Household savings : an estimation for Sri Lanka

SDF- 27 CONFIDENTIAL STUDIES IN DOMESTIC FINANCE No. 27 HOUSEHOLD SAVINGS: AN ESTIMATION FOR SRI LAiKA By Katrine W. Saito* Public and Private Finance Division Development Economics Department * I wish to acknowledge the helpful comments received from Dr. J.K.S. Ghandhi; the responsibility for remaining errors, however, is mine alone. October 1976 CONTENTS P Pa'ge -No. I. Introduction 1 II. Measurement of Saving: General Concepts and Definitions h III. Procedure for Deriving Esimates of Household Sector Saving of Sri Lanka: 1l 1. Currency 14 2. Net Deposits with Commercial Banks 18 3. Net Deposits with Savings Banks 23 4. Net Saving with Long-term Credit Institutions 24 3. Net Saving with Hire Purchase Finance Companies 27 6. Net Claims on the Government 29 7. Net Claims on Co-operatives and Rural Banks 30 8. Net Corporate Securities 31 9. Net Saving through Provident Funds 33 10. Net Saving in the Form of Life Insurance 36 11. Tangible Asset Holdings 39 IV. Summary Of Findings- 43 Annex Tables. TEXT TABLES Table No. I. Household Sector Savings in Current and Constant Prices: 191-74. II. Household Sector Savings in the form of Financial and Tangible Assets: 1951-74. III. Change in Assets and Liabilities of the Household Sector 191-7. IV Structure of Net Saving in Financial Form of the Household Sector: Annual Averages over 5 year periods. V. Structure of Gross Saving in Financial Form of the Household Sector: Annual Averages over 5 year Periods. VI. Comparison of Savings Estimates. ANNEX TABLES Table Number 1. Household Sector Saving in the Form of Currency. 2. Household Sector Gross Saving in the Form. of Commercial Bank Deposits. 3. Household 'Sector Liabilities with Commercial Banks. h. Household Sector Net Saving in the form of Commercial Bank DeDosits. Household Sector Net Saving with Savings Banks. 6. Hovsehold Sector Net Saving with Hire Purchase Finance Companies and Long-term Credit Institutions. 7. Household Sector Net Saving with the Government. 8. Household Sector Net Saving with Co-operatives and Rural Banks. . Household Sector Net Saving in Corporate Securities. 10. Household Sector Net Saving through Provident Funds: The 1nployees Provident Fund. 11. Household Sector Net Sa Tg through PIovident Funds: Private Provident Funds. 12. Household Sector Net Saving in the Form of Life Insurance: The Insurance Corporation of Ceylon. 13. Salary Loans made by the Insurance Corporation of Ceyloi. 14. Household Sector Net Saving in the Form of Life Insurance: Private Life Insurance Companies. 15. Household Sector Saving in the form of Tangible Assets. 16. Household Sector Net Saving 1951-75. I., Introduction Despite the significance of saving in the process of economic development, relatively little is known on the country-specific-basis about- the magnitude of saving, about the relative contribution of different groups of savers, and of the forms in which savings are accumulated. When savings data are available, they have usually been derived from the National Accounts, a method which, in many developing countries, is notorious for its margin of error. This is particularly true of the estimates of the household sector, which are usually derived as a residual in the Accounts. Since it is saving of the household sector which supplies the major portion of total savings of most non-communist countries, this paucity of reliable estimates of household saving is all the more surprising. An additional disadvantage of estimating household saving from the National Accounts is that no information is provided on- the structure of household savings, i.e., on the distribution of saving among financial and tangible assets. Without reliable information on its magnitude and structure, it is difficult to see how quantitative advice on policies relating to the mobilisation of household savings can be given. This paper presents a systematic procedure for estimating saving of the household in a developing country within a system of social accounts. Saving is defined as the change- in earned surplus and measured primarily from balance sheets as the change in that sector's asset holding less the change in liabilities. By estimating in this way information is provided on the various forms of saving of the household sector, and on Trends in the components, both individually and in relation to each other, over the period of study. The conceptual matrix and accounting techniques for estimating savings within a system of social accounts were first developed on a comprehensive basis by R. W. Goldsmith using the United States as a case study. This was the first investigation of such a comprehensive and detailed nature into saving, and estimates were derived for saving on a sectoral basis for a period of fifty years. The details of this system, however, are too complicated and too anbitious for application in devel- oping country- where the data limitaticns can be severe. Furthermore, the institutional framework in developing countries can be quite distinct, so that the estimation procedures such as those devised by Goldsmith would be inappropriate. There have been few attempts to modify Goldsmith's methodologr and develop direct estimates of saving for developing countries. One exception is the excellent work done by the National Council for Advanced 2/ Economic Research in India which was appropriate to the data base and the 1/ R. W. Goldsmith (1995). Estimates of saving of the U.S. were also - developed by Friend ( 1954) and Friend and Jones (1960 ); of Shinohara S(1989);..Mizoguchi ( 1970 ) and Blumenthal have done similar work on Japan. Relatively for cuch studies have been made for United Kingdom, although some empirical work was done by Radice (1939 ) and the Oxford institute of Statistics in the 1950' s on the structure of consumer assets. The OECD ( 1968) have also published a comprehensive series of savings data for several member countries. 2/ N.C.A.E.R. ( 1965). -3- institutions there. Among the attempts to develop savings estimates using a system of social accounts, for other developing countries, the E.C.A.F.E. 3/ studies deserve mention, although they are generally of a le,- refined and comprehensive nature than either Goldsmith's or the N.C.A.E?.R. work. In this paper, a procedure for estimating househ6ld sector savings with specific reference to Sri Lanka is described. It is believed that this estimation has a general applicability to developing countries. The methodology used in described in some detail since it does involve several features which are quite different from the estimation rccedures developed by Goldsmith and the N.C.A.E.R. although the conceptual matrix around which the syste is built remains the same. Hous-ehold savftings are estimated on an annual basis for a period of twenty-four years, thus providing a sufficient number of observations for further puantitative studies. Section II discusses in general. terms the procedure for measuring savings, examining first the different concepts and definitions of saving and then briefly outlining the procedure for deriving household savings and the data problems likely to emerge in the process of estimation. Section III, discusses in more detail the standard procedure for deriving house- hold sector saving, and the specific procedures adopted in the Sri Lanka estimation. The final section presents a summary of the derived savings figures. 1/ For example, the savings studies on the Phillipines (1 963) on Japan (1958), on Malaysia (1963 ), and on Si Lanka ( 1961 ). II. Measurenemt. of Saving: General Concepts and Definitions A detailed description of the actual procedures adopted in deriving the estimate of household sector saving of Si lanka is provided in the folLowing section Rather than repeat or condense this description, the purpose hore is to first explain the different concepts and definitions of saving and then describe broadly the estimation procedures for deriving household sector saving and the problems likely to emerge. In its simplest form, saving of a sector can be defined as the excess of that sector?s income over its consumption expenditures. Such an excess of income manifests itself.in an increase in the stock of assets helcl b7 that sector. Alternatively. therefore, saving may be defined as the increase in assets less any increase in liabilities held by that sector . in A.ciounting terms, this is the equ4valent to earned surplus, provided that the definitions are consiRtent. Saving of an economic unit can thus be estimated in two ways: (i) E1om the income account as the difference between current income and current non-capitalizable expenditures, including distribution to owners; (ii) from the balance sheet as the net result of changes (ex- cluding valuation changes) in all types of assets and liabilities. The numerical results of the two calculations are necessarily equal if the accounting methods employed are consistent. This equality is expressed in three fundamental relationships: (i) current revenue - current expenditures (including distributions to owners) = retained income; (ii) change in assets - (change in liabilities plus change in paid in capital) = change in liabilities plus change in paid capital) = change in earned surplus; (iii) retained income change in earned surplus. The distinction between saving and investmient can lead to some confuasion. At the sectoral level, saving in the form of tangible assets constitutes that sector's investment, and this can be financed tb obtaining credit from financial institutions as well as by direct saving. For a sector, therefore, the amount of investment may be quite different from saving, which is made up of not only real investment (i..e., tangible assets) but also includes the net increment in financial asset holdings. At the national level, for a closed economy, saving will .be equal to investment. National saving is the -result of the consolidation of saying or dissaving of all sectors within a nation. The important point is that national saving is a net concept, i.e., the dissaving of some sectors should be treated algebraically as an offset against the saving of others. Moreover, in a closed economy, if the saving of different sectors has been estimated on a consistent basis, then national saving in the form of financial assets must be zero, since one group's financial assets constitute another's liabilities. National saving thus becomes equivalent to the change in national tangible wealth, or to national investment. - 6 - Estimates of national saving can be derived using a "direct" or an "indirect" z-pproach. According to the former, estimates of saving are derived sepazately for the government, corporate and household sectors and these are then aggregated to form national saving, This is the approach adopted in this paper. National saving derived according to the rindirectlt approach is undertaken in the following manner: (1) The foreign balance on current transaction plus (2) Capitalizable purchases by domestic units from domestic units (i.e., gross domestic capital formation or gross domestic investment as it is usually defined) less (3) Depreciation allowances. It is not possible to derive an estimate of household saving of S&i Lanka for* the period 1951-74 using'this "indirect"' approach. To sarte 'ith5 no uo:ptal foimalion data are available for the year priior to 1959, and seccndly, the capital formation data available from 1959 have no depreciation allowances. As emphasised earlier, the basic advantage of estimating house- hold savings using the "direct" approach is that information is provided on th' saving of different groups and on the forms in which savings are accuiulated. From a theoretical viewpoint, the grouping of savers and the division of saving into different assets and liabilities should, to quote Goldsmith: 1/ An ecellent attempt has been.made by T. Savundranayagan (1974) to develop the National Accounts for the years 1950-59. However, only product data-have been derived, no income or capital formation data. "be carried to the point where both groups and forms become heterogeneous. This point is reached when each of the series that reflects the saving of one group of savers through one form of saving is made up of cells... which behave in a reasonably similar way." In practice the limitations of data can be such that the grouping of savers and the forms of saving adopted can be far from homogeneous. Some compromise must be reached between the theoretical and the feasible. The 2/ grouping adopted in this study is not so refined as in Goldsmith. The household sector is defined rather broadly to include not only households, but non-corporate enterprises and non-profit organisations also. Fom a theoretical viewpoint it would have been interesting to. treat unincorporated enterprises and non-profit organisations separately, especially the former, since their activities are likely to be significant in Sri Lanka. In practice, it is extremely'difficult to collect data on the operations of unincorporated enterprises since within the household sector a number of economic activities are often pursued by different members of the household either simultaneously 3/ or during different time periods as joint enterprises.- I/ Goldsmith (1955), p. 88. 2/ Goldsmith divided -the household sector into the following three groups: cultural, household and individuals; (2) Agricultural households; (3) Un- incorporated enterprises. 3/ See M. Mukherjee, U. Datta, R. Choudhury and D.S. Prasada Rao (1975) p. 419 on this problem. The forms of saving which are distinguished are both financial and tangible. The financial assets (ntof iblte)tknit con are currency, bank deposits, holdings of corporate and Government securities, savings, through pension funds and life insurance, and gold. The tangible assets included are residential constructiori, machinery and equipment, vehicles and inventories of goods. Saving of the household sector can be derived from balance sheets or inecome accounts, given data availability. However, to obtain information on the forms in which savings are accumulated, the balance sheet approach is necessary. A major problem is obtaining the required data,since unlike eith9r the corporate or the government sectors, there are no series of household accounts which can be used to estimate household sector savings. Recourse mu,3t therefore be made to the accounts of the financial institu- tions, the gover=,nI.n and. corporations. The accounts of the financial institutions are generally reliable and accessible, and tend to be the main source of data on household savings. However, they rarely have the necessary household/corporate breakdown. This was frequently the case with the Sri Lanka estimation; the accounts of the financial institutions provided the basic data series on the savings of Athe private sector (i.e., household and corporate sectors together)., and it was necessary to use other sources of data, such as benchmark survey data, to derive the household share. Government accounts are also easily accessible, and to the limited extent that they have relevance to the estimation of household sector savings, they can be useful. This rele- vance may be restricted to holdings of government securities by houeeholds -9- and to any credit extended by the government directly to the public. There may be problems in obtaining data on the activities of all registered companies. The annual financial statements of joint-stock companies are generally available, but information on their activities in aggregate may not. The balance sheets of private companies are much less accessible, yet they may well comprise the major share of all registered companies. The information needed for estimating h9usehold sector assets and liabilities is primarily holdings of corporate shares by households, which in the case of Sri lanka, were impossible to.obtain directly and had to be estimated using sample data provided by a major brokerage firm. When the accounts of the financial institutions do not give an adequate breakdown of their assets and liabilities to derive the house- hold sector share, then this share may have to be derived as a residual, bhose OfPth corporate aid poem~~t Ti'~ cases the government and corporate sectors accounts are useful. For example, household holdings of currency in Sri Lanka were derived by this proce.dure. The data series published by the Central Bank entitled "Currency held b the public" includes not only household holdings, but also those of companies, co-operatives -and local governments. Holding. by households only were derived as a residual after deducting the estimated holdings of companies, co-operatives -and local governments, all -of which, were derived from their respective accounts. 1/ In Sri lanka, for example, the Government makes loans to its employees. -10- The first stage in compiling the basic data for estimating household sector saving using the balance sheet approach consists of identifying data series on the major financial assets and liabilities and tangible assets held by households. Data selection should be based on recurrent availability, reliability and coverage. The gaps which emerge in the data series must be filled through different statistical processes. In the Stii Lanka data, the gaps were due mostly to data series which did not extend over the entire period or to data which lacked a corporate/household sector breakdown. When a related data series which extended over the entire time period was available the non-continuous data were extrapolated by assuming a constant ratio between the two data series, or through regressing one series on the other and using the derived coefficients to extrapolate. When no related data series were available, the basic data were exGended through the most appropriate trend. Derivation of the household sector share was usually achieved through utilising benchMark azurvey data which did give such a breakdown, and extra- polating to derive shares for the other years. The forms of household sector saving which were distinguished in the Sgi, Ianka estimation were the following: (1) Currency. (2) Demand, time and savings deposits with commercial banks net of credit extended. (3) Time and savings deposits with savings banks, net of credit extended. (1) Time and savings deposits with long-term credit institu- 2/ tions net of credit extended.- ($*) Claims on the government net of credit extended. (6) Claims on companies. (7) Claims on co-operatives net of credit extended. 3/ (8) Saving through life insurance net of credit extended. (9) Saving in the form of provident funds net of credit 4/ extended.- (10) Gold. (11) Residential construction. (12) Machinery and equipment. (13) Vehicles. (14) Planting and replanting of land. *and 1/ Includes deposits with/a credit received from the Post Office Savings Bank, Ceylon Savings Bank, Savings Certificate Fund, and after 1972, the National Savings Bank. 2/ Tncludes deposits with tnd credit received from the following: State Mortgage Bank, Agricultural and Industrial Credit Corporation, National Housing Fund, and Hire Purchase Finance Companies. 3/ Includes both private and State insurance corporations. 4/ Includes both private and the State provident funds. A debatable point is whether or not to include the accumulation of consumer durablesp thereby deducting such expenditures from current consumption and regard them as a part of saving in the form of tangible assets. In favour of this proposal, it has been argued that since these goods provide services which are not exhausted in the a.ccounting period,.they should be included. In other words, present holders of consumer durables enjoy services from such goods.included in the gross national product of previous years, while currently-produced goods will continue to provide services to holders in forthcoming years. On the other hand, the argument hos been put forward that expenditure on consumer durables is not regarded as saving by the consumer, so that in a behavioural sense, such expenditure should be treated as con- sumption. While this ma, well be true in a developed country, in those which are less developed it is likely that expenditure on at least some consumer durables is a capital investment. Transportation vehicles, for example, are I_;RE:Y to be--ze. not Gnly-I'cr porcnal ploasure but aic fo wcrk-rr,Dntr ri pirposes These are soine of the conceptal issues involved; there are also two important practical considerations. First,the anticipated use of the end- product zhould be borne in mind. If these savings figures are to be used in relation to the National Accounts, then to obtain c6nsistency, expenditure on these durables which comprise a part of gross domestic capital formation should be included. Secondly, the data limitations may be such that no data on consumer durables can be derived. In the case of Sri Lanka, both these considerations outweighed the conceptual arguments, and the only consumer durable good in- cluded was transport vehicles, expenditure on which is included in the National Accounts, and so data were available. This-inclusion can also be defended on the conceptual grounds described above. This inventory of forms of saving differs from Goldsmith's in several respects. Beginning with the tangible assets, Goldsmith included twelve types of new consumer durables on the basis of "decisions which, to 13 some extent, are matters of convenience and judgement rather than logic. The N.C.A.E.R. study, by comparison follows the same procedure as this current study, and includes only vehicles. The main differences in the financial forms of saving are due to institutional diversity. Judging by the magnitude of credit which they extend, co-operatives are an important credit institution in Sri lanka. They also issue share capital, though they are predominantly net extenders of credit. Because of their relative ins ignifi cance as institutions in the U.S., Goldsmith could ignore their activities in his estimates; the N.C.A.E..R. study, however, did take them into account. While non-institutional credit is of substantial magnitude in Sri Lanka, in common with other developing countries, reliable estinates can be extremely difficult to obtain. For Sri Lanka a new data series was developed in this current study on the net credit extended by hire-purchase finance 2/ companies sing company balan c sheets, original questionnaire returns to a Central Bank survey, and, for recent years information provided by the Central Bank. While this does not account for all the non-institutional credit ex- tended, (e.g., friends, relatives, traders, landlords, commission agents, private money lenders are other important sources), it was considered to comprise a major portion,. If no data were available on hire-purchase finance companies, some estimate of non-institutional debt in Sri Lanka could be derived frow surveys of consumer debt, and extrapolation undertaken on the b asis of the relationship of this data to some continuous data series, e.g., commercial bank credit or some index of earnings. It is interesting to note that 1/ Goldsmith (1955) -p. 412. 2/ Both private and joint-stock hire purchase finance companies were taken into account. Goldsmith took certain forms of non-institutional credit into account,- while the N.C.A.E.R. ignored it, despite its significance in India. III. Procedures for Deriving Estimates of Household Sector Saving of Sri Lanka The purpose of this section is to provide an explanation of the procedure for deriving estimates of the components of household sector saving, and outlining the specific procedures followed in the case of Sri Lanka. Com- parison is made to the estimation methods used in two previous studies of Sri Lanka savings, by W. Rasaputram (1961) who derived estimates for the period 1950-1959 and by S. Kahagalle (1975) for the period 1960-1972. The procedures adopted in these two studies are referred to in order to illustrate alternative methods of estimation and to demonstrate why the procedures used in the current study are preferred. This section does not contain an exhaustive description of data sources nor does it give a detailed step-by-step description of how these estimates were derived. The methodological framework described here should have a broad applicability to other countries, although certain adjustments may be necessary to accomodate the kind of data available and the institutional set up. 1. Currency Conceptually, the estimation of saving by the household sector in the form of currency is simple. It consists of the difference between the stock of currency held by that sector at the end of the year and that held at the end 1/ Estimates were derived for two categories of non-institutional debt; instal- lment credit obligations represented by notes given when commodities are pur- chased, and service debt (mostly debt for medical services). of the previous year. What is difficult, however, is the derivation of holdings of only the household sector, -since the data series on holdings of the "public" of "currency in ci:culation outside the banks and the Treasury" will contain non-household sector holdings. . In the case of Sri Lanka, the end of the year currency holdings by the "public" as presented in the Central Bank, Bulletin, include currency held by companies, co-operatives and local governments. The proced- ings by which these holdings were deducted from the Bulletin figures to arrive at household sector holdings are described below. a) Currency held bY companies A rather lengthf procedure was necessary to derive estimates of currency held by companies.. First the ratios of total currency plus demand deposits to total paid-up capital for approximAtely 200 companies J were derived for five benchmark dates. By fitting a curve to these ratios, a series of ratios were derived annually for the years 1950-7h. Using these ratios and the total paid-up capital of all registered companies,/ the holdings of currency and demand deposits for this sample were blown-up to represent holdings of all registered companies in Sri Lanka. To derive only currency holdings it was necessary to deduct from this data series holdings of demand deposits by companies. These were 1/ These data were taken from the Handbook of Rupee Securities, published annually by the Colombo Stockbrokers Association. 2/ Values were derived using an ordinary osculatory interpolation, Karup-King formula. 3/ The total paid-up capital of all registered companies is given in the Administrative Report of the Registrar of Capanies. *I 1/ derived- as a residual from the Monthly R turns by commercial banks to the Central Bank and subtracted from the smn of currency and demand deposits data series to derive currency held by the corporate sector. b) Currency held by co-operatives Estimation of currency holdings of co-operatives encountered the same problem of deducting demand deposits.from the data series as with the esti- mate of corporate sector currency holdings. The balance sheets in the Administrative Reports of the Co-operative Movement give end of the year holdings of currency plus demand deposits of the co-operative societies. These were adjusted from a financial year to a calender year basis. To derive only currency holdings, it was necessary to deduct demand deposits held by co-operatives. These data were available on an annual basis from 1964 in the commercial bank Monthly Returns. Earlier data were extrapolated by applying the average ratio of currency to .currency plus demand depccits 2 / for the years 1964-1969 to the currency plus demand deposits data for the years 190-63. c) Currency held by local governments To develop estimates of the petty cash held by local governments, -information was sifted from the Government Gazettes on the holdings of currency 1/ There is some minor problem here of comparability of estimates. Demand deposits are end of the year totals, whereas the currency -plus demand deposits data collected from balance sheets refer to the various financial years of com- panies. The error involved, however, is not anticipated to be large. 2/ The ratios for the years 1970-74 were not taken into account since a declining trend in the ratio is quite apparent from 1970; during the earlier six years there was less fluctuation in the ratios. -17- by a sample of urban, town and village councils, and municipalities for benchmark dates. The total income received by this local government sample was also recorded. Assuming the applicability of a constant ratio, the avorage of these ratios were multiplied to the annual income of all local authorities to derive the annual currency holdings of the local authorities for the years .1950-74. d) Currency held by the household sector Estimates of household sector saving in the .form of currency were arrived ;. by deducting the sum of the estimated changes in the currency held by companies, co-operatives and local government from changes in the currency by the public. These data are presented in Table 1 of the Annex. The methodology adopted here to derive household sector saving through currency bears some similarity to that of Rasaputram in that currency holdings of companies, co-operatives and local authorities are first estimated and househcld holdings arc then derived a- a rcsidual by deducting the sum of thcse non-household holdings from those of the "public'" Kahagalle derived currency held by the household sector by multiplying the estimated share of ho,sehold sector demand deposits to the total for each year to the cirrency holdings of the public. The underlying assumption, that currency held by households movus in the same proportion and direction as demand deposits, is 2/ open to serious question. 1/ There was little variation between them. 2/ See, for exampl,e, the empirical analyses of the currency to demand deposit ratio by D.J. Khazzoom (1966 ) and K.A. Lewis and F.F. Breen (1975 ). 2. Commercial Bank Deposits Net saving in the form of commercial bank deposits by the household sector is defined as the change in the end of the year holdings of bank deposits by that sector over the previous year's holdings less the change in liabilities of that sector with commercial banks. Estimation problems are similar to those which arose in the currency estimation, i.e., how to.'isolate the assets and liabilities of the household sector from non-household elements. a) Demand deposits The Central Bank Bulletin presents data on demand deposits held by the resident public which include not only holdings by households but also by companies, co-operatives and local authorities. Rather than attempt to estimate the holdings of these non-household sector groups and deduct from th, holdings of the resident public (in the same manner that currency holdings were derived), it was decided to use a different procedure. From 1996 the Central Bank undertook Quarterly Surveys on the holders of deposits and the recipients of advances. For the years 1956-74, the sum of deposits held at the end of each year by two categories in -these Surveys - non-business institu- tions and individuals - were interpreted as comprising household sector holdings. Holdings prior to 196 were extrapolated fron the total deposits held by the 1/ resident public, given in the Bulletin. A linear least-squares regressiorr 1/ DD = -162.37 + 0.74 DDP (Q.Q8) 2 R = 0.91 where DD = demand deposits held by household (data derived from Quarterly H Surveys) DD = demand deposits held by the resident public (data extracted from the Bulletin). (standard error shown in parentheses). -19 was run for the period 19.6-74 and the derived coefficients applied to the annual data on demand deposits held by the resident public from 1950-55. In this way holdings of the household sector could be derived. These data are presented in Table 2 of the Annex. This method differs from that adopted by Rasaputramn and Kahagalle. These authors utilised the Central Bank Quarterly Surveys to derive the ratio of deposits held by companies (Rasaputram) or households (Kahagalle) to total deposits and applied this ratio to the end of the year demand deposit holdings of the public given in the Central Bank Bulletin. Specifically, Rasaputram derived the average ratio of demand deposits of companies to total demand deposits for the years 1956-1959 from the Quarterly Surveys. Assuming a con- stant ratio, he then extrapolated the data for pre-1956 years by multiplying thJs average ratio with the Bulletin data. No adjustment was made for hold- iigs by 4a-op-eratives or local authorities. There are also problems associated -uith the use of a constant ratio. Kahagalle used the same Survey data to derive the ratios of demand deposits held by individuals, non-business institutions and unincorporated enterprises to total demand deposits. He then applied the average of these ratios to the end of the year demand deposit holdings of the public as given in the Bulletin. Both Rasaputran and Kahagalle preferred to use the Survey data to derive average ratios of household sector holdings of demand deposits to total demand deposits, and then develop a continuous data series of household sector holdings by multiplying the average ratio to the data provided in the Bulletin. Certainly, the Bulletin data has the advantage of being continuous from 1950 to 1974, whereas the Survey data is only available from 1956 to 1974. However, the consequent need to use a constant ratio to derive the household -20- sector breakdown for the years prior to 1956 is a serious drawback; furthermore, it is not altogether clear why the Bulletin series of data on deposits should' be preferred to that provided in the Quarterly Surveys. The difference between the two series of figures on deposits held by the public is due primarily to the inclusion of interbank deposits in the latter. Even taking into account these interbank doposit, there still exists some discrepancy, which commercial banks explain by the fact that the monthly statements are prepared within a shorter period of time than the Quarterly Suverys, uhich are prepared over the time span of 1 to 11 months. According to Centrhl Bank authorities, the Quarterly Surveys -could be considered more reliable, although the Central Bank does use the monthly data for its money supply statistics, primarily for historical reasons (the Monevary Law Act of 1953 specifies that the money supply should be pre- pared on a monthly basis). It was, therefore, decided to utilise the Quarterly sw d.ta stacc thaT did provide a more useful breakdown of deposit holdings than the monthly data, and since the relative reliability of the monthly series over the quarterly was open to question. Questions can also be raised regarding Kahagallets interpretation of deposits held by .'other businesses" (item v in the Quarterly Surveys) as those of'unincorporaLed enterprises. Certainly holdings by unincorporated enter- prises neQd to be included in the household sector, along with those held by individuals and non-business institutions. There appears little basis, however, for assuming that deposits held by business institutions other than financial institutions, tea, rubber and coconut plantations, trading and manufacturing establishments (i.e., items i to iv in the Surveys) are those held by un- incorporated enterprises. - 21 - b) Savings and time deposits Estimation of time.and savings deposits held at commercial banks by the household sector encounters the same problem as in the derivation of demand deposits, i.e., how to identify the share held by the household sector. The method used resembles that used to derive holdings of demand deposits. Rom the Qaarterly Surveys end of the year holdings of savings and time deposits by non-business institutions and individuals were extracted for the years 1956-74. Earlier data were extrapolated from the total savings and total time deposits held by residents as given in the hulletin, Linear least-squares regressions were run for the period 1996-74 and the estimated coefficients applied to the annual data on time and savings deposits held by the public for the period 1950-55. Holdings of the household sector could then be derived. These data are presented in Table 2 of the Annex. The mpthodologies used by rsapitram 2nd Kahaglle were basicll the same as those used to derive data on demand deposits held by the household sector. They both derived ratios of savings adi time deposits of' the household sector to the total held at commercial banks, using data given in the Bulletin. "2 1/ (1) SD = -14.434 + 1.033 SD R =0.980 ,H (0.008) P (2) TDH = -0.469 + 0.252 TDp = 0.983 (0.023) where SDH savingb deposits held by households (data derived from Quarterly Surveys). SDP savings deposits held by resident public (data extracted from imonthly Bulletin). TD = time deposits held by households (data derived from Quarterly Surveys). H TD = time deposits held by resident public (data extracted from Monthly Bulletin). (Standard errors shown in parentheses). -22- In addition, Kahagalle made the same assumption regarding holdings of savings and time deposits by unincorporated enterprises as he did when estimating holdings of demand deposits. c) Liabilities to Banks To derive net household saving with commercial banks the change in aVanCeS and .oans made by banks to households needs to be deducted from the change in bank deposits held by households. Estimation of household liabilities encounters the familiar problem of identif7ing only those loans going to house- holds. Where no data are available on the corporate/non-corporate status of the borrower, information on loans granted according to purpose can be used. Certain purposes, such as housing and consumption, clearly 4_nply that the funds will be used by the household sector; with other purposes, such as agriculture or trading, further information may be necessary to derive some estinata of what proportion of funds used for such purposes are used by households. In the.case of Sri Lanka, information on loans granted by purpose available in the Quarterly Surveys was used to identify credit extended to households. The following purposes were interpreted as being directed to the household sector: wholesale and retail trading, housing, and consumption and certain percentages of loans made for agriculture and the processing of agri- cultural commodities. These percentages were derived from censuses of agri- culture which gave a breakdown of plantation estates by ownerships (individuals, and partnerships and corporate).. For paddy loans it was assumed that all were directed to households. One weakness of this methodology is its implicit assumption of constant credit allocation per acre of plantation. It may well be the case that the larger plantations receive a disproportionate share of total credit. In the absence of any suitable information on credit allocation - 23 - by size of plantation, it was decided to use the methodology described above. Since the Quarterly Surveys are only available from 196, data were a:trapolated from the Monthly Bulletin data. The outstanding loans and adrances to residents are available from 1949 - 74; extrapolation was done by using the coefficients derived from regressing the household sector liabilities with commercial banks on those of residents.1/ The complete data series for household sector liabilities with commercial banks for the years 1950-74 is givan in Table ). 3. Not Deposits with Savings Banks Until their takeover in 1972, the Post Office Savings Bank, the Ceylon Savings Bank and the Savings Certificate Fhnd accepted savings and time deposits from the public. In particular, the Post Office Savings Bank had con- siderable success in attracting deposits. Deposits held at the Yost Office and ' the Svig Cartificate Fund were as.umied to be owned ennircly by households. The Ceylon Savings Bank both received deposits from and extended credit to the household sector; after its take over in 1972 its activities have 2/ been primarily confined to the household sector-. In April 1972, the Ceylon 1/ Regression results were as follows: L = -1o.95o + 0.228 L R2 = 0.967 H (0.030) R where LH = household sector sector liabilities with commercial banks (depived from Cuarterl Surveys) and L a liabilities of residents with commercial banks (extracted from Monthly Bulletin data). (Standard errors in parentheses). 2/ The single exception is Rsl J.2 million loan to a Corporation in 1974. Savings Bank was amalgamated with the Post Office Savings Bank and the Savings Certificate FVnd and renamed the National Savings Bank. Data on the assets and liabilities of these institutions are available in the Central Bank's Annual Reuorts, and household saving with these saving banks is presented Table 5. '. Net Saving with Long-Term Credit Institutions In Sri Lanka, outside the commercial banks, these are three Lrnpor- tant institutions which are concerned only with extending credit, usuallyon on a medium - or long-term basis. These are the State Mortgage Bank, the Agricultural and Industrial Credit Corporation, and the National Housini Fund. For each of these institutions, estimates have been made of that portion of credit extended which goes to the household sector. This distinguishes these estimates from those of previous studies, where all the credit extended by these institutions was assumed to be directed to households. The rerhvPd estimates are presented in Table 6. a) State Mort,age Bank The State Mortgage Bank extends credit for residential construc- tion and for agricultural purposes. For the five years 1970-197h a break- down ofloans extended for purposes of residential construction, purchase of agricultural properties, and other agricultural loans was available. While there was no subdivision of loans by household/corporate sectors, authorities at the State Mortgage Bank estimated that approximately sixty percent of all agricultural loans ( including those for the purpose of agricultural proper- ties) have gone to the household se.ctor. All loans for residential construc- tion are, by definition, extended to the household sector. Cn the basis of this information, it was assumed that sixty percat of all agricultural loans were made to the household sector. For the years 1970-74 annual ratios were derived of the sum of agricultural and housing and housing loans made to households as a percentage of total made by the State Mortgage Bank. The average of these ratios was multiplied to the total loans of the State Mortgage Bank outstanding at the end of each financial year between 1950-69._ In this way, estimates were obtained on the Bank's credit to houiseholds outstanding at the end of each financial year for the years 1950-69. The data were then adjusted to a calendar year basis. b) Aricultural and Industrial Credit Corporation (A.I.C.C.) The A.I.C.C. extends credit for agricultur'al, industrial, mining and other purposes, but cannot finance housing construction, except for agricultural and development purposes. It also makes loans in denominations as small as Rs. 500. Previous savings estimates assumed that all A.I,C.C. credit went to the household sector; this clearly is erroneous, though to what extent is difficult to ascertain. It was not possible to obtain on an annual basis a breakdown of credit extended to the household and corporate sectors. Authorities at the A.I.C.C. did; however, provide the total amount of loans granted between October 1, 1949 and December 31, 1975 to 1joint-stock companies" and to "individuals, parternships and firms with unlimited liability". The ratio of loans granted ."to inividuals, partnerships and firms with unlimited liability to the total loans granted by A.I.C.C. over this period was then applied to the total loans outstanding of the A.I.C.C. for the end of each year between 1950 and 197?. In this way annual estimates of the household sector share could be derived These data were then adjusted from financial to calendar years. 1/ Given in the Annual Report of the State Mortgage Bank, various issues. - 26 - Utilising this ratio to derive advances made to the household sector obviously involves some overstatement, due to the inclusion of loans to private firms. However, it is unlikely that the margin of error is substantial, and in the absence of any additional information, it was felt that reasonable estinates could beJderived this way. c) National Hoising Fund In 1 954 the National Housing Fund was established under the National Housing Act. The Fund was to finance residential construction and to a lesser extent, the importation and manufacture of building materials. Funds were obtained from the issue of debentures and from- Govermient grants. As in the previous savings estimates, all advances from the National Housing und have been assumed to go to the household sector. Any overstatement whichi this involves, due, for example to credit extension to building societies or to manufacturing or trading firms involved in the construction business, is likely to be small. This is corroborated by the very high percentage of loans in small denominations (in the mid-1960ts more than 90 percent of all loans were betieen Rs. 2000 and Rs. 200000- ). Data were obtained from the Central Bank, Annual Reaort,various issues, and adjusted from the financial to calendar year. 5. Net saving with Hire Purchase Finance Companies Hire purchase finance companies have become increasingly active in extending credit and mobilising deposits in Sri Lanka. Most of the credit operations are for hire-purchase business, primarily transactions involving motor vehicles and agricultural machinery. In 1970 the financing of industrial machinery, household durables etc., accounted for less than 1 percent of the 1/ H. N. S. Karunatilake (1968) p. 153. - 27 - outstanding volume of hire purchase business. Certain companies do also extend loans, though this is not so common as hire purchase financing (out of 41 companies extending hire purchase finance in 1970 only 29 were granting loans). In terms of the total volume of credit outstanding, the hire purchase finance companies have ranked considerably higher than any other non-bank financial institution during recent years, and the volume of out- standing credit to households had risen to almost one quarter of that extended by commercial banks by 1970 although declining to about one-eight by 1974. -During the 1970's they have also been increasingly successftl in attracting deposits. Considering the significance of hire purchase finance companies as a source of household credit and mobilizer- of funds, it is unfortunate 2/ thnt the datp on their activities are so incomplete. The Central Bank began to systematically collect such information only from 1969; this is published in the Annual Reports from 1970 onwards. It is clear from the amounts of hire purchase credits and loans already outstanding as of December 1969 (Rs. 79.4 and 16.5 million, respectively),that the companies had been.in operation for some time . Besides the Central Bank Annual Report, additional information on the activities of hire-purchase finance companies was obtained from two sources: 1/ Central Bank of Ceyltn, Annual Report, 1970, p. 118. 2/ The inadequacy of data obliged the authors of the previous savings studies to ignore deposits and advances of hire purchase finance companies in their estimates. - 28 - (a) Handbook of Rupee Companies. A few of the companies were public and their balance sheets were published in the Handbook, an annual publication put out by the Colombo Brokerz; Association. From this source deposits received and outstanding credit extended by this group of companies were obtained annually for the period 1950-1974. (b) Questionnaire returns for a Frivate Investment Survey con- ducted by the Central Bank of Ceylon, which provided balance sheets for 14 hire-purchase finance companies for the years 1964, 1965 and 1966. For each of these three years, therefore, data was available on the deposits received and outstanding credit extended by this group of companies. The summation of these two groups was considered to be reasonably representative of all the hire purchase companies in operation during these three years. From these three data sources, the deposits received and credit extended by a21 hire purchase companies can thus be derived for nine years 1964 to 1966 and 1969 to 19714. For each of the years 1969 to 197)4, the deposits and the credit extended by those hire purchase companies specified in the Handbook comprised almost sixty percent of the deposits and credit extended by all hire purchase companies, as given in the Central Bank, Anntal Report. By multiplying this ratic to the data on the "Handbook'" group of companies for the years 1967 and 1968, the net deposits received by all companies could be derived. In order to obtain data on net deposits prior to 1964, it was-assumed that the "Survey" group of companies all began operations in 1958 - and intermediate data were then interpolated. This derived data were summed with net deposits of the "Handbook" group of companies to derive the net deposits of all companies. For the years prior to 1958, the net deposits of the "Handbook"group of companies were assumed 1/ The median year of registration for this saimple was, in fact, 198. - 29 - to represent those of all companies. The procedure used to develop these estimates,presented in Table 6, has been explained in some detail since it does provide an interesting example of how, when faced with no basic data series, data can be built up using previously untapped sources. Construction of such data, however, should be accompanied by an independent check for accuracy. in this case, corroboration was obtained from census material. 6. Net Claims on the Government Net saving by the household sector with the Government consists of the change in holdings of Government securities less the change in any credit extended by the Government to households. The estimation is not usually a problem, since information on the ownership of public debt is generally available. For Sri Lanka claims on the Government by the house- hold sector were derived from registers in the lublic Debt Department of the Central Bank. These registers give a breakdown by ow:aership of holdings of Government debt. Individuals, clubs and non-profit institutions were considered to constitute the household sector. The claims on the Govern- ment consisted. of Rupee Loans, War Loans, Treasury Bills, and, from 1956 ownwards, Tax Reserve Certificates. Adjustment to a calendar year basis was necessary. Net saving with the Government is arrived at after deducting the 2/ change in household sector liabilities with the Government. Direct credit 1/ Corroboration of the derived figure for the total amount of outstanding credit extended by hire purchase finance companies in 1951 was obtained from the Census of Financial Institutions (1952) conducted by the Depart- ment of Census and Statistics. 2/ The Government.is important as a .source .of secondary finance channelling funds through such institutions as co-operatives, the National Housing Fund and the A.I.C.C. -30- 1/ granted by the Government to the household sector consists primarily of loans made to Government employees for the following purposes: salary loans, transport loans, property loans, and loans to relieve indebtedness. Information on the total outstanding credit from the Government to its employees was obtained from a new centralised authority, the Credit Council. After adjusting to a calendar year, net saving of households with the Government could be derived; these data are presented in Table 7. 7. Yet Claims on Co-operatives and Rural Banks. Net saving with co-operative societies (including rural banks) consists of the change in deposits with these societies plus the change in share capital purchased less the increment in credit extended by these societies. In Sri lanka, the co-operative societies have been active in the iL.d. oi credit extien6ioii, particuldrly in rural areas, throughout the period under study. Prior to 197, credit was extended by the co-operatives Irimarily through credit societies through the Co-operative Agricultural Production and Sales Societies (C.A.P.S.), and through retail consumer societies. These societies obtained funds from the public through the issue of share capital, and from the Government in the form of loans. Re- organization in 1957 concentrated credit extension, as well as retail sales nd purchase of rural produce in the hands of the Multi-Purpose Co-operative Credit Societies (M.P.C.C.S.). Information on outstanding loans by these various types of co-operative societies, together with their share capital 1/ In addition; the government has directly assisted various agricultural development schemes, such as the peasant colonization schemes, the middle class land alienation schemes and renabilitation schemes. The annual amount-of credit granted, however, has been relatively small. Further details are given in Karunatilake (1968). p. 100-102.. -31 - issued, were obtained from the Administrative Reports of the co-operative societies and are presented in Table 8. Shortly after its information in 1961, the People's Bank selected certain M.P. 0. C. S. which were to set up Rural Banking Departments for accepting deposits from rural households. From 1964 this scheme of rural banks became operative, and their activities have subsequently expanded rapidly. Rural banks obtain funds by means of overdraft facilities with the People's Bank and by receiving redeposits from certain M.P.C.C.S. which accept time and savings deposits from both members and non-members. Rural banks extend credit and also engage in pawnbroking, which, in 1973, amounted to 58 percent of all advances. One can reasonably assume that all- credit c.tand d by r -1 banks wnt to households. The Central Bank publishes information on outstanding loans and deposits of rural banks in its Annual Report*. Net saving through rural banks-is presented in Table 8. 8. Net Corporate Securities In Sri Lanka household sector holdings of corporate shares and debentures comprise a relatively minor share of that sector's asset holdings, the small investor being discouraged by a high divident tax (33 1/3 percent) as well as by the lack of an easily accessible, active stock market. -32- It can be extremely difficult to estimate holdings of corporate securities by households. Information on the ownership (i.e., household/cor- porate) of equity stock are not readily available due to confidentiality r.gula- tions. Goldsmith measured household savings throughcorporate secirities in- directly, by subtracting from the net purchases or sales of securities by their issuers, the purchase or sales balance of non-individual holders and by adding to the difference commissions and taxes paid by individuals on their transactions on already outstanding securities.- Insufficient data made this approach im- posible ,for Sri Lanka. Instead, the followaing rather lengthy procedure was -necessary. Information on changes in the paid-up capital of all registered com- panies in Sri Lanka were obtained from the Administrative Report of the Registrar of Companies. In order to derive the share of securities held by the corporate sector, one of the leading brokerage firms in Sri Lanka supplied a breakdown C)~ w16,ult bIt.~ It- D-ned 'n'o pa,bases 11% cvYYian*,es and by fos~oPsTr cextain benchmark dates, fitting a curve to these benchmark ratios, a series of ratios could be derived anual3y for the years 1950-1974. The application of these ratios to the total paid-up capital for the corresponding year provides the household sector holdings of corporate shares.(presented in Table 9). 1/ See Goldsmith ( 1955) p. 284-357. 2?/ In SIi Lanka, both fiscal as well as control -considerations motivate companies to purchase shares issued by other companies. 3/ Values were derived by using an ordinary oscularoty interpolation, Karup- King formula. -33- Ra8aputram and Kahagalle derived data on household sector holdings of corporate securities using an average ratio of investment to paid-up capital taken from balance sheets of a group of companies presented in the Handbook of Rpee Companies. They assumed that such "investments" consisted only of the holdings of the securities of other companies, and blew up these ratios on the basis of total paid-up capital to obtain total "investments" by companies. Annual holdings of corporate shares by companies were then subtracted from the total; changes in the residua.l, household sector holdings, could then be derived. A major weakness of this methodology is its assumption that all items described as "investments" in companies' balance sheets consist only of corporate securities. In fact, the breakdown adopted in these balance sheets is such that a variety of financial assets are likely to be included in this "investment" category. The methodology of all three studies on Sri lanka suffers from the disadvantage that the balance sheets of only public companies are used to derive the ratios, private companies are not taken into account and it may well be the case that new issues of private companies are held primari2y by individuals. While no accurate assessment is possible, the estimates of the household sector's holdings of corporate shares may well understate that sector's holdings. 9. Net Saving through Provident Funds The procedure for estimating savings through provident funds requires some elaboration. Net savings through provident funds are defined as the difference between total receipts and total expenditures. by providenmWt funds during an accounting period. They consist of the surplus genexated. This is measured as the sum of the contributions received frozz employees and employer and interest received on investments less the sum of the refunds made to employees, administrative expenses and interest paid. In balance sheet terminology this is the equivalent to the increase in total assets, less valuation changes. In Sri Lanka there are two kinds of provident funds currently in operat.on one which is run by the Government (the Diployees Provident Fund or E...F.) and several which ate privately organized. a) Saving Through the Employees Provident Fund By far the most important of the provident funds in operation in Sri Lanka is the E.P.F., which was established in 1958. Employers of those workers covered by the E.P.F. are obliged to contribute at least 7 percent of the employeets gross salary, while employees contribute a minimum of 6 percent. Em- ployees are permitted to contribute up to 40 percent of their salary, although employers are not obliged to match it. All investments made by E.P.F. are in gcvernment securities; no advances are made to policy holders. While the E.P.F. is administered by the Commissioner of Labour, the Central Bank of Ceylon has the responsibility for maintaining the accounts. Information on the amount of contributions, interest received on investments, 1/ See Goldsmith (195) p. 281 - for an elaboration of this definition. Also Reviglio ( 1967 ) P* 325. refunds, working expenses and "interest paid" were available on an annual basis for the years 1958-1974 from the E.P.F. Department of the Central Bank. This last item "interest paid" constitutes the interest of 2.5 percent which 1/ is paid on the closing balance at the end of each year.- Annual inflows, outflows and net saving of households through E.P.F. (saving is already- in net terms since no credit is extended by E.P.F. to its members), are shown in Table 10 . Net saving has been grow,7ing steadily over these 17 years, rising from 11 million rupees in V959 (or 2 percent of total household saving) to more than 120 million rupees in 1974 (or 7 percent of total household saving). In the previous estimates of Sri Lanka savings, total investments by- E.P. F. in government securities were interpreted as household sector saving through this institution. While this estimate may, under quite specific circumstances, be equivalent to that derived by the method described in the first paragraph, it is more likely that the former estimate would involve a margin of error. Unfortunately, no direct comparison is possible since the previous estimates give no breakdown for savings through E.P.F. and private provident funds, only the total figure is published. b) Saving Through Private Provident Funds In addition to the E.P.F., certain large firms operate provident funds on a private basis. Total contributions received by- these private pro- vident funds in 1973 constituted approximately 20 percent of those received by the E.P.F., compared with 45 percent in 1960. 1/ As Karunatilake (1968) p. 181,.has pointed out, by paying interest on a per annum basis, members are effectively receiving a higher rate than 2.5 percent per annum. -36- While it is true that the private provident funds have become increasingly less important relative to the E.P.F over the last decade, they do, nevertheless, constitute a form of household sector saving and require estimation. Data on inpayments and outpayments of these private provident funds are not readily available. The Commissioner of Labour did, however, provide information on contributions, refunds, and outstanding loans and advances on an annual basis for all the private provident funds operating during the years 1958--to 197-. The private provident funds, unlike the Employers Provident Fund, do extend credit to policyholders. Using the same methodology as for E.P.F., net household saving through private provident funds could be derived from these data annually for the years 1959-1974 (Tablell ). For earlier years, numerous attempts were made to try and extrapolate data. Various regression techniquer, were used (with both linea and quadratic functions), but the standard errors were too high to permit even a moderately good estimate to be derived. Fitting an exponential function was also tried, tfsing least squares method of estimation. The resulting standard error was again too high. Consideration was also given to using growth rates; however, given the high degree of variation in the data for the years 1958-1974 the use of an annual average growth rate to Extrapolate was deemed inappropriate. Rather than develop data for years prior to 1957 which would be subject to a wide margin of error, it was assumed in this study that there was no net saving through private provident funds in earlier years. 10. Net Saving in the form of Life Insurance Conceptually, one of the confusing estimation procedures concerns - 37 - 1/ saving in the fon of life insurance. Several studies7 have interpreted such saving as being the total premia paid for life insurance policies within a given period lecs benefit payments paid and less consumption loans granted by the life ,insurance companies. The point to remember is that saving through life insurance occurs when an individual purchases a policy for full life coverage, and that this purchase involves not only an-annuity but also the coverage for risk. It is this annuity element that comprises a part of household sector saving. Saving in the form of life insurance should, therefore, be measured not as net premia paid, but as the change in the life insurance 'fund less the change in outstanding loans made to policy-holders. This is the 2/ 3/ estimation procedure described by Goldsmith and the N.C.A.E.R.~ and has been used in the current estimates. 1/ Example, the series of studies of savings commissioned by E.C.A.F.E., and includ- ing Hooley'4 work in Philippines (1963), Rasuputram's work on Sri Lanka (1961 ) and a'study on Malaysia (1962). A more recent work on Sri Lanka done by Kahagalle (1975) also used this procedure. 2/ Goldsmith k95, pp. 268-278. 3/ Actually, N.C.A.E.R. used the revenue accounts to derive savings through life insurance. Saving was measured by the n surplus of total premium income and other incomes, net of payments for current expenses, including charges on re- serves, as given in the revenue accounts, represents the addition to the life fund ever,r year" (p. 68). - 38 - In Sri Lanka both private and public insurance companies are engaged in the business of life insurance. The nationalised company, the Insurance Corporation of Ceylon, was founded in 1961, and had begun its operations by 1962. At that time it was agreed that the private life insurance companies then in existence could not subsequently open any new accounts; they -could, however, maintain operation of thosef-policies already in existence. The Insurance Corporation of Ceylon has expanded its business rapidly, both in terms of policies issued (it deals not only with life insurance; policies are issued for accident, fire, marine as well as motor insurance), and in terms of loans extended. The latter are comprised of loans to policy-holders for purposes not specified and (since 1970) for housing purposes, loans to I.C,C. staff, and loans to field officers. Field officers consist of the agents and organizers scattered over the island. While they do not receive an/ salary, they do receive commission on any business they bring to I.C.C. and are entitled to loans from I.C.C., with repayments and interest being deducted from their commission. Net saving by the household sector in the form of life insurance with I.C.C. is derived from the change in the life insurance fund,less the change in outstanding loans made to policy-holders. Loans made. to staff and field officers, while they certainly should be included in the liabilities of the household sector, are not granted because the borrower is an I.C.C. life insurance policy-holder; they are granted because the borrower is employed by I.C.C. They are, in effect, salary loans. Only loans made to policyholders, whether for purposes not specified or for housing (housing loans have been -39 - granted since 1970), are therefore deducted from the change in the life insurance fund to derive net saving in the form of life insurance with I.C.C. This information, together with the liabilities of policy holders with the I.C.C. n-r presented in Table 12o Salary loans made by I.C.C. are given in Table 13. Information on the operations of the private life insurance companies was particularly difficult to obtain. The data- provided in Statistical Abstract, which was utilised by both Rasaputram andKahagalle, was felt to be insufficient to derive satisfactory estimates. Additional information was obtained from the Controller of Insurance giving details on the activities of the private life insurance companies operating between 1958 and 197L. These data provide6d a basis for estimating the changes in the total life insurance fund of private life insurance companies and variations in loans and advances which these Comr1paie6 extended during the period 1958-7L. Information on net premia received and crdit extended by these companies available in Statistical Abstract from 1950 was used tQ pxtrapolate data prior to 1958. The cmplete data series is presented in Tablelh 11. Tangible asset holdings Lack of reliable data makes investment by households in tangible assets one of the most difficult estimates to derive. The problem is serious in view of the fact that savings in the form of tangible assets usually comprise a rather large share of the total. This was true in Sri Lanka during the period under review. It should be borne in mind, however, that the household sector as de- fined in this study, includes unincorporated enterprises. In view of the signi- ficant activities of the unincorporated business sector in Sri Lanka, it is likely that such enterprises are responsible for much of the direct investment of the household sector. No consistent series of data on the tangible asset holdings of 1/ the household sector- is available for Sri Lanka. Partial information, however, can be obtained from three different sources. These are: a) The Central Bank of Ceylon, Annual Report, various issues, which present annual data on the gross fixed capital formation of the "the private sector" as a part of the National Accounts for the years 1959-1974. Private sector here includes the household and corporate sectors, as well as Government corporations.. For most years information is available for the private sector exclusive of Government corpora- tions. For those years where they were not excluded, the average ratio of gross fixed capital formation of the private sector only to that of the private sector including Government corporations was applied to the data for. the private seobor inclusive of Government corporations. In this way the gross fixed capital forma- tion of the private sector only could be derived. (b) The National ccounts estimated 2/ by the Central Bank of Ceylon provide data on GDP to current factor cost prices for the period 1950-1973, but those of gross fixed capital formation only from 1959. The average share of the gross fixed capital formation of the private secto. to GDP for the years 1959-197h was therefore multiplied by the GDP data for the years 1951-1958 to derive the gross fixed capital formation of the private sector for these years. c) Private Investment Surveys conducted by the Central Bank of Ceylon in 1954 and 1966 provided the information necessary to derive the household sector share of the private sector gross fixed capital formation data. 1/ In this study, expenditure on residential construction, vehicles, machinery and equipment and land planting and replanting was considered household saving in the form of tangible assets. 2/ See T. Savundranayagam (1974). d) It would be desirable to develop estimates of the respective shares of residential construction, transportation vehicles, plant and machinery, land improvement, etc., in these gross fixed capital formation figures. Attempts to derive such data for residential construction frcm information presented in the Statistical Abstract were unsuccessful. This annual publication by the Departmont of Census and Statistics. givce annual figures for new residential construction by Government Departments, municipalities,. urban and toin councils and under various building sahemes. The summation of these data, however, comprises only abo-t 10 percent, of construction data presented in the Central Bank Survey of Private Investment for the financial year 1966/67. This amounted to FRs. million 315.i comnpared with the Statistical Abstract figure for 1965 (later years are unavailal,le) which amou'bed to Rs. million 38.9. It was decided, therefore, to use the Survey of Private Investment to derive the ratios of the shares of investment in residential construction, vehiclc, machiney ad equipment and land (including planting and replanting) in the total household sector investment to the total household sector investment data derived earlier. Assuming the applicability of constant ratio throughout the study period, these ratios have been applied to the household sector gross domestic capital formation figures derived earlier. In this way, the household sector's investment in residential construction, machinery and equipment and land, including planting and replanting for the period 1950-1974 can be derived. These data are presented in Table 14. Depreciation Appropriate adjustment for description on the gross value of tangible asset is necessary to develop estimates of household sector net investment in such assets. However, few developing countries have adequate information on the relevant depreciation rates or magnitude of depreciation for the various types of tangible assets. Such adjustment then becomes extremely difficult. In any single year depreciation on a fixed asset is ccmposed of the depreciation provided on the toal value of assets which have accumulated up to the period, as well as the depreciation allowed on the Assets newly acquired in that particular year. Depreciation in year t would thus be estimated on the total value obtained by the addition of the accumulated value of the asset up to year t to the gross investment in tangible assets in year t. For the year 1951, to give an example, net capital formation would be derived by subtr*acting the depreciation from gross investment; the net value of accumulated assets at the end of 1950 is obtained by deducting depreciation from the total gross value of assets. Net investment in 1951 is, therefore, derived from the net value of assets existing at the end of 1950, and the gross capital formation for 1951, depreciation being deducted at the same rate for both years. This is the methodology adopted by Goldsmith and N.C.A.E.R. in their studies. - Most of the other savings estimates referred to earlier were unable to develop depreciation Estimates due to the lack of adequate data. This was the case for the current estimate, where the data base was such that it was not -ossible to develop any reasonable estimate of the appropriate rates of depreciation, or of the value of accumulated assets. It was decided, therefore, to leave tangible asset holdings in gross form. 143 IV. Summary of Findings 1/ Net saving of the household sector grew from Rs. 4011 million in 1951 to Rs. 1830 million by 1974 in current prices (see Table I),. corresponding to an increase at an annual compound rate of 6.5 percent. In constant prices,. the compounded rate of growth was only 3.4 percent. To eliminatc short run fluctuations, five-year averages of total household saving to GNP were estima- These average savings ratios rose steadily, from 7.5 percent for the period 1951-5 to 11.3 percent for the period 1971-74. Household savings have been primarily in the form of tangible asset3 throughout this twenty-four year period. This is evident from Table II, which presents a breakdown of household sector saving into net financial anti tangible assets, and annual averagles over five-year periods. Saving, in the form of tangible assets (primarily residential construction) accounted for an average 2/ of 7 percent of total household saving over the entire period.- This percentage did fluctuate during these years, with the share of saving in financial form in- creasing over the period 1962-67 (when it averaged 37 percent) and 1970-73 (when it averaged 29 percent). The share of saving in tangible form may be an over- I/ As explained earlier, net saving of the household sector is the change in assets less the change in liabilities of households. Gross saving refers to the change in asset holdings only. 2/ This percentage is similar to that estimated in other studies; for example, the N.C.A.E.R. estimated 67 percent of household saving in India over the years 1950-61 was held in tangible form, the estimated percentage in Malaya for the years 1954-58 was 58 percent, for Philippines the average percentage for the years 1951-60 was h9 percent, although this includes only residential construction and consumer durables. The percentage was much lower in Japan for the years 1951-58 (22 percent); this was the only estimate where depreciation allowances were deducted. estimation for two reasons: first, no depreciation estimates cocld be derived for tangible asset holdings, so they remain in gross forn, and secondly, for reasons explained earlier, unincorporated enterprises are included in the house- hold sector, and savings of these enterprises are likely to be-reinvested in the business rather than in some form of financial asset. A closer examination f the structure of household sector saving is possible from Tables III, IV and V. Table III presents annual changes in assets and liabilities held by households over the twenty-four years, while in Table IV and V, annual averages over five-year periods of the share of different financial assets in total net financial saving (Table IV) and in gross financial saving- (Table V) are presented. From these tables it is clear that the pattern of financial asset accumulation of households has changed considerably over the quarter century. While some of these changes have little observable trend, ono of the most strik- ing developments.is the increasing share of savings in contractual form, particu- larly through provident funds. While there are deficiencies of data on the 3/ operations of private provident funds it is apparent that with the formation 1/ That is the change in financial assets held by households less the change in credit extended to households. 2/ That is the change in financial assets held by households. 3/ As explained in the previous section, no information on the activities of private provident funds was available prior to 1958, so this upward trend during the 1960's. may exaggerate the actual situation. However,. the extent of the error is unlikely to be large since the coverage of these private provident funds was limited and the rates of -contribution generally small. of the State-owned Ehployees Provident Fund (E.P.F.) in 198, contractual savings rose rapidly, reaching a peak during the period 1966-70 (when they- comprised 60 percent of total net financial savings of hous,holds). On t1l average, contractual savings accounted for more than 40 percent of net financial savings of households over the entire period. Questions can be raised regarding the impact of this rapid increase . 1/ in saving in the form of provident funds on the composition and volume of s&ving Did it induce people to reduce saving by guaranteeing retirment benefits, or stimulate saving by "setting a pattern for retirement rather than a lifetime of 2/ work'?- There is also the question of asset substitution. If one assumes Ihat households' saving behavior is deter=ined by sme savings or wealth target, then saving in contractual fonmay well be a substitute for holdings of otiier fin9ncial assets. Substitutability is, however, limited by the fact that contractual savings possess attributes which are quite distinct from other financial assets. In the case of the E.P.F., savings with this Fund are illiquid, both in the sense that withdrawal-is possible only on retirement, and in the sense that r credit 3/ is extended to policy holders.- While further research would be necessary to 1/ There is a growing body of literature dealing with this issue; see, for example, Cagan (1965), Aaron (1967), Katona (196), and Munnell (197,1). 2/ Munnell (1974) p. XV. 3/ In this sense, insurance funds have a greater degree of liquidity in Sri Lanka since it is possible to borrow against the policy. to investigate the relative impact of these substitution and scale effects, it does appear that the introduction of the E.P.F. has had a positive effect on the volume of household savings. One indicator which supports this is the ratio of total household savings to GNP which, on average,- has been steadily increasing since the introduction of E.P.F. However, since employees must also contribute to the E.P.F., there may be a negative effect on corporate savings. Besides provident funds, deposits with commercial and savings banks constitute the other major component of aousehold saving in Sri Lanka. The volume of household assets held as bank deposits has fluctuated widel7 over these twenty-four 7ears, and on average, has comprised one third of total 2/ household saving.- In the early 1950's household deposits with commercial and savings banks were particularly large. Thereafter, the accumulaiton of savings bank deposits remained sizeable, and consisted primarily of deposits with the Post-Office Savings Bank until the Bank's takeover in 1972, and of deposits with its successor, the National Savings Bank. The success of the Post Office Savings Bank in attracting deposits is worthy of special note, especially with regard to the early 1950's. During this period, the commercial banks and the provident funds had not yet become significant mobilizera of house- hold saving, and the Post Office Savings Bank proved to be a popular savings in- stitution with households, particularly small savers. The volume of savings with the Post Office Savings Bank remained fairly constant until its takeover, although the share of these savings in total household savings 1/ That is, average ratio over 5-year periods; see Table I column (h). 2/ This share applies to both gross and net financial saving. - 47 - 1/ decll'ned7 as savings through provident funds and commercial bank;s grew. In contrast to deposits with savings banks, there has been a noticeable trend of households shifting away from holding deposits with commercial banks from the mid-1950's. Household saving with commercial banks never regained its share in the total which was achieved during the period 1951-55.(50 percent of net and 40 percent of gross financial saving of households) During the following ten years, households preferred to save through provident funds, savings banks and in the form of currency rather than with commercial banks. This trend was reversed somewhat during the second half of the 1960's, a period marked by a rapid expansion of bank branches and by an increase in the deposit rate offered by commercial banks. (The minimum rate offered on 12-month fixed deposits rose in stages from 2.50 percent in 1967 to 4.50 percent in 1970). From 1971, however, net household saving with camercial banks became much less important, as deposits grew at a slower rate and bank credit ex- tended to households increased markedly. Not only has there been a decline in the share of household saving with commercial banks, but the composition of this saving has also been changing. From the mid-1950's household saving with commercial banks has been increasingly in the form of time and saving deposits. Holdings of demand deposits, while highly volatile over the ienty-four years, have become a much less important form of household saving, declining from an average share of almost 5 percent of gross financial saving of households over the period 191-55 to an average 1/ The year before its takeover, 1971, was exceptional to this trend, and deposits with the Post Office Savings Bank rose sharply, apparently in response to an increase in the rate of interest offered on such deposits. h8/ share of 0.4 percent over the'years 1971-74.- It appears that this latter development can, in part, be attributed to a switching from demand deposits to currency as a means of transaction, currency holdings by households showing a marked tendency to increase as a share of total household savings 2/ from the mid-1950's: A possible explanation of this declining importance of demand deposits and the growing preference for holding currency may lie in changes in income redistribution in favour of the lower income groups which was substantial over the last two decades in Sri Lanka. According to surveys of consuner finance undertaken by the Central Bank of Ceylon in 1953, 1963, nnd 1973, the Gini coefficients for spending units declined from 0.46 in 1953 to 0.45 in 1963 to 0.35 in 1973. It is likely that this redistribu- tion of income would exert a substantial impact on the composition of the money stock, since those benefitting from the redistribution, i.e., the lower income groups, could be expected to prefer using currency rather than demand deposits for transactions. The "long-term credit institutions" (i.e., the National Housing Fund, the Agricultural and Industrial Credit Corporation, the State Mortgage Bank and the Hire Purchase Finance Companies) were a major source of household 1/ The huge increase in demand deposits held by households in 1971 was exceptional, and can be attributed to the demonetization of currency notes which was undertaken in that year with the intention of taxing undeclared wealth in the form of currency. See Kahagalle (1975), p. 57. 2/ Looking at annual averages over 5-year periods, the exception to this trend is the period 1966-70 when currency holdings by households comprised a much reduced share of total household savings -49- credit during the decade of the 1950's when they supplied one-quarter of all credit extended to households. In later years this role was taken over by pension funds (during the 1960's pension funds supplied almost one-half of household credit) and, in the late 1960's and 1970's, by co-operatives. During the period 1971-74 this group of "long-term credit institutions" became a net mobilizer rather than a net supplier of funds to households reflecting the increasing success that the hire purchase finance companies were having in attracting deposits. No trend is observable in household savings in securities, and, on average, securities comprised a negligible share of total net financial saving of households over the twenty-four years (2 percent), though a rather larger share of gross financial saving of households (9 percent). The difference is due to the net outflow of funds from co-operative institutions to households, which became especially important during the second half of the 1960's and the 1970's. The tendency for households to dissave with the Government is explained by the practice of the Government extending salary loans to its employees, and the annual amount of credit extended exceed, on average, the purchases of Government securities by households. A Comparison with Previous Estimates Table VI presents the three sets of estimates of household sector saving in Sri Lanka. (1) those derived by W. Rasaputram (1961) for the period 1951-1959; (2) those derived by S. Kahagalle (1975) for the period 1960-1972; and (3) those derived in this study for the period 1951-1974. The most obvious difference between the three series of estimates is the consistent understatement of Rasaputram's figures compared with those of the current study, and also compared with Kahagalle's estimates for subsequent years (note the jump between Rasaputram's figure for 1959: Rs. 275 million, and Kahagalle's for 1960: Rs. 687 million). While this can be attributed to a variety of methodological differences, the most significant difference lies in the estimation of tangible asset data. Unlike in the other two studies Rasaputram estimated the change in tangible asset holdings in net terms and as a result, his estimates of the change in these assets are substantially lower than in the other two studies. In this study, as in Kahagalle's, it was decided that the limitations of data were such that it was not possible to develop any reasonable 1/ depreciation estimate.- Tangible asset holdings do, therefore, remain in gross form. Compared with the estimates of the change in tangible assets, the discrepancy between the estimates of the net change in financial assets is not so large, although the present estimates are generally smaller than those of both Rasaputram and Kahagalle. The methodological differences accounting for these discrepancies were explored in detail in the previous section. In summary, the principle differences are accounted for by the different methods of deriving estimates of saving in the form of currency, commercial bank deposits, and provident funds, together with the incorporation of previously un-tapped data concerning the operations of life insurance companies, provident funds and hire-purchase finance companies. 1/ Rasaputram himself emphasises the inadequate nature of the data base; see pages 20 and 21. -51- Table I Household Sector Savings In Gurrent and Constant Prices: 1951-1974 Current Prices Constant . O() Over (1959) Prices GNP : Period specified (Rs. illion) (per cent) (1)(2 ) (3 ) (4 ) 1951 403.7 452.6 8.9 1952 206.2 219.0 4.7 1953 . 210.4 243.1 4.5 7.5 1954 454.6 496.4 9.2 1955 498.6 527.59.5 1956 408.1 135.0 7.8 1957 343.6 367.6 6.3 1958 390.1 396.7 69 1959 624.5 624.5 10.6 . 1960 556.7 556.9 8.9) 1961 503.9 512.8 8. 0 1962 525.8 542.6 8.1 1963 730.3 71.3 10.7 9.0 1964 723.1 730.3 9.9 1965 621.6 627.2 8.3 1966 650.0 66m. . 8.· 1967 893.9 880.0 10.8 1968 909.0 819.3 9.2 10.0 1969 - 941.6 817.9 8.8 1970 1436.0 1203.1 12.4 1971 '1580.8 1299.7 13.4 1972 1581.9 1252.2 12.5 11.3 1973 1733.8 1187.9 11.4 1974 1830.3 997.2 9.3 Source: Annex Tables. - 52 - Table II: Household Sector Savings in the form of Financial and Tangible Assets: 1951-1974 (Rs . Million ) Changes in Holdings of Total Saving (1) 1(3) Net Financial Tangible (1) + (2) Assets Assets .._--.__ (1) Annual (2) Annual nua Annual (2) 3) Pnual 1(4) Average Average Average Average 1951 106.7 297.0 ho3.7) 26.4 1952 -82.5 288.7 206.2 ho. O 1953 - 96.8 399 307.2 i31.8 210.4( (35L-7 -6.o 711.2 1954 124.6 330.0 454.6 27.4 1955 147.4 351.2 498.6) 29.6 1956 53.1 355.0" 408. 13.0 1957 -26.2 369.8 o.32 h6h .6 7.6 13.3 61.6 l 103. 0\,L163. 1958 3.9 386.2 390.1 1.0 1959 129.7 h94.8 624.5! 20.8 1960 147.6) 409.1Y 556.7 26 . 1961 57.9' h46.0 503.9 11.5 1962 99.1 1 426.7 ,52.8 6 913. S03417.3 11 kl6909 i1828 1963 288.3 k203.7 4420 3303 1 39.8 1964 288.8 434.3 23.1' 39.9 1965 284.2) 337.4 62 l6 45.7 1967 30> 158216j 1966 159.9 l 490.1 650.0 )24.6 1967 309.7 -.2427 58D.2 723.4 893.9 ý966.1 34.6 . 1968 247.7 A2.7 661.3 - 7909.0 27.2 25.1 1969 77.8 863.8 . 941.6 1 8.3 1970 418.2 J 1017.8 1436.0 29.1 1971 537.6 1043.2) 1580.8 34.0 1972 503.4 497 107 870 i 1581.9"6P1.7 31.8 29 1973 508.1 • 1225. ' 1733.8 29. 9 1974 429.4 1lc0.9 1830.3 235 (1951-1974 4717.9 .14040.6 18758.5 25.2 Source: Annex Tables. - 53 - Table III: CHANIE IN ASSITS AND LIABILITIE OF THE IIOEHIOLD iECOrH 1951-714 (par- 1) (hs. Million) 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1. Change In currency holdings 21.8 -136.5 - 56.2 18.5 31.7 79.1 - 96.7 9.1 39.3 -6.9 87.3 43.8 2. Change in Net Conumercial Bank Deposits 29.8 -54.9 - 32.2 92.8 70.2 - 12.7 7.0 -.13.1 18.6 27.14 1.5 - 13.4 Change in Gross Deposits 3 -5T- -29.5 97.1 71.92 --i12 20.1 -2W 30.7 - 1.3 365 Change in Liabilities 8.5 - 1.3 2.7 4.3 1.7 - 1.5 13.1 5.1 9.8 3.3 - 2.8 19.9 3. Change in Net Savings Bank Deposits 46.9 18.5 - 11.7 - 7.0 114.4 26.7 22.6 19.9 31.6 25.9 8.1 9.14 Change in Gross Deposits - 2TF --W5 -l-- 27.1 20.5 31.7 -2 -5. 10.7 Change in Liabilities 0.9 1.1 2.3 2.5 0.2 0.4 - 0.8 0.6 0.1 - 0.1 0.7 1.3 1t. Change in Net Deposits with Long-term Credit Institutions - 6.4 - 6.5 - It.3 - 3.5 - 3 - .7 - 6.8 - 6.2 - 12.6 - 13.5 - 3.1 - 3.9 Change in Gross Deposits 0.3 0.1 - 0.2 0.1 0.3 F 0.9 1.2 0.7 -7 02 - 0.3 Change in Liabilities 6.7 6.6 4.3 3.7 3.6 50 7.2 7.1 13.8 14.3 2.9 3.6 5. Change in Net Claims on Government 0.8 - 3.2 2.7 0.6 -27.8 -11.1 - 4"8 0.1 5.1 13.8 1.2 Gross Claims T -2.3 - - 2 - 3.3 - 3.1 - 0.2 - 1.7 S 6 47-U 33.3 0.3 Change In Liabilities 5.4 0.9 - 5.5 - 1.8 -213.5 8.0 - 0.2 6.5 0.7 - 0.5 - 0.5 - 0.9 6. Change In Net Claims on Co-operatives -0.9 2.6 -25.2 -3. 106 -11.1 1.5 3.6 2.3 1.8 1.2 1 Gross Claims 2.3 3.7 ~ 1- 2.1- 2.3 0.9 -T. F 3.5 -2.2 3.0 10 i Change in Liabilities 3.2 1.1 26.6 5.1 - 8.3 12.0 0.3 - 0.) - 0.1 .1.2 2.8 5-1 7. Change in Ilet Holdings of Corporate Securities 111.2. .8.9 .22.2 33.9 39.1 23.8 53.5 - 6.5 15.? 38.1 -LO. 3 .1.9 8. Net Saving through Life Insurance Policies 0.5 - 0.9 7.7 9.7 - 6.9 - 7.3 - 7.4 5.7 3.11 3.6 Gross Savings 2.3 3.1 2.3 3.4F 3.0 3.1 -3.1 3.5 2.3 -3.9 -- 6. 7.9 Change in Liabilities 1.8 1.5 1.11 11.1 - 6.7 10.0 lo.1 10.9 - 3.5 2.9 2.9 4.3 9. Net Saving trhough Pension Funds - - - - - - - - 0.3 29.5 68.7 6.0 64.6 Gross Saving - - - - - - - 21.8 )5.1 - 7 -IT3-.T6- Change in Liabilities - - - - - - - 22.1 15.6 15.1 88.7 48.9 10. Change in Tangible Assets 297.0 288.7 307.2 330.0 351.2 355.0 369.8 386.2 491.8 109.1 1,6.o 126.7 11. Total Net Saving of Household Sector (1 to 10) 103.7 206.2 210. 4511.6 498.6 408.1 343.6 390.1 62.5 556.7 503.9 525.8 Gross Saving V30.2 219.] 213.2 V7635 513.6 Vi12.0 373.9 T2 T- 59-28 578-.6 1T Change in Liabilities 26.5 12.9 31.8 21.9 15.0 33.9 30.0 39.2 36.4 36.2 91.7 112.2 1e TableIII: CHANGE IN ASSETS AND LIABILITIES OF THE H0OUSEHOF-D S1CTOR 1951-71, (page 2) (Rs. Million) 1963 1964 3965 1966 1907 1968 1969 190 197I 19(2 1973 1974 1. Change in Currency Holdings 118.3 63.9 60.2 - 31.3 13.9 55.8 - 45.1 -56.2 214. 270.3 230.7 7.8 2. Chunge i Net. Comercial Bank Deposits 50.3 19.0 15.8 289 - 1ii.6 4.6 5.8 219.0 -49.( -8 . 6 20.5 62.4 ig in Gross Deposits 2 -65 - 17 143.6 Change In Liabilities 18.2 46.3 20.14 9.1 79.8 34.6 76.0 42.9 115.1 61.9 129.3 81.2 3. Change in Net Savings Dank Deposits 1A.7 32.1 33.3 30.6 31.9 14.1 133. 5 99.5 111.0 109.3 213.1 203.5 Change in Gross Deposits is101. 11.7 19.3 3.1 203.5 Change in Liabilities 1.8 2.1 0.9 - 1.5 2.0 1.0 1.9 0.7 - - - 4. Change in Net Deo1ts witLong-term Credit Institutions -2.2 3.9 0.2 8.4 -31.3 4.4 -10.5 5.6 16.1 0.5 - 4.2 Chnei rss Deoits -60.3 --0.5 -0.3 3. 13.7 3.7 1~ 25.1 7.2.5 -1.U. Change in Liabilities 1.9 3.14 - 0.5 12.0 111.7 45.0 20.8 35.6 2.2 6.1 - 2.0 5.2 5. Chang ms on Government 0.3 - 0.5 2.9 13.3 - 2.6 - 1.4 -10.3 11.3 3.5 2.8 -13.7 - 3.1 Change n Gaross Claims 0.5 1.1 5.7 15T1 Change in Liabilities 0.8 1.6 2.8 2.4 5.6 14.6 6.6 1.3 0.8 3.4 9.6 3.1 6. chaaigei Claims on Co-operatives - 82 - -12.2 14.2 15.8 -25.2 11.1 -42.2 -18.4 6.o -170.8 -75-11 Chnei ross Claims 0U.5 3.5 1.7 -. - T- 9.3 7.3 -- 7.2 2~ .2~ 20.1~: Change in Liabilities 8.7 12.2 3.3.9 17.0 20.7 34.5 18.4 35.0 15.6 65.6 195.0 95.5 7. Change in Net Holdings of Corporate Securities 3.1 3.2 9.5 12.4 314.0 37.8 15.14 40.3 54.0 77.8 68.0 61.0 8. Net Saving through Life Insurance Policies 20.5 23.5 23.0 21.5 1.2 19.7 17.3 1.8 23.0 37. L 41.9 17.8 G-r-oss Savi ng -17 2.9 -21.6 2T3.1 -197~ 3 ~ ~ 25.8.91 3TEZ --3B-U --M-. 51.0 Change in Liabilities - 2.1 - 1.6 - 1.4 1.6 1.6 5.7 11.5 13.3 11.8 1.7 7.0 .3.2 9. Net Saving through Pension Funds 62.1 130.2 151.5 Ib7-1 185.7 111.3 96.9 142.2 163.9 112.6 118.2 2.6 Gross Saving 120.7 -n-Tli 19-5.l -15-3.6 261.9 -176T.0- 186.2 - 593 -3M . 2rl.7 3 12.0 713 Change in Liabilities 58.6 141.2 44.0 46.5 76.2 64.8 89.14 117.2 144.7 182.0 193.Y 234.2 10. Change in Tangible Assets 442.0 4314.3 337.14 1490.1 584.2 661.3 863.8 1017.8 1043.2 1078.5 1225.4 1100.9 11. Net Saving of the Household Sector 730.3 723.1 621.6 650.0 193.9 909.0 941.6 136.0 1580.8 1581.9 1733.8 I83 cGange Saing Lb e92s.3 70015.6 788. 23. 1 1 15.2 .2 1683.1 171.8 19).2 2266.5 2252.7 Change in Liabilities 87.9 205.2 80.1 88.6 230.1 191.2 216.7 247.2 290.9 317.2 532.6 1122.4 - 55 - Table IV: STRUCTURE OF NEZ SAVR.G IN FB-ANCTAL FORM OF THE HOUSEHO) SECTOR: ANNITAL AVERAGES 0VER -EAR FRDS. Average Annual Share in Total Net Saving in Financial Form of Periods: 1951-55 1956-60 1961-65 1966-70 1971-74 1951-74 1. Currency -13.9 7.8 39.6 5.0 36.6 25.0 2. Net Deposits Held In: 75.0 ,6.0 18.5 31.1 31.8 30.2 Coimmercial Banks 53.0 ~'T~64 Savings Banks 34.2 61.1 9.6 18.0 33.7 25,,0 long-term Credit Institutions -12.1 -1)4.2 -1.3 -7.1 0,9 --3,2 3. Net Securities of' Which: 37.9 39.3 6.0 3.4 3.9 2.1 Government Securities -13.5 ~-~5 1.7 ~~~ ~i~ ~ Cooperative Securities -8.o -0.6 -3.2 -8.9 -6.6 - Ifl.? Corporate Securities 59.3 C-O.3 -6 11.5 13.2 2É 4. Contractual Savings (net) Through: 1.·0 26.9 48.0 6o.4 35.6 h2.6 Lif e Insuran e1.0 73 -76- Pension Funds -31.8 40.7 53.0 28.0 36.2 Total Net Saving in Financial Form 199.5 308.0 1018.5 1213.5 1978.8 h717.9 (Period totals ,i us. million) - 56 - Table V: STRUCTURE OF GROSS SAVING IN FINANCIAL FORM OF THE HOUSEHOLD SECTOR: ANNUAL AVERAGES OVER 5-YEAR PERIODS Average Annual Share in Total Gross Saving in Financial Form of Period: 1951-55 1956-60 1961-65 1966-70 1971-74 1951-74 1. Currency -9.1 4.9 25.2 2.7 20.4 14.6 2. Gross Deposits Held In: 64.2 32.5 21.1 35.9 29.1 30.9 Commercial Banks 39.5 5.6 14.7 22.3 9.4 14.8 Savings Banks 24.I 26.2 6.5 10.3 18.8 11.8 Long-Term Credit Institutions 0.2 0.7 -0.1 3.3 0.8 1.3 3. Gross Savings in Securities: 40.2 28.1 -0.9 8.6 8.6 9.1 Government Securities -2.1 0.1 1.3 1.4 - 0.6 Co-operative Securities 3.8 2.4 0.7 0.8 1.2 1.2 Corporate Securities 38.5 25.7 2.9 6.4 7.4 7.3 4. Contractual Savings (Gross) through: 4.6 34.4 54.5 52.8 41.9 45.4 Life Insurance 4.6 3.3 4.8 5.3 4.9 4.9 Pension Funds - 31.1 49.8 47.4 37.0 1o.6 Total Gross Saving in Financial Form 307.5 483.8 1598.4 2186.7 3542.2 8118.6 (Period totals in Rs. million) - 57 - Table VI COMPARTSON OF SAVINGS ESTIATVES (HIs. Million) Difference in Household Rasaputram KahagAl1e Current. Study Sector Saving Household Net change Household Niet change Household Net change Absolute Relative Sector Saving in Financial Sector Saving in Financial Sector Saving in Financial (Percent) Assets Assets Assets 1951 246.1 169.4 - - 403.7 106.7 157.6 64.0 1952 16.0 27.1 - - 206.2 - 82.5 190.2 1188.8 1953 68.0 24.0 - - 210.4 - 96.8 142.4 209.4 1954 192.0 132.0 - - 454.6 124.6 262.6 136.8 1955 296.0 1111.0 - - 498.6 117.4 202.6 68.4 1956 210.0 127.2 - - 408.1 53.1 198.1 911.3 1957 117.0 78.1 - - 343.6 - 26.2 226.6 193.7 1958 238.0 175.9 - - 390.1 3.9 152.1 63.9 1959 275.0 224.7 - - 624.5 129.7 349.5 127.1 1960 - - 687.1 257.8 556.7 147.6 -130.4 - 19.0 1961 - 557.8 118.4 503.9 57.9 - 53.9 - 9.7 1962 - - 653.4 222.8 525.8 99.1 127.6 - 19.5 1963 - - 637.8 262.5 730.3 288.3 92.5 14.5 1964 - - 759.3 354.7 723.1 288.5 - 36.2 - 4.8 1965 - - 618.7 281.7 621.6 284.4 2.9 0.5 1966 - - 577.3 202.1 650.0 159.9 - 72.7 - 12.6 1967 - - 855.1 453.7 893.9 309.7 38.8 4.5 1968 - - 764.9 189.3 909.0 247.7 14.1 18.8 1969 - - 1265.9 321.1 941.6 77.8 324.3 25.6 1970 - - 1312.1 358.0 1436.0 1118.2 123.9 9.4 1971 - - 1412.2 534.0 1580.8 537.6 138.6 9.6 1972 - - 1404.2 332.9 1581.9 503.4 177.7 12.7 1973 - - - - 1733.8 508.4 - - 1974 - - - - 1830.3 429.11 Bibliography Aaron, H., "Social Security: International Comparisons," In Studies in the Economics of Income Maintenance, ed., by 0. Eckstein, Washington: Brookings Institute, 1968. Bhatt, V.V., "Estimates of Saving and Investment in the Indian Economy: 1950- 1951 to 1958-1959", Reserve Bank of India Bulletin, August 1961, pp."1200- 1213. Blumenthal, T., "Life Cycle .Analysis of Saving Behaviour in Japan", Hitotsubashi Journal of Economics, Volume 9, No. 2, 1969, pp.61-67. Cagan, P., The Effect of Pension Plans on Aggregate Savings, New York: NBER., 1965. Census of Financial Institutions, 1952,* Department of Census and Statistics, Government of Sri Lanka. ECAFE, "Saving of the Federation of Malaya, 1954-1958: A Preliminary Estimate," ECAFE Bulletin, June 1962, pp. 17-33. Friend, I., Individuals' Saving: Volume and Composition, New York: J. Wiley and Sons, 1954. and R. Jones., (ed.), Study of Consumer Expenditures, Incomes and Savings, Proceedings of the Conference on Consumption and Saving, University of Pennsylvania, 1960. Goldsmith, R.W., A Study of Saving in the United States, Volume II, Princeton University Press, New Jersey, 1955. Hooley, R.W., Saving in the Philippines 1951-1960, Institute of Economic Develop- ment and Research: University of Philippines, Quezon City, 1963. Kahagalle, S., "An Estimate of Savings and its Dete=minants in the Sri Lanka Economy, 1960-1972", Staff Studies, Central Bank of Ceylon, Volume 5, No. 2., September 1975, pp. 33-76. Karunatilake, H.N.S., Banking and Financial Institutions In Ceylon, Central Bank of Ceylon, 1968. Katona, G., Private Pensions and Individual Savings, Ann Arbor: Survey Research Center, Institute for Social Research, University of Michigan, 1964. Khazzoom, D.J., The Currency Ratio in Developing Countries, New York: F.A. Praeger, 1966 Lewis K.A., and Breen, F.F., "Empirical Issues in the Demand for Currency: A Multinational Study," Journal of Finance, Volume XXX, No. 4, September 1975, pp. 1065-1079. -2- Nizoguchi, T., Personal Savings and Consumption in Postwar Japan, Economic Research Series. No. 12, Hitosubashi University: Institute of Economic Research, Tokyo, 1970. Mukherjee, M., U. Datta, R. Choudhury, and D.S. Prasada Rao, "Economic Accounts for Developing Countries," Peview of Income and Wealth, series 21, No. 4, December 1975, pp. 411-455. Munnell, A., The Effect of Social Security on Personal Savings, Ph.D. dissertation, Harvard University, 1973. N.C.A.E.R., Saving in India 1950-1951 to 1961-1962, New Delhi, 1965. O.E.C.D., Committee for Invisible Transactions, *Capital Markets Study, Paris, 1968. Radice, E.A., Savings in Great Britain, Oxford University Press, London, 1939. Rasaputram, W., "Savings of the Ceylon Economy, 1950-1959," Central Bank of Ceylon Bulletin, January 1961, pp. 10-31. Reviglio, F., "Social Security: A Means of Savings Mobilization for Economic Development", IMF Staff Papers, 1967, pp. 324-365. Savundranayagam, T., Sources and Methods Used in the Estimation of National - Product and Expenditure, mimeo, to be published, 1974. Shinohara, M., Consumption Function, Keiso Shobo: Tokyo, 198. "The Structure of Savings and the Consumption Function in Postwar Japan" -Journal of Political Economy, Vol. LXVIII, No. 6, 1959, pp. 589-603. ANNEX: TABIES TABLE 1: HOUåEHOLD SECTOR SAVING IN THE FORM OF CURRENCY HOLDINGS (Rs. million) Change in Change in Currency Held by Change in Currency Currency Held Companies Co-operatives Local Held by Household by Govt. Sector Public (1) - (2+3+4) (1) (2) (3) (4) (5) 1951 52.0 29.5 0.6 0.1 21.8 1952i -20.8 24.9 0.6 0.2 46.5 1953 -21.3 35.0 -0.2 0.1 - 56.2 1954 6.5 -12.7 0.7 - 18.5 1955 42.7 6.8 1.4 - 0.2 34.7 1956 16.6 -64.7 1.7 0.5 79.1 1957 33.8 125.9 4.6 - - 96.7 1958 94.9 81.7 3.7 0.4 9.1 1959 35.2 4.9 0.6 0.2 ' 39.3 1960 28.3 36.0 -1.1 0.3 - 6.9 1961 98.9 8.5 3.0 0.1 87.3 1962 20.4 -21.4 -2.2 '0.2 43.8 1963 115.8 - 30.5 -2.2 0.2 148.3 1964 24.6 -41.6 2.3 - 63.9 1965 48.4 -27.9 15.6 0.5 60.2 1966 -18.9 21.8 - 9.3 -0.1 -31.3 1967 97.4 -33.0 - 7.1 0.6 136.9 1968 86.3 27.9 2.0 0.6 55.8 1969 17.7 56.0 6.3 0.5 -45.1 1970 -148.8 -80.9 -11.8 0.1 -56.2 1971. 180.2 - 28.1 - 7.4 1.0 214.7 1972 87.0 -186.6 2.4 0.9 270.3 1973 234.4 9.3 - 6.2 0.6 230.7 1974 103.6 96.8 - .2.5 1.5 7.8 Source: see text. TABLE 2; HOUSEHOLD SECTOR GROSS SAVING IN THE FOR OF COIMRCIAL BANK DEPOSITS (Rs. million) Change in Change in Change in Total Demand Savings Time (1) + (2) + (3) Deposits Deposits Deposits (1) (2) (3) (4) 1951 27.7 8.5 2.1 38.3 1952 -62.7 3.3 3.2 -56.2 1953 -31.8 3.5 -1.2 -29.5 1954 80.9 9.8 6.4 97.1 1955 54.0 18.1 -0.2 71.9 1956 -66.8 20.1 2.5 - 44.2 1957 -2.3 11.4 11.0 20.1 1958 -22.8 12.2 2.6 -8.0 1959 12.5 11.4 4.5 28.4 1960 11.3 7.1 12.3 30.7 1961 -7.2 2.6 3.3 -.1.3 1962 18.5 8.8 9.2 36.5 1963 19.7 22.9 25.9 68.5 1964 56.8 27.2 11.3 95.3 1965 12.9 31.0 - 7.7 36.2 1966 -6.4 16.3 28.1 38.0 1967 -7.4 25.9 11.7 30.2 1968 11.9 47.1 16.2 75.2 1969 21.4 52.5 7.9 81.8 1970 82.8 151.6 27.5 261.9 1971 -23.2 68.4 20.2 65.4 1972 2.0 -8.5 -18.2 -24.7 1973 31.2 132.2 -13.6 149.8 1974 41.7 99.6 2.3 143.6 Source: see text. TBa`R 3: lIOUSEHOLD SECTOR LIABILITIES WITII COMMERCIAL BANKS ( Ra. million) Outstanding Loans to lHousehold Sector: Change in Total For Commerce For Agriculture For Industry For Consumption For lousing TOTAL (1) to (5) (1) (2) (3) (4) (5) (6) (7) 1956 32.47 4.24 5.96 13.46 56.13 1957 31.64 8.43 7.27 21.86 69.20 13.07 1958 39.55 4.38 4.64 25.69 74.26 5.06 1959 42.31 4.22 5.10 32.46 84.09 9.83 1960 45.07 4.20 4.36 33.74 87.37 3.28 1961 42.09 4.36 5.42 32.73 84.60 - 2.77 1962 67.90 11.06 5.99 49.53 134.48 49.88 1963 Q5.68 17.17 3.25 36.59 152.69 18.21 1964 125.54 18.72 5.81 48.88 198.95 16.26 1965 130.01 21.58 3.10 64.69 219.38 20.43 1966 124.13 25.50 3.66 75.17 228.46 9.08 1967 126.68 80.12 4.90 96.60 308.30 79.84 1968 120.98 -119.17 7.51 95.27 342.93 34.63 1969 159.82 147.92 12.34 98.81 118.89 75.96 1970 185.62 161.55 15.53 99.07 461.77 42.88 1971 23h.62 189.39 19.27 133.55 576.83 115.06 1972 254.16 214.94 21.51 151.15 641.76 64.93 1n73 290.13 243.98 21.96 138.86 76.14 771.07 129.31 1974 343.33 235.68 19.04 143.62 110.62 852.29 81.22 Source: see text. TABLE 4i: HOUSEHOLD SECTOR NET SAVING IN THE FORM OF COMMERCIAL BANK DEPOSITS (Rs. million) Variations in Variations in Net Saving Total Deposits Out"tanding Loans (1) - (2) Held (1) (2) (3) 1951 38.3 8.5 29.8 1952 -56.2 -1.3 -54.9 1953 -29.5 2.7 -32.2 1954 97.1 4.3 92.8 1955 71.9 1.7 70.2 1956 -44.2 -1.5 -42.7 1957 20.1 13.1 7.0 1958 -8.0 5.1 -13.1 1959 28.4 9.8 18.6 1960 30.7 3.3 27.4 1961 -1.3 -2.8 1.5 1962 36.5 49.9 -13.4 1963 68.5 18.2 50.3 1964 95.3 46.3 49.0 1965 36.2 20.4 15.8 1966 38.0 9.1 28.9 1967 30.2 79.8 -49.6 1968 75.2 34.6 40.6 1969 81.8 76.0 5.8 1970 261.9 42.9 219.0 1971 65.4 115.1 - 49.7 1972 -24.7 64.9 -89.6 1973 149.8 129.3 20.5 1974 143.6 81.2 62.4 Source: see text. Table 5: HOUSEHOLD SECTOR NET SAVING WITH SAVINGS BANKS ( Rs. million) Post Office Ceylon Savings National Total Savings Savings Certifi- Savings Net Bank Bank cates Bank 1/ Saving 1951 b3.2 3.9 -0.2 - 46.9 1952 16.3 1.5 0.7 - 18.5 1953 0.2 -4.0 -0.9 - -4.7 195h -2.9 -3.0 -1.1 - -7.0 1955 13.9 1.2 -0.7 - 14.4 1956 23.5 1.h 1.8 - 26.7 1957 17.4 2.6 2.6 - 22.6 1958 18.5 0.9 0.5 - 19.9 1959 27.1 3.0 1.5 - 31.6 1960 22.5 2.4 1.0 - 25.9 1961 8.5 -0.9 0.5 - 8.1 1962 10.4 - 1.2 0.2 - 9.h 1963 13.2 - 1.5 - 1.7 1964 26.8 2.3 3.0 - 32.1 .1965 24.8 4.1 4.4 - 33.3 1966 11.3 4.6 14.7 - 30.6 1967 12.2 0.5 22.2 - 34.9 1968 22.7 5.1 12.6 - 40.4 1969 . 17.2 1.3 -5.0 - 13.5 1970 103.9 8,8 -13.2 - 99.5 1971 65.1 83.7 - 7.8 .. 111.0 1972 1/ 1/ -0.1 109.b 109.3 1973 1/ 1/ 32.8 180.3 213.1 1974 1/ 1/ 8.9 194.6 203.5 1/ The National Spvings Bank was formed in 1972 when it took over the assets and liabilities of the Post Office Savings Bank and the Ceylon Savings Bank (effective April 1, 1972). Table 6: HCUSEHOLD SECTOR NET SAVING WITH HIRE PURCHASE COMANIES AND LCGTERM4 CREDIT INSTITUTICNS (Rs. million) Hire Purchase Finance Comranies Long-Term Credit Total Change in Change in Net Institutions 1/ (3) + (4) deposits Liabilities Saving Net Saving (1) - (2) (1) (2) (3) (4) (5) 1951 0.3 0.3 - -6.4 -6.4 1952 0.1 0.3 - 0.2 - 6.3 - 6.5 1953 - 0.1 -0.1 - 4.2 -4.3 1954 0.2 0.3 -0.1 -3.4 -3.5 1955 0.1 0.3 -0.2 - 3.3 -3.5 1956 0.3 1.8 -1.5 - 3.2 -14.7 1957 0.4 0.6 -0.2 -6.6 -6.8 1958 0.9 1.9 -1.0 - 5.2 -6.2 1959 1.2 8.7 -7.5 -5.1 -12.6 1960 0.7 5.4 -14.6 - 8.9 -13.5 1961 - 0.2 -2.0 1.8 - 4.9 -3.1 1962 -0.3 -0.3 - -3.9-3.9 1963 - 0.3 -0.2 -0.1 -2.1 -2.2 1964 -0.5 2.7 -3.2 -0.7 -3.9 1965 -0.3 -0.5 0.2 - 0.2 1966 3.6 2.1 1.5 - 5.9 - 8.4 1967 13.7 25.8 -12.1 -18.9 -31.0 1968 13.7 31.7 -18.0 -13.3 -31.3 1969 16.h 13.2 3.2 - 7.6 - 4.4 197D 25.1 22.8 2.3 -12.6 -10.5 1971 7.8 -10.2 18.0 -12.4 5.6 1972 22.5 11.7 10.8 5.3 16.1 1973 -1.5 -11.2 9.7 -9.2 0.5 1974 1.0 -14.3 15.3 -19.5 -14.2 1/ A.I.C.C., National Housing Fund and State Mortgage Bank. TABLE 7: 1OUSEHOLD SECTOR NET SAVING WITH THE GOVERNMENT (Rs. million) Variations in Net Saving Household Outstanding (1) - (2) Sector Claims Loans from on the Govt. the Govt. (1) (2) (3) 1951 6.2 5.4 0.8 1952, - 2.3 0.9 -3.2 1953 - 2.8 -5.5 2.7 1954 -4.2 .-4.8 0.6 1955 -3.3 24.5 1956 -3.1 8.0 -i.1 1957 -0.2 -0.2 1958 -1.7 - 6.5 4.8 1959 0.8 0.7 0.1 1960 4.6 -0.5 5.1 1961 13.3 -0.5 13.8 1962 0.3 - 0.9 1.2 1963 0.5 0.8 1964 1.1 1.6 - 0. 1965 5.7 2.8 2.9 1966 15.7 2.4 13.3 1967 3.0 5.6 -2.6 1968 3.2 4.6 -1.4 1969 -3.7 6.6 -10.3 1970 12.6 1.3 11.3 1971 4.3 0.8 3.5 1972 -0.6 -3.4 2.8 1973 4.1 9.6 -13.7 1974 3.1 - 3.1 Table 8: FCUSEHOLD SECTOR NL' SAVING WITH CO-OPEATIVES .AN' RURAL BANKS Rs. million) Co-oneratives (excluding rural banks) Pural Banks Total Change in Change in Net Change in change in Net 3) - (6) Share Outstanding Saving deposits outstanding Saving Capital Loans 1/ (1) - (2) held loans (1) (2) ~ (3) (h) (5) (6) (7) 1951 2.3 3.2 - 0.9 - - - -0.9 1952 3.7 1.1 2.6 - - 2.6 1953 1.h 26.6 -25.2 - - - -25.2 1954 2.1 5.1 - 3.0 - - - _ 3. 1955 2.3 -8.3 10.6 - - - 10.6 1956 0.9 12.0 -11.1 - - - 11.1 1957 1.8 0.3 1.5 - - - 1958 3.5 -0.1 3.6 - - - 3.6 1959 2.2 -0.1 2.3 - - - 2.3 1960 3.0 1.2 1.8 - - - 1.8 1961 4.0 2.8 1.2 - - -2 1962 0.8 5.0 - 4.2 - 0.1 -0.1 - 3 1963 0.3 8.6 - 8.3 0.2 0.1 0.1 _ 8.2 1964 3.3 12.0 - 8.7 0.2 0.2 - - 8.7 1965 1.6 13.1 -11.5 0 .1 .0.8 -0.7 -12.2 1966 2.6 16.4 -13.8 0.2 0.6 -o.h -1.2 1967 L.2 19.8 -15.6 0.7 0.9- -0.2 -15.8 1968 7.4 32.4 - 25.0 1.9 2.V . -0.2 -25.2 1969 4.h 16.0 -11.6 2.9 2.h 0.5 _11.1 1970 -17.9) 33.0 -50.9 10.7 2.0 8.7 -42.2 1/ 1 1971 - 3.8 - 15.41 -19.2 1.0 0.2 0-8 -18 1972 - 2.8 59.7 -62.5 5.4 5.9 -0.5 -63.0 1973 2.3 171.3 -169.0 21.9 23.7 -1.8 -170.8 197h 2.6 67.0 -64.4 17.5 28.5 -1.0 -75.4 1/ Interpolated2 no data available for 1971 due to reorganization. Table 9: HOUSEHOLD SECTOR NET SAVING IN CORPORATE SECURITIES (Rs. million) Net Saving in Corporate Securities 1,91 14.2 1952 8.9 1953 22.2 1954 33.9 1955 39.1 1956 23.8 1957 53.5 1958 -6.5 1959 15.2 1960 38.1 1961 -60.3 1962 - 1.9 1963 3.1 1964 3.2 1965 9.5 1966 12.4 1967 34.0 1968 37.8 1969 15.4 1970 40.3 1971 54.0 1972 77.8 1973 68.0 1974 61.0 TABL2J HOUSEOLD SECTOR NIiI' SAVING THROUIL PROVID2NT WI-DS. Enployees Provident Fund. (Rs, i@aand - Contributions Interest Total Refunds Working Interest Total Net Saving Received Received Impayments Made Expenses Paid Outpayments of louseholds (1) + (2) (4)+(5)+(6) (3) - (7) (1) (2) (3) (4) (5) (6) (7) (8) 1958 - - - - 45 - 45 - 45 1959 12015 108 12123 87 1057 - 1144 10979 1960 55407 1212 56619 164 2574 1293 4031 52588 1961 56543 3562 60105 464 3509 3133 7106 52999 1962 57714 5983 63697 962 3716 4761 9439 51258 1963 53606 8758 62364 2012 3891 6295 12198 50166 1964 56049 11661 67710 3045 1933 7903 12881 54829 1965 58565 14771 73336 4112 4429 9617 18158 55178 1966 68073 18172 86245 5693 3964 184 20841 65404 1967 68548 21778 90326 9779 3258 13297 26334 63992 1968 79696 26111 105807 9618 3562 15475 28655 77152 1969 84016 31614 115630 20530 3957 25797 50284 65346 1970 109802 37554 176011 33508 4947 31800 70255 105756 1971 147762 48211 195973 41064 4941 42783 88743 107230 1972 177493 63659 241152 54247 5434 57064 116745 121407 1973 188862 80090 268952 7G,361 5503 72372 148236 120716 1974 208754 96995 305749 89141 5518 90583 185242 120507 SOURCE: Employees Provident Fund. TABLE 11: HOUSEHOLD SECTOR NET SAVING THROUGH FROVIDENT FUNDS: Private Provident Funds (Rs. Thousand) Contributions Interest Total Refunds Loans and Total Net Received Received Iapayments Made 1/ Advances Outpayments Saving (1)+(2) ()+(5) (3)-(6) (1) (2) (3) (4) (5) (6) (7) 1958 19857 1978 21835 21092 1046 22138 -303 1959 30644 2349 32993 9027 5437 1464 18529 1960 24685 2486 27171 7525 3576 11101 16070 1961 32735 1902 34637 9424 72198 81622 - 46985 1962 40999 8852 h9851 33237 6297 39534 10317 1963 36836 21458 58294 34857 11542 46399 11895 1964 154318 49406 203724 15657 112635 128292 75432 1965 92515 29614 122129 12010 13833 25843 96286 1966 60973 6450 67423 12818 12893 25711 41712 1967 163839 7741 171580 17802 32060 49862 121718 1968 58505 11722 70227 17749 18347 36092 34131 1969 60389 10248 70637 17213 21867 39080 31557 1970 74172 9165 83337 22822 24118 46940 36397 1971 97913 14779 112692 19214 36814 56028 56664 1972 68206 15319 83525 32625 32692 65317 18208 1973 34920 8062 42982 20186 25301 45487 - 2505 1974 48785 9275 58o6o 26169 22830 48999 9061 1/ Includes interest payments Source: Commissioner of Labour. Table 1' RI ousehold Sector Net Saving in the Form of' Life Insurance: T-nsurance Corporation of Ceylon -(Rsb thiouWand} Change in Change in total Net Li.fe Assurance loans outstanding Saving Fund (l)-(2) (1) (2) (3) 1962 770 -770 1963 4187 53 4134 1964 9151 96 9055 1965 14683 450 14233 1966 20816 1272 19544 1967 21692 1626 20o66 1968 26539 2394 2414-5 1969 30040 4670 25370 1970 40935 10858 30077 1971 46323 8354 37969 1972 51664 6337 45327 1973 6240.0 5879 56521 1974 66oOo 8357 57643 Source: Insurance Corporation of Ceylon. Table 13: Salary Loans irade by Insurance Corporation of Ceylon Loans Outstanding Change in to to Total (3) Staff Field Officers (1)+(2) (1) (2) (3) (4) 1962 23 - 23 23 1963 67 12 79 56 1964 150 120 270 191 1965 185 47 232 -38 1966 396 51 447 215 1967 653 332 985 538 1968 602 853 1455 470 1969 1476 223 1699 244 1970 54h 426 5870 4171 1971 11936 506 12442 6572 1972 12281 525 12806 364 1973 12834 567 13401 595 1974 11723 . 467 12190 1211 Source: Insurance Corporation of Ceylon Table 1-: Household Sector Net Saving in the Form of Life Insurance Private Life Insurance Ccmnanies Rs. Thousand Change in Life Outstanding Loans Change Net Insurance Fund to Policy holders in (2) Saving (1) - (3) 1951 2250 15760 1783 467 1952 3063 20262 4502 - 1439 1953 2289 21622 1360 929 195k 3356 32726 1110h -77h8 1955 3023 26065 -6661 968LL 195e6 3078 36056 9991 - 6913 1957 3133 L6U0 10384 - 7251 1958 3L,55 57321 10881 -7426 1959 2269 538h3 - 378 5747 1960 3888 56799 2936 952 1961 6277 59659 2860 3417 1962 7129 63957 1298 2831 1963 1h206 617h5 -2212 16418 196h 12755 59886 -1859 14611L 1965 6915 58028 -1858 8773 1966 2301 58115 87 2214 1967 -2970 57618 - 497 - 2293 1968 -1113 60438 2820 -3933 1?69 -8h07 60056 - 382 - 8025 1570 -12789 58330 -1726 -11063 1971 -11504 55243 -3087 - 817 1972 -12830 50263 -4980 -7850 1073 -13ih69 50870 607 -.1h070 1974 -14991 L6978 -3892 - 11099 SOURCE: Data supplied by the authorities. Table 15: Saving in the Form of Tangible Assets of the Household Sector (rs. million) Residential Machinery and Land: Planting Construction Equipment and Replanting Gold Other- Total (5) (1) (2) (3) (4),- (4) (1) to (4) 1951 157.5 24.7 37.8 2.0 75.0 297.0 1952 152.5 23.9 36.6 3.1 72.6 288.7 1953 163.7 25.6 39.3 0.5 78.1 307.2 1954 175.3 27.5 42.1 1.6 83.5 330.0 1955 186.5 29.2 44.8 1.9 88.8 351.2 1956 187.4 29.4 45.0 3.9 89.3 355.0 1957 195.2 30.6 • 46.9 4.2 92.9 369.8 3,958 205.5 32.2. 49.4 1.7 97.9 386.7 1959 262.5 41.1 63.1 3.0 125.1 494.8 1960 216.7 33.9 52.1 3.2 103.2 409.1 1961 236.3 37.0 56.8 3.5 112.5 446.0 1962 225.9 35.4 54.3 3.5 107.6 426.7 1963 233.8 36.6 56.2 3.9 111.5 442.0 1964 230.5 36.1 55.4 2.4 109.9 .434.3 1965 178.2 27.9 42.8 3.5 85.. 337.4 1966 260.2 40.8 62.5 2.6 124.o 490.1 1967 309.8 48.5 74.5 3.8 147.6 584.2 1968 347.8 55.0 84.4 3.8 170.3 661.3 1969 458.1 71.7 110.1 5.6 218.3 863.8 1970 540.4 84.6 129.8 5.5 257.5 1017.8 1971 556.6 87.2 133.8 0.5 265.1 1043.2 1972 574.0 89.9 138.0 3.1 273.5 1078.- 1973 650.9 101.9 156.4 6.1 310.1 1225.4 1974 743.5 116.4 178.7 8.0 354.3 1k00.9 1/ Consists mostly of vehiLcles and inventories. Source: See text. Table 16: HOUSEHOLD SECTOR NET SAVIN (1951-711) Page 1 1951 1952 1953 1951 1955 1956 1957 1958 1959 1960 19 1 1962 A. Financial Assets 106.7 -82.5 -96.8 124.6 1117.4 53.1 -26.2 3.9 129.7 147.6 57.9 99.1 1. Currency 21.8 -46.5 -56.2 18.5 34.7 79.1 - 96.7 9.1 39.3 - 6.9 87.3 43. 2. Net Commercial Bank Deposits 29.8 -54.9 -32.2 92.8 70.2 - 42.7 7.0 -13.1 1.6 27.4 1.5 -13.11 3. Net Savings Bank Deposits 46.9 18.5 - 4.7 - 7.0 14.4 26.7 22.6 19.9 31.6 25.9 8.1 9.11 4. Net Saving with Long-term Credit Institutions 1/ - 6.4 6.5 - 4.3 - 3.5 - 3.5 - 4.7 - 6.8 - 6.2 -12.6 -13.5 - 3.1 - 3.9 5. Net Claims on Government 0.8 3.2 2.7 0.6 -27.8 - 11.1 - 11.8 0.1 5.1 13.8 1.2 6. Net Claims on Co-operatives 2/ - 0.9 2.6 -25.2 - 3.0 10.6 -11.1 1.5 3.6 2.3 1.1 1.2 - 1.3 7. Net Corporate Securities 14.2 8.9 22.2 33.9 39.1 23.8 53.5 - 6.5 15.2 38.1 -60.3 - 1.9 8. Net Saving through Life 0.5 - 1.4 0.9 - 7.7 9.7 - 6.9 - 7.3 - 7.4 5.7 1.0 3.4 3.6 Insurance 3/ 9. Net Saving Through Pension - - - - - - - - 0.3 29.5 68.7 6.0 61.6 Funds 4/ B. Tangible Assets 297.0 288.7 307.2 330.0 351.2 355.0 369.8 386.7 491.8 o09.1 1b6.0 1126.7 Saving of the Household Sector hQ-.7 2QL2 a 451.6 198.6 03.1 343.6 390.6 624.5 556.7 503.9 525.8 (A + B) 1/ National Housing Fund, State Mortgage Bank, Agricultural and Tndustrial Credit Corporation, and Hire Purchase Finance Cru-rations. 2/ Including rural banks. 3/ Both private life i-surance companies and the Insurance Corporation of Ceylon (E.C.C.). Prior to 1958, no data are available on saving through the private life insurance companies, and the I.C.C. was n3t then in operation. 11 Both private pension funds and the Employees Provident Fund. Taole 16 :HOUSEHOLD SECTOR NET SAVNG (1951 - 7) Page 2 1963 1964 1965 1966 1967 196 1969 1970 1971 1972 1973 1974 A. Financial Assets 288.3 288.8 284.4 159.9 309.7 247.7 77.8 418.2 537.6 503.-1 509.4 129.4 1. Currency 148.3 63.9 60.2 -31.3 136.9 55.8 -45.1 - 56.2 214.7 270.3 230.7 7.8 2. Net Commercial Bank Deposits 50.3 49.0 15.8 28.9 -49.6 40.6 5.8 219.0 -49.7 -89.6 20.5 62.4 3. Net Savings Bank Deposits 14.7 32.1 33.3 30.6 3.9 40.4 13.5 99.5 141.0 109.3 213.1 203.5 4. Net Saving with Long-term Credit Institutions 1/ - 2.2 - 3.9 0.2 - 8.4 -31.0 -31.3 - 4.4 -10.5 5.6 16.1 0.5 - 4.2 5. Net Claims on Government - 0.3 - 0.5 2.9 13.3 - 2.6 - 1.4 -10.3 11.3 3.5 2.8 -13.7 - 3. 6. Net Claims on Co-operatives 2/ - 8.2 -8.7 -12.2 -14.2 -15.8 - 25.2 -11.1 -42.2 -18,4 -63.0 -V70.8 -75.4 7. Net Corporate Securities 3.1 3.2 9.5 12.4 34.0 37.8 15.4 40.3 54.0 77.8 68.0 61.0 8. Net Saving through Life Insurance 3/ 20.5 23.5 23.0 21.5 17.2 19.7 17.1 14.9 23.0 37.1 41.9 47.8 9. Net Saving through Pension iands 4/ 62.1 130.2 151.2 107.1 185.7 111.3 96.9 142.2 163.9 142.6 118.2 129.6 B. Tangible Assets 442.0 434.3 337.4 490.1 584.2 661.3 863.8 1017.8 1043.2 1078.5 1225.4 1400.9 Saving of the Household Sector 730.3 723.1 621.6 650.o 893.9 909.0 941.6 1436.0 1580.0 1581.9 1733.q 1830.3 (A + B) 1/ National Housing Fund State Mortgage Bank, Agricultural and Industrial Credit Cornorati)n, and Hire Purchase Finance Cmpanies. 2' Including rural banks. 3' Both private life insurance companies and the Insurance Crporation- of Ceylon (I.C.C.). Prir to 1959 no data are available )n saving through the private life insurance companies and the I.C.C. was n)t then in 1perati'n. 4' Both private pension funds and the Emoloyees Provident Fund.