of borrowing no longer applies. 21. An attempt was made to compute FISIM using the method outlined above. We could derive fairly reasonable effective interest rates (from the ratio of interest paid by banks on deposits over the average balances on deposits fo993 1995) for FISIM on deposits for different sectors, ranging from 4 5 per cent. Effective interest rates on loans are expected to be higher than that paid out for deposits in order for banks to earn a margin. However, in this case, the effective interest rates on loans was about 5 per cent. The published minimum lending rate for banks (average fo0 leading banks in Singapore) was 6.26 per cent in 1995. Further examination on the data/data sources will be required to resolve this apparent inconsistency in results obtained. Conclusion 22. Singapore is keen to implement the recommendation on the estimation and allocation of FISIM as recommended in SNA 93 and further developed by a number of statistical agencies. 23. However, for historical data, it will not be feasible to obtain the detailed stock and reference rate for specific sectors and financial institutions. As a second best solution, one possibility is to allocate FISIM to the respective users based on available financial indicators, for example assets and liabilities as recommended in SNA 93. This method needs to be further assessed and the estimation of historical data further discussed to obtain a viable solution. 24. For current and future estimates, we agree that the reference rate method is conceptually appealing. However, there are practical difficulties in filling the data gaps. These difficulties will have to be resolved if we were to adopt this approach to estimate and allocate FISIM. 25. We will work with the Monetary Authority of Singapore to improve our data sources. However, there are concerns that we should not impose too heavy a burden on our banks and financial institutions. In this regard, the experience of other countries will be useful. Authors: Dr Soon Teck Wong, Ms Ho Poh Ching and Ms Lilian Tan Seak Eng Department of Statistics, Singapore May 1998 References Hill, P., The services of financial intermediaries Revisited or FISIM Revisited, Australian Bureau of statistics paper presented at the Working Party of National Accounts of the European Communities, June 1996 quoted from Arpaia, A. and Scafuri, E., The impact of FISIM on GDP: An empirical evaluation on Italian data, Eurostat and Istat presented at the OECD National Accounts meeting, June 1997. OECD Secretariat, FISIM, to be presented at the OECD/ESCAP meeting on national accounts, 4 8 May 98, Bangkok. United Nations, System of National Accounts, 1993. Disk C:\SNA93\FISIM 5