SWP463 The views and interpretations i n t h i s document are those of the author and should not be attributed t o the World Bank, t o its affiliated organiza- tions, or t o any individual acting kn their behalf. WORIl) B r n Staff Working Paper No- 463 June 1981 WRAT IS A SAM? A LAPMAN'S GUIDE TO SOCIAL ACCOUNTING MATRICES This paper is intended as an introduction for readers who have no previous familiarity with social accounting matrices. It assumes nothing more from the reader than a rudimentary knowledge of national income accounting. Starting ftom the simplest formlation, the paper introduces step by step many of the main features of a social accounting matrix i n its fully developed and more complex form. An explanation is given of the relationship between social accounting matrices on the one hand and national accounts, input-output matrices and multipliers on the other. Prepared by: Benjamin 9. King Development Economics Department Development Policy Staff Copyright @ 1981 The World Bank 1818 H. Street, N.W. Washington, D.C. 20433, U.S.A. PREFACE This paper has its origin i n a conference held i n Cambridge, England i n 1978, which was sponsored by the World Bank External Research Program under research project 671-27. The author was one of the two participants a t that conference who did not belong to the "subculture" of practitioners and connoisseurs of SAMs or social accounting matrices. Graham Pyatt, who is responsible f o r the research project and who organized the conference, subsequently persuaded the author t o write a guide t o SAMs f c r the layman. The author's principal qualification was that he shared with his intended audience an almost complete ignorance of the SAM framework; without that ignorance he would never had. had the temerity t o undertake such a task. Once it was launched, Graham Pyatt commented on each successive draft and edited extensively with good reason. Many useful comments were also made by Clive B e l l , Alan Gelb, Sherman Robinson, Alan Roe and Jeffrey Round. The f i n a l product, as usual, remains the author's reluctant responsibility. Table of Contents Page No Introduction..................................................... 1 A Primitive Example: the Robinson Crusoe Economy ................ 3 Sri Lanka 1970................................................ 6 Diagonal Entries ............................................ 11 Indirect Taxes............................................... 12 Savings and Investment....................................... 15 A General Framework.......................................... 18 An Input-Output Matrix ....................................... 27 Botswana 1974175: The Flow of Funds ............................. 31 Purchase and Sale of Assets .................................. 34 Decomposition of the Capital Account......................... 36 ................................ Decouiposing the Financial Transactions....................... 39 Decomposition of Institutions 41 ................................................. The Uses of SAMs 44 Learning by doing................................^.. 54 References....................................................... 59 List of Tables .Countrv . Table Page No .................................. 1 Robinson Crusoecs Island 4 ................... 2 Sri Lanka An I n i t i a l Aggregate SAM 7 .............. 3 Sri Lanka A Revised Aggregate SAM (including diagonal entries 13 ......... Sri Lanka Further Development of the Aggregate SAM (introducing indirect taxes) 14 ...................... Sri Lanka AFinal Aggregate SAM 16 .............. Sri tanka A Partially Disaggregated SAM 19 ....... Sri Lanka Consolidated Accounts for the Nation 21/22 ....................... Sri Lanka Reconciliatton Table 24 ....................... Sri Lanka Reconciliation Table 26 ................. Sri tanka Complete Disaggregated SAM 28 ........................ Sri Lanka Input-Output Matrix 30 ........................... Botswana An Aggregate SAM 33 ................... Rotswana Introduction of a Ftnancial Account into the Aggregate SAM 35 .............................. ... Botswana A SAM with Disaggregated Capital Accounts 37 ............ Botswana Extending the SAM t o Include Separate Accounts for Financial Assets 40 ................................ Botswana A Final SAY 42 .......................... Botswana Multiplier Effect i n the Form of an Increnental SAX 48 ............................... Yemen Arab A Simple SP;M 55 Republic Introduction 1. A social accounting matrix, familiarly known a s SAM t o the limited fraternity familiar with it, has two principal objectives. The f i r s t is concerned with the organization of in£ormation, usually in£ormation about the economic and social structure of a country i n a particular year, though it could as well be about a region in a country, a city or any other unit which one is interested i n and the unit of time, though convenient, is arbitrary. Complaints about the inconsistency and unreliability of economic and social data i n developing countries have reached the point of being trite. While there is justification i n these complaints, they are not the whole story. There is of ten idormation, dispersed ,r fragmentary, which is not used for lack of a framework t o make the maximum use of the available information and t o pinpoint with greater accuracy and specificity than is usually done the salient gaps and inconsistencies. 2. Once the data i n a particular country f o r a particular year have been organized i n the form of a SAM, they present a s t a t i c image which can reveal much about economic structure. Even so, the image is only a "snapshot." I n order t o analyze Row the economy works and predict the effects of policy intervenrions i n order t o change it, more is needed thnn just a s t a t t c image. A model of the economy has t o be created which can simulate the effects of interventions, f a r example. This is the second I objective of a SAM: t o provide the s t a t i s t i c a l basis for creation of a plausible model. 3. The principle of a SAM is really nothing more than that of double entry bookkeeping. It is a series of accounts in each of which incomi~gs and outgoing. for income and expenditure i n many cases) must balance. Furthermore, what is "incoming" into one accouct must be "outgoing" from another account. I n t h i s respect, a SAM resembles traditional national accounts. In fact, as w i l l be demonstrated later, a SAM embodies the information normally included i n national accounts and nuch more. I n a SAM the double entries are achieved by only a single entry .in a matrix which resembles an oversized chessboard. Each account consists of one row across the board and one column down it, identically numbered. We shall explain how this works shortly. 4. How large the matrix is depends on the limitations of the available data and the motivation one has for constructing it. In principle, there is no limit t o the fineness of detail. In practice, both the data and the effort available f o r constructing the SAM impose limita- tions. One of the original motivations for the elaboration of SAMs has been the growing interest i n issues of poverty and basic needs. If one wishes t o show how different actfvities affect or are affected by different socioeconomic groups i n society, the amount of detail must correspond t o the differentiation one wishes t o make. 5. This paper does not attempt t o examine the problems arising i n the construction of SAMs nor the methods of economic analysis which use the assembled data. In other words, it is not addressed t o specialists but t o the broader audience of those who need an introduction t o SAM'S. I n the rest of the paper we shall proceed with simplified examples of SAMs, based on more elaborate ones published elsewhere. W eshall s t a r t a t the very simplest level with a purely imaginary economy and proceed by increasing size and complexity with examples of SAM'S worked out for real economies, Sri Lanka and Botswana. The reason for using two different vehicles, rather than carrying the reader non-stop on one, is that they illustrate the fact that SAM'S may be constructed i n different gays--or, more properly, with different accents--for different purposes. Some points of interest FAY occur i n one and others i n a.r.=ther. A Primitive Examle: the Robinson Crusoe Economy 6 As an expository device, the Robinson Crusoe economy has perhaps become rather shopworn. Nevertheless, for want of something better, we shall use it again as a point of entry into the description of SAW+. We shall assume that Robinson engages i n only one production activity, the picking of coconuts. In some given period, he picks 1,000 coconuts. This represents a t the sane t i m e the level of production, the level of his income and the level of his demand for products (sometimes called "wants"). A l l three are equal, as they must be i n such an economy. 7. The structure of this economy is set out i n Table 1, i n which two columns and rows have been l e f t blank because they are not being used for the time being. The final coluam and row around the border shav the total of each row or column. Within the border there are three entries, each of them equal t o 1,000. These constitute the SAM for Robinson Crusoe- 8. In a SAM the rows represent incomings and the colucms outgoings. For example, R w 1 which is labelled "Income," receives 1,000 from Column 4 labelled "Production." In other words, Robinsones incosee derived from production and equals 1,000. Now, turning t o Columa 1, we see the corresponding entry 1,000 which represents the outgoings of incotne in Row 2, which is labelled "Demand." Column/Rw 1, i n effect, describe Crusoe's i . role as income-earner 9. In Column/Row 2, we look a t him as consumer. H i s demand arising from income i n Row 2 is balanced by what he spends on production i n Column 2 (Row 4). The third leg of this process is in Row 4 and Column 4, where he turns up a third time as producer: his demand f o r production is 1,000 and income arising from production is also 1,000. These various identities could be s e t out i n the form of double-entry accounts, but, although not so apparent yet, the matrix is more economical since it requires one entry f o r each item, whereas conventional accour?ts require two. 10. It w i l l be noted that there is a circular process. If one puts the three entries i n the table i n coordinate form with the row f i r s t and column second, they would appear l i k e this: (1, 4 ) ;(4, 2); and (2, 1). Thus, the matrix illustrates the circular process of demand leading t o production leading t o income, which i n turn leads back t o demand. 11. Of course, this rather complicated way of setting out the t r i v i a l structure of the Robinson Crusoe economy might well be considered much ado about nothing. We present it this way, however, because it is so self- evident and can serve as an introduction t o the more complex relationships i n an actual economy. In real l i f e Robinsoe Crusoe as a member of a society may, indeed, f i l l a l l three roles, as income earner, consumer and producer, but he would do so as a member of different sorcs of units or subdivisions, according t o his function. In the accounts for a whole society, income may be subdivided into many different categories, among which income t o labor and income to capital are only the f i r s t tier. That income accrues t o a variety of domestic institutions, which are the source of demand: households with different characteristics; firms; and government (central or local). The outgoings or expenditure of these institutions is spread over a variety of products, as indeed Robinson Crusoe's must have been; production thus can be divided into as many sectors or subsectors as is desirable or practical. In the next section, we turn to a description of a SAM which has been worked out f o r an actual economy. Although progressive complexities are brought in, the fundamentals remain the same. Sri Lanka 1970 12. Among tho f i r s t SAMs constructed was that for Sri Lanka. Sri Lanka is a cwuntry with a law average income per head, but an unusually equitable distribution of income and high standard i n meeting "basic needs." Raving successfully achieved a high quality of l i f e for such a .low income, Sri Lanka's need f o r more rapid growth of income and reduction of unemployment implied structural change and it also implied better understanding of the existing economic structure. 13. A l l the data used i n this section are taken or adapted from one of the earliest publications documenting construction of a SAM.^/ We s t a r t with a highly aggregated and simplified version, shown i n Table 2. In t h i s table a l l s i x pairs of row and column are used, although one of them (3) has no entry i n the colum. Apart from the border totals, there are now 11 entries. Two of the t i t l e s used i n the Xobinson Crusoe economy were not the conventional ones used i n SAM'S, because i n such a primitive economy they would have sounded rather fatuous. Column/Row 1 is now called "Factors of Production" and ~olumn/Row 2 "Institutions." A word of explanation is i n order about these two fundamental ingredients of a SAM. Factors of productj.on consist primarily of the labor and capital which are -11 Pyatt, G., and Thorbecke, E. (1976). Planning Techniques for a Better Future, ILO, Geneva. used i n the process of production and receive income from it. Rut the production process draws these factcrs from where it can, without being overly concerned with the entities t o which their owners belong. It is these entities which constitute the "institutions" i n Colum/~ow 2. Foremast among them are households, which we may wish t o study i n different categories. Households may supply labor and capital through one or more of their members, Sut act as n unit when it comes Eo spending the income from it. Other institutions are firms or corporations, both public and private, which provide capital. A third type of institution is government, central or local.- It, too, may provide capital, but it has another important role i n the production process--at least as f a r as the determination of the value of prod~ctionis concerned--namely, the levying of indirect taxes. 14. The two new accounts 3 and 5 (i-e., the pairs of row and column) are there for different reasons. The account for the Rest of the World (5) is necessary because Sri Lanka, though an island, is not an tsslztcd one l i k e Robinson Crusoe8s but has many transactions with the Rest of the World. The "Surplus or Deficit" (3) is a direct consequence of this; the current transactions are normally not equal and must be balanced by borrow- ing or lending or use of reserves. Haw these two new qccounts f i t i n w i l l be explained i n a tour of the matrix. 15. We s h a l l s t a r t with the Production account (4). In the row are the "incomings": the proceeds of sales, a t producers 8 prices, of 11,312 million rupees t o institutions within Sri Lanka i n Column 2 and of exports of 2,113 t o the Rest of the World in colunrn 5. These are exactly balanced -1/ It w i l l be noted that these are a l l domestic institutions. by the "outgoings" or cost of production i n Column 4: i n Row 5 payments of 1,067 for imports of materials going into production; i n Row 2 payments of 885 made t o institutions during the production process, which are i n fact indirect taxes on intermediate goods or imported materials; and finally, the value added by factors of production of 11,473 i n Row 1.- 11 16. I f we move t o the income account (I), the 11,473 just referred t o is now interpreted as income t o factors of production (again not differentiated as yet). In Column 1we see rhe disposition of this income t o factors of production, consisting mainly of income of 11,360 accruing t o institutions i n Sri Lanka and the small balance of 113 t o the Rest of the World. 17. Moving now t o the third main account, for institutions, we have i n Row 2 two incoming items already mentioned: income from,factors of produc- tion (11,360) i n Column 1 and indirect taxes (885) i n Column 4 . There is one additional receipt--trans£ ers from the Rest of the World (97) i n Column 5. The components of this w i l l be described later. 18. Column 2 shows hcw the receipts of institutions are spent. I n Rows 4 and 5 there are goods and services purchased: 11,312 being the value of goods produced i n Sri Lanka, and 1,455, the value of goods imported in final form as opposed t o those used as materials i n the course of produc- tion. In addition, 3.n Row 3, the deficit of -425 is due t o the fact that expenditure by institutions i n Sri Lanka exceeds their receipts. 19. In the course of this brief tour, we have now referred t o every item i n the two new accounts, 3 and 5, with one exception., This is the -11 A l l figures are i n millions of rupees. surplus of 425 in Row 3 of the account of the Rest of the World (5). The latter's source of income in the row exceed its expenditure in the column by this amount; the Rest of the World's surplus exactly balances Sri Lanka's deficit. Consequently, Sri Lanka had to borrow or use its reserves to cover its deficit of 425 with the Rest of the World. 20. It is worth pointing out that the three largest entries in the matrix, by a wide margin, are in boxes (2, 1); (1, 4); and (4, 21, which were the only entries in the Robinson Crusoe SAM. The core of the matrix is still the circular process of demand (2), production (4) and income (1). Most of the remainder are required by the existence of the Rest of the World. It is worth, at this juncture, noting the duality of this circurar process. One can move round a circle clockwise or anticlock- wise. If we go in the anticlockwise direction, corresponding to the order just given, we are implicitly followng the flow of money. In (1, 4) factors of production (e.g., labor) receive money from production; in (2, 1) institutions (egg., households) from factors of production; and in (4, 2) production from institutions (e.g., households). We may think of the other dtrection as the supply of goods or services: factors of production to the production process; production to institutions; and institutions (by a slight stretch of the meaning) to factors of production. 21. We can now give an exauple of how the SAM could be set out in the form of double entry accounts, of which there would be five. For example, the production account: Revenue (Row) Expenditure(Column) Domestic sales 11,312 Payments to factors of production 11,473 Exports 2,113 Indirect taxes (on production) 885 Imported materials 1.067 Total 13,425 Total 13,425 22. This is more or less the form of traditional national accounts. The complete accounts are fully articulated i n this double entry form i n the sense that each item i n one account apvears on the opposite side of another account. It can be readily appreciated that, as the number of accounts is multiplied, their interconnections can become rather hard t o follow. A SAM is more economical i n that it has only half as many entries as a series of double-entry accounts, and the interconnections between accounts are obvious. Diagonal Entries 23- There are two modifications t o Table 2 i n the next version--Table 3. Both are diagonal entries: one a t (2, 2) and the other at (4, 4). Nothing else i n the table has been changed except, of course, the totals for ~ow/~olumns 2 and 4, which are each increased by the amount of the corresponding diagonal entries. 24. What meaning is attached t o a diagonal entry, which appears both as a revenue and as an expenditure to a particular account? It clearly can only mean that, i n one case, institutions make certain payments t o themselves or, i n the other case, production units do the same. If we were only concerned with institutions or production units as a group, this would not serve much purpose. But that is not the case. In fact, we shall i n due course s p l i t both accounts into several subaccounts. The diagonal elements represent the t o t a l of transactions between these subaccounts plus new diagonal elements within the subaccounts themselves, if they, too, are aggregations. 25 It should be noted that, by including diagonal entries, we have changed the meaning of the totals. In Table 2, the total of Column 4 was -Table 3 - A Revised Aggregate SAM for Sri Lanka in 1970 (million rupees) 2. Institutionsa/c 3. Surplus or Deficit 5. Rest of the World a/c 6. Total the t o t a l value of a l l goods and services (value added i n Sri Lanka plus import content) produced i n Sri Lanka without duplication. In Table 3, it represents the sum of the outputs of a l l production units. There is duplication i n t h i s sum t o the extent that one production unit s e l l s t o another; the diagonal entry is the amount of this duplication. Similarly, the diagonal amount of 2,441 i n Row/Column 2 represents monies paid from one institution t o another. Such transfers are a part of total incomes and t o t a l expenditures. Indirect Taxes 26. I n Table 4 we have made an i n i t i a l s p l i t i n the institutional account (i). Instead of one line, there are now two: 2abc (foreshadowing further s p l i t s into 2a, 2b acd Zc), which we have rather lamely given the same t i t l e as before; and 2d, which is labelled indirect taxes. Part or a l l of each of the entries i n the old l i n e 2 have been extracted and entered on line 2d. Thus i n Column Zabc, the original figure of 2,441 has been s p l i t into two entries, 2,052 and 389; i n Column 4 the old entry of 885 is now on line 2d; and the old entry i n Column 5 has similarly been s p l i t up. The amounts so extracted from Row 2 have been replaced by a single figure equal t o the sum of the parts extracted- 1,368 i n Column 2d. Consequently, the tocal for Row (and Column) 2abc is the same a s before. 27. Why do we need a separate account for indirect taxes? The main reason is that they should be distinguished from direct taxes. The l a t t e r are extracted from the income stream and, therefore, constitute a transfer t o the Government from other institutions. Indirect taxes, on the other hand, are levied on expenditure by the f i n a l purchaser of goods and services or, earlier, on intermediate goods purchased by producers. In Column 2 of Table 3 the total cost of goods purchased at home(11,312) and abroad (1,455) is p& what the buyer pays. Indirect taxes (or their opposite, subsidies) must be added in (or subtracted). The same is true for exports in Column 5. The symmetry with Column 4, where indirect taxes on intermediate goods or materials are clearly separated, should be evident. 28. At the same time indirect taxes which appear as an expenditure in Columns 2, 4 and 5 are also a source of revenue to the government (one of the institutions of Sri Lanka) and so appear in Column 2d on the appropriate Row (2abc). 29. Of course, indirect taxes are not an "institution" in any reasonable sense. But separating them out has sufficient advantage in understanding the structure to justify a separate subaccount. The logic of the matrix is in no way altered. The total still appears as part of the revenue of the government, while the parts are allocated to the relevant types of expenditure. This is an illustration of the flexibility of a .-1-/ SAM Savings and Investment 30. Before proceeding to a subdivision of the various accounts, we will make one more change- In Column 2 (or 2abc) expenditure so far has included all kinds of goods and services, whether for consumption or investment. The change introduced in Table 5 in effeet separates consump- -I/ Another way of dealing with indirect taxes, which may be preferable in some contexts, is to insert all of them into the production process (Column 4). tion and investment goods. Column 2abc now becomes a true current account for institutions, and has been so labelled. 31 The initial change is a reallocation between Columns 2abc and 3. We extract the investment goods in Rows 4 and 5 from Column 2abc and enter them in Column (3), renamed "combined capital account." For example, in line 4, the entries in Columns 2abc and 2d--9,350 (consumption goods) and 1,962 (investment goods)--correspond to the total 11,312 in line 4 of Table 4. The related indirect taxes similarly shift on line 2d. There is also a small shift on line 2abc for institutional reasons which are specific to Sri Lanka and do not merit detailed explanation here.-I/ 32. The total of all these changes in Column 3 is equal to expenditure on investment--namely, 2,639. Expenditure in Column 2abc is now reduced by this amount. The balance of -425, which formerly appeared on line 3, is - altered accordingly by the amount spent on investment to -425 + 2,639 = 2,214. This is now the difference between income and current expenditure or consumpticr, rather than income and total expenditure. It constitutes the savings of demestic institutions. Total savings derive from domestic sources (2,214) and foreign (425). These together finance investments (account number 3). In double-entry form, the combined capital account is: Revenue (Row) Expenditure (Column) Domestic savings 2,214 Domestic investment goods 1,962 Foreign savings 425 -- Foreign investment goods 364 --- Indirect taxes 270 Other payments 43 Total 2,639 Total 2,639 -1/ They have to do with the system of Foreign Exchange Entitlement Certificates (FEECs), a system granting or charging a premium for foreign exchange, here applied to debt payments. A General Framework 33. I n Table 6 we have divided two accounts (1 and 2) into subaccounts. The account for factors of production (1) is divided according t o the two main factors: labor (la) and capital (lb). In account number 2 the three main s e t s of institutions have been distinguished: households (2a); corporations (2b); and government ( 2 ~ ) . Indirect taxes (2d) have already been separated i a Table 4. 34. There is l i t t l e t o say about the subaccounts for factors of production, which show income t o labor and capital separately. Income t o capital is more than half the total, but, as w i l l be seen from Column lb, the greater part of it accrues directly to households. This consists largely of income from small enterprises including farms, plus a substan- t i a l element of impute& income from housing. 35. The entries i n the "box" bounded by Rows and Columns 2a, 2b and 2c are Zn the aggregate equal t o ihe diagonal entry (2, 2) i n Table 5. Payments between institutional sectors a l l f a l l into the category of transfers, since they do not constitute income directly received from production. They arise from a variety of causes: ownership of certain assets (e-g., debt or equity investment i n firms); direct taxation or government subsidies t o households (not related t o goods or services); or even voluntary transfers. Payments t o the government include, of course, direct taxes by both households and corporations, but they also include social security contributions, pension fund contributions (actual or imputed) and dividends and the l i k e from corporations (mainly public ones i n Sri Lanka). Payments by the government include pensions, interest on the public debt and a substantial amount of direct transfers t o public corporations. The large payments from corporations t o households a t (Za, 2b) are mainly payments t o owners of capital (debt or equitlr), although they include some private corporate pensions. 36 It w i l l be noted in this table that, on line 4, there is no entry for consumption i n Column 2b (corporations). Following standard national accounting practice, only households and the government consume. Goods and services used by f i r m as inputs i n the course of production are included i n the f i n a l value of output; they are only consumed when the f i ~ e zztput l is consumed. Firms may invest, but their investments are included i n Column 3. 37. Some observations are i n order about some of the other accounts. First, we can see why account 3 is now labelled combined capital account. It combines the capital. accounts of households, corporations, government and the r e s t of the world. In principle there is no reason why each institution should not have a capital account of its own. I n practice, this was not possible i n t h i s case because of data limitations. Later, we w i l l show an example of a disaggregated capital account i n another SAM. 38. The production account (4) has not yet been subdivided. The blank lines suggest, correctly, that it w i l l be i n the following table. The princi-pal reason for not dividing it a t this stage is that we do not wish t o make too many changes a t once. There is also a second reason. The SAM i n its present form gives a l l the information necessary to compile the "Consolidated Accounts for the Nation'' as outlined under the UNSNA systemLJ The accounts i n t h i s form are as s e t out in Table 7. They do -1/ United Nations Statistical Office (1968), A System of National Accounts, Series F, No. 2, Rev. 3, United Nations, New York, Table 2.14, page 29. Table 7 Consolidated Accounts.for the Nation: Sri Lanka, 1970 (a) Domestic Product and Expenditure Incomings Government consumption 1,649 Income to labor Private consumption 8,911 Income to -capital Investment 2,596 Indirect taxes Exports 2,207 - Less Imports -2,522 12,841 (b) National Disposal Income and @itlay Incomings Income to labor 5,569 Government consumption 1,649 Income to capital 5,904 Prtvate, c~nsuqtion 8,911 Indirect taxes 1,368 Savings 2,214 Current transfers from abroad 3 Transfers from capital a/c 43 - Less property income transferred abroad -113 12,774 Table 7 (cont.) Consolidated Accounts for the Nation: Sri Lanka, 1970 (c) Capital Account Incomings Outgoings Savings 2,214 Capital formation - Less transfers from capital a/c -43 Foreign borrowing 425 2,596 (d) Rest of the World Outgoings Imports 2,522 Exports prope;ty income Transfers transferred abroad 113 Surplus (of ROW) not contain any information which is not already included i i ~ Tablo, 6. The l a t t e r is therefore a more compact form i n which to present these consolidated accounts. It is also perhaps more useful insofar as the interconnections between accounts are more obvious i n Table 6 than they are i n the double-entry format. 39. Each of the Consolidated Accounts can be identified with either a single column/row i n Table 6 or a combination of them. The two accounts which correspond t o a single Column/Row can easily be compared. The Capital account (c) corresponds t o Column/Row 3; the Rest of the World account (d) corresponds t o ~olumn/Row 5. Outgoings i n the Consolidated Accounts correspond to columns and the incomings to rows. The other two accounts are less easy t o identify, since they correspond t o combina- tions. The reader not interested i n the details of the identification which follows i n the next three paragraphs is advised t o skip t o paragraph 43. 40. The Domestic Product and Expenc',i'cuce account (a) corresponds t o SAM accounts 2d and 4 combined with two adjustments. In order t o s h w the correspondence, we have reproduced Table 6 with only the entries for Column/Rows 2d and 4 i n Table 8. The two w i l l be added together i n order t o make a combined rowfcolumn, but f i r s t W-I?have t o make the two adjust- ments reEerred to. These are: ( i ) The elements common t o both row combination and column combination have been omitted; these are the figures 885 and 4660 i n Column 4 of Table 6. Since they would together form a diagonal element i n the combined column/ row, they would appear on both sides of the account and so can be . if t out. ( i i ) Secondly, we have added an extra row and column, 7, whlch does not appear as such i n Table 6. The row corresponds t o direct imports for f i n a l consumption and for invest- ment (the figures 1091 and 364 appear i n Row 5 i n Table 6); the column has the t o t a l 1455 required t o balance the account. 41. Row 8 is the sum of the eonbined entries i n Rows 2d, 4 an4 7. This row corresponds to the incomings o r left hand side of Consolidated Account (a) except for the imports, which have, i n effect, been transferred, changing the sign, from the opposite side. Column 8 includes these imports and the outgoings or right hand side of account (a). 42. The National Disposable Income and Outlay (b) -corresponds t o SAM accounts Za, 2b and 2c. Table 6 has been reproduced with only these rows and columns i n Table 9, but eliminating the diagonal transfer elements between them. Here we represent the t o t a l s by two rows and columns 8a and 8b, instead of one, because, t n several cases, the simple addition of the elements i n a particular combination does not correspond t o an entry i n Consolidated Account (c). For example, income t o capital i n Colum l b appears not as the sum of 4216 and 1535 (5791) but a s the difference between t o t a l capital income (5904) and capital income to the Rest of the World (113); that the two are the same is obvious from Column l b i n Table 6 . In Column 8b, the three elements of private consumption marked by an asterisk appear combined i n Consolidated Account (b): indirect taxes (119); domestic production (7701); and imports (1091) making a total of 8911. Once one has performed these arithmetic tricks, the elements i n the row again correspond t o the Consolidated Account incomings and the column t o outgoings. An Input-Output Matrix 43. The f i n a l step i n t h i s exposition of the S r i Lanka table is t o subdivide the production account (4) into s i x subsectors, which are listed i n Table 10 i n Rows 4a t o 4f. The "box" formed by these rows and their equivalent columns constitutes the core of what is commonly known as an input-output matrix. The accounts of each subsector follow essentially the same logic a s before. 44. For example, i f we take the agricultural subsector (4a), the column i n t o t a l is equal t o gross output at produceres prices (excluding sales taxes and the like). This t o t a l of 5,903 includes not only value added (in Rows l a and lb), taxes on intermediates (in Row 2d) and imported materials (in Row 5), but also inputs from other subsectors. The largest volume of inputs is a t the diagonal (4a, 4a) and therefore consists of internal purchases and sales within the sector; i n fact, it is largely paddy (one subsector) sold t o ricemills (another subsector). 45. Just as Column 4a shows the sources of gross output t o which payments are made, Row 4a shows the destinations of gross output from which the production units derive their revenue: i n Column 2a the amounts going t o consumption; i n Column 3 t o investment; i n Columns 4a t o 4P a s intermediate good^ t o other subsectors; and i n Column 5 to exports. 46. Table 10, with 85 entries i n it, is a long way from the pristine simplicity of the Robinson Crusoe economy. The circularity of the process of production-income-expenditure is no longer so evident. But this fundamental attribute of the SAM is still there. This can be illustrated by reference to Input-Output tables per se. Table 11 shows an input-output table with six sectors corresponding to those i n Table 7. Table 11 is the same as Table 10, with two exceptions. The f i r s t , which is t r i v i a l i n substance, is that it has been rearranged. Columns and Rows 4a t o 4f, which w e r e i n the lower right corner now occupy the top l e f t corner; Columns and Rows 1, 2, 3 are now below or t o the right of them. The substantial difference is that the lower-rlght hand corner of Table 11 is blank. The input-output table captures only the relationships between the production accounts and the other accounts (factors of production, institutions, capital account and rest of the world). Inter-relationships between these accounts, most of which are i n the top l e f t corner of Table 10 are not there. 47. If we were t o specify a new set of final demands different from those i n Table 11, techniques, based on specific assumptions about intersectoral relationships, exist for deriving the implied pattern of production i n each subsector. That implied pattern, of course, w i l l include income t o factors of production (Rows 1 and 2). The results, however, give no guarantee that there is any relationship between incomes generated and the ensuing demand- The complete SAM, in principle, provides the missing link--or a t least the data to establish it. . 48 Table 10 is as f a r as we shall go here i n subdiriding the accounts. In the original source, much greater detail is shown. In one complete matrix, the following appear: ( i ) Three labor groups (urban, rural and estate); ( i i ) Three capital groups (public, private and housing); (iii) Three household groups (urban, rural and estate); (iv) Two kinds of corporations (private and state); (v) Eleven production sectors. However, still more detail lies behind that matrix. For example, the production account is based on detailed accounts for 48 subsectors. Household data are based on information for six income brackets with each group. 49. This completes the first part of this introduction to the exposition of social accounting matrices. For those who may, unexpectedly, have been titillated by the subject, we have added one more section on a SAM for a different country, Botswana. This illustrates an attempt to bring into the framework of the accounts not only the various transactions already described but also the financial counterpart of these real transac- tions and the flow of capital funds into investment- In some ways this is one of the most intriguing uses of a SAM, as it brings together two related aspects of development, tbe real and the financfal. The parsimonious use of financial data and the neglect of its relationship with the real econony have often been noted. Botswana 1974175: The Flow of Funds 50. Our purpose at this point Is to illustrate how to introduce financial transactions into a SAM. The example we use is taken from a ll report on a SAM constructed for Botswana- C.C. Greenfield, Harry Fell and associates, Social Accounrinp:Matrix, Botswana 1974175, Ministry of Overseas Development, Statistics Division, Economic Planning Staff, London. November 1977. 51. The Botswana SAM shown as Table 12 is conceptually the same as Table 5 for Sri Lanka, subject t o three differences: ( i ) There is no separate line for indirect taxes; for simplicity they have been included with the Central Government, which is i t s e l f included under institutions ROW/CQPUUIU 2. ( i i ) The combined capital account has been moved down t o Row/Colunm 5 and the two accounts previously on lines 4 and 5 promoted t o 3 and 4, respectively. ( i i i ) Wew 1ines 6 and 7 have been added. One of them t o be explained later has been l e f t blank for the moment. The other consists of errors and omissions. Totals are now shown i n Rows and Columns 8. 52 The reason f o r not having a separate line for indirect taxes is t o avoid unnecessary clutter. Having made the point once, we do not need t o repeat it. 53. The change i n order is net a change i n substance, since there is no magic i n any particular order. The best order is the one that follows a reasonable logic and makes the SAM intelligible t o the reader. About this, there can be differences. The order here con£orms rather closely to the original, more detailed version of the SAM. 54. The account for Errors and h i s s i o n s is there for two quite different reasons. The f i r s t s e t of errors is attributable entirely t o rounding. Many individual figures in the original table have been added together t o form subaggregates; this inevitably involves rounding errors. However, there are several large specific errors, which appear as such i n the original SAM, mainly i n Row/Column 4 and 5. These errors are akin t o residual errors often l e f t i n the balance of payments or national accounts. A l lSAMs have such errors at some stage of their construc- tion. How these particular errors arose and why they were l e f t i n are explained i n detail i n the original source. Purchase and Sale of A s s e t s 55. The main change i n Table 13 from Table 12 is that Row/Column 6 has been labelled Financial Account with one entry each i n the column and the row, apart from the Errors and Omissions. In addition, there is a new diagonal element i n Row/Colum 5. These are basically the only differences. 56. A l l previous tables, whether on S r i Lanka or Botswana, have dealt with the consequences of current activity during the year, the production- income-expenditure cycle. Savings are savings out of current income and the invest w n t they finance encompasses only new investment. However, t h i s does not exhaust the totality of transactions. Institutions may buy or sell ,existing physical assets, particularly land and buildings. They also lend or borrow, creating financial assets or l i a b i l i t i e s as the case may be. The new entrfes are intended to acknowledge these facts. 57. The diagonal element of 30.3 i n Row/Colum 5 expresses the fact that institutions bought existing assets of that amount and also that they sold them. Obviously, the two must balance. When looked a t as an entry i n the column, the transaction appears as a purchase; when looked a t as an entry i n the row, it appears as a sale. 58. The new account (6) reflects a l l financial activities on capital carried out by balrks and financial enterprises.- The two new entries, at the intersection of Row 5 and Column 6 (127.3) and the intersection of Row 6 and Column 5 (126.6), are identical except for the errors and omis- sions. In principle they must be. They express the fact that institutions incur financial liabilities to the financial sector (for example, by borrowing) and also acquire financial assets from that sector (for example, currency-or bank deposits). These two must balance, since a liability automatically creates a corresponding asset. They must balance in toto, but, as we shall see, they need noc balance for any individual subset of institutions. 59 . We can now set out the capital account implied by the entries in ~ow/Column5 in the familiar double-entry form as follows: Incomings(row) Out~oings(column2 Domestic Savings 55.4 New Investments(rows 2-4) 110.0 Foreign Savings 69.4 Purchase of existing assets 30.3 Sale of existing assets 30.3 Acquisition of financial assets 126.6 Financial liabilities , Errors and Omissions incurred 127.3 Errors and Omissions -8.9 -.. Total 273.5 Total 233.5 Decomposition of the Capital Account 60. In Table 14 we have divided Row/Column 5 into three parts. The first is labelled 5abc and includes the group of institutions identified in Sri Lanka, namely, Households, Enterprises (or Corporations) and -1/ Current activities and new physical investments of these enterprises are already included as part of the accounts of "fnstitutivns." Government; the fact that it has an "abc" a t the end foreshadows the fact that it w i l l be broken down still further a t a l a t e r stage. The second, labelled 5d, consists of the new category of financial enterprises such as banks, which loom quite small i n the production processes we have hitherto considered, but loom much larger i n the' financial transactions that we are now considering. They do, i n fact, represent the capital and money markets through which most of the financfal transactions take place. The f i n a l element i n Row/Column 5e is the capieal account of the Rest of the World. 61 The sales of existing physical assets are now identified (fiabc, 5e) mainly as sales by institutions i n Botswana t o the Rest of the World and, t o a smaller extent, vice versa. This is explzined by the Interest of foreign corporations i n mining enterprises i n Botswana. 62. W e see that individual accounts do not have to balance insofar a s the financial transactions are concerned. Those who save may put their savings into real or financial assets. Investors i n real assets may borrow i n order t o finance them. This is the purpose of a capital market. 63. For example, we can set out the capital account of domestic institutions. other than financial enterprises to i l l u s t r a t e this point: Incomings (row) Outgoings (column) Savfngs 53.9 New investment (rows 2-4) 108.6 Sale of physical assets 23.8 Purchase of existing assets 6.8 Financial l i a b i l i t i e s Acquisition of financial assets 53.7 incurred 90.7 Errors and Omissions -0.5 Errors and Omissions -0.2 Total 168.9 Total 168.9 64. A similar account could be made for financial enterprises (5d). It would include rather modest amounts for savings and new investment (1.5 and 1.4, respectively). The principal elements i n this account would be the entries i n Row or Column 6, that is, the sale or purchase of financial assets. In practice, these are bound t o balance, except for the minor difference between savings and new investment, but, i n the table, balancing is achieved through Errors and Omissions. This arises, for example, because occasionally i n the year-end accounts or a borrower and a lender , a l i a b i l i t y and its corresponding asset may be valued differently. Decomposing the Financial Transactions 65. I n Table 15, the account for financial assets and l i a b i l i t i e s (6) 11 has been broken down into four different categories: domestic currency;- bank deposits and the like; domestic borrowing or lending; and foreign borrowing or lending. These are identified by Row/Columns 6a through 6d. 66 A s might be expected, institutions acquire additional resources (Row 5abc) by incurring l i a b i l i t i e s through domestic or foreign borrowing (Columns 6c and 6d). To the extent that they do not spend these resources on physical assets (old or new), they retain them for the most part in the form of financial assets: either currency cr deposits (Rows 6a and 6b i n Column 5abc). 67. The accoulnt of financial enterprises (5d) is of particular interest here. I n presenting it i n the usual form of outgoings and incomings, we make a minor modif-lcation, changing t i t l e s t o "Changes i n Assets" and "Change in Liabilities," respectively. We could do the same -I/ Actually, a t the time there was no independent Botswana currency; the currency then circulating was the Rand. for other capital accounts but the change i n this case brings out more clearly the nature of outgoings (acquisition of assets) and incomings (incurring of liabilities). The account thus reads: Channe i n Assets (column) Chanpe i n Liabilities (row) Physical assets 1.4 Savings 1.5 Currency 0.4 Deposits received 22.9 Deposits made 0.5 Domestic borrowing 3.2 Domestic lending 14.5 Foreign borrowing 5.7 Errors and Omissions -6.9 Errors and Omissions - -9.6 Total 23. 7 Total 23.7 68. Two itens appear on both sides of the account: deposits, and domestic lending or borrowing. There are a t least two reasons for this. Financial enterprises cover more than commercial banks. , Some of them make deposits i n commercial banks. These deposits appear a s a l i a b i l i t y t o the latter, but an asset to the depositing enterprises. Secondly, financial enterprises may both borrow on the market and lend t o their customers. They, therefore increase their assets by lending and their l i a b i l i t i e s by borrowing. 69. The account is, i n fact, analogous t o the change i n an enterprise's balance sheet from one year t o the next, except that it applies t o a s e t of enterprises. Savings here correspond to the increase i n equity investment attributable to retained earnings. Decom~ositionof Institutions 70. In Table 16, the final one for Botswana, institutions have been broken down into the same three constituent parts as for Sri tanka: Households; Enterprises (or Corporations); and Central Government. This has been done both f o r the current account (2) and the c a p i t a l account (5abc). 71. Except f o r the greater amount of i n s t i t u t i o n a l d e t a i l , there is no change i n principle from Table 15. However, t h i s new data now shows how each set of i n s t i t u t i o n s contributes t o the flow of c a p i t a l funds through the system. This information shows, f o r each of the three s e t s of i n s t i t u t i o n s and, f o r the t o t a l , t h e changes i n assets and l i a b i l i t i e s . I n an alternative format, ft can be presented as set out below: Households Enterprises Government Total Change i n A s s e t s (column) Physical assets (new) 15.7 67.7 25.2 108.6 1/ Physical a s s e t s (existing)- - 6. 7 -23. 7 -17.0 Financial assets 9.5 15.4 28.8 53.7 Errors and Omissions - - - - -0.2 0.1 -0.1 -0.1 Total 25.0 89.9 30.2 145.1 Clhanae i n ZiabPhities (row) Sa~bintgs 23.1 11.7 1981 53.9 Ptnanciel l i a b i l i t i e s 1.7 76.5 12.5 90.7 Errors and ~ i s s i o n s - - - - 0.2 1.7 -1.4 0. 5 Total 25.0 89.9 30.2 145.1 72. Households, as is of ten the case, saved more than they invest i n physical assets and, consequently, put the difference i n t o financial assets* Xnterprises were the principal investors i n physical assets and, -l/ Sales of existing assets have been brought over from the " liability" side a s a negattve item. This is a net figure. since their savings (or equity participation) is only a small part, most of it had to 5e financed by borrowing. In this particular year (1974/75), the role of the government was unusual, i n that it financed most of its new investment by the sale of existing assets. The Uses of SAMs 73. The effort required t o put together a SAM is not trivial. Data must be ferreted out, wherever it may be available. Conflicting sources must Somehow be Treconciled- Rows and columns do not conveniently come to the same t o t a l i n the f i r s t instance2' What does one get out of it a l l except a rather complicated and impressively tldy collection of numbers? 74- Because social accounting matrices have not been i n existence for long and there are not many of them, t o say what they are useful for is partly an exercise i n conjecture. Nonetheless, there seems t o be sufficient foundation t o make a few plausible suggestions. In the f i r s t place, a SAM is clearly a step forward i n the upgrading of statistics. Recent comparisons of microeconomic information obtained from household surveys with national accounts have shown that the discrepancies between these two sources of information can be very large. How do we choose between them? Or, perhaps should we ask whether we can choose between them? While construction of a SAM is certainly not going t o reveal the ultimate truth, a t least it forces attention on inconsistencies in a way that brings one closer to the root of their cause. Judgment, t o be sure, -1/ Techniques exist for making adjustments t o achieve balance a t minimum cost in terms of variation from the original; and new or improved techniques are being worked on. has to be used in imposing ultimate consistency, but it can be done in such a way as to keep adjustments within plausible limits and so avoid a purely Procrustean process of fitting one set of data to the dictates of another. 75. The concept of a SAM goes further than the improvement of statistics for their own sake. It could 3e said to be the common ground of economic planners or development economists on the one hand and statisti- cians on the other. A SAM is cast in a form that, given the fineness of detail with which it is constructed, makes the most of existing informa- tion. Economic models of an economy, which may be designed for quite particular purposes, nevertheless imply the existence of an underlying S M . Parts of this implied SAM may be aggregated and parts highly dis- aggregated, but it is nevertheless a SAW. The existence of an actual SAM, against which to test the behavioral assumptions of a model and the SAM they imply, is, on the face of it, a useful way of testing the modelos validity. Much has been and could be said about the relationship between models and SAMs.- Here we shall only touch on some of the simpler applications of a SAM to the understanding of the way in which an econony - works 76. Uses of SAM can be thought of in two categories: those in which the whole carpus of inforanation in the SAM is used; and those in which only a part is used. Of course, in the latter case, it is not necessary to have the complete SA??. But the construction of the complete articulated SAM means that one has at one's disposal a multipurpose tool and does not have to construct separate subsets of accounts for each purpose ad hoe. An -l/ Examples from the growing literature are de Melo (19791, Dervis(1978) and Dervis(1981). illustration of the use of part of a SAM has been documented in the case o. the Sri Lanka SAM which we have described in aggregate terms. The purpose of the exercise in question was to establish the order of magnitude of the total fiscal incentives for exporting in various sectors.- At the time (in 1970) substantial direct fiscal incentives were given to encourage nontraditional exports. At the same time, many industrial subsectors in Sri Lanka, among which, it might have been expected, would be found some potential exporters, received fairly high nominal protection- The input- output matrix within the Sri Lanka SAM was used to convert nominal protection rates into effective protection rates to these industries; in most cases, these were substantially higher than the nominal protection rates. The incentives in the tariff system that implicitly encouraged production for the domestic market could then be compared with the export incentives. In many cases, they greatly outweighed the export incentives and, in a number of others, reduced them to fairly small proportions. These findings, which were unexpected, could of course have been reached without a SAM; its existence, however, m d e the task easier. 77. One of the principal ways in which the whole corpus of information in a SAM ;an be brought to bear is through multiplier analysis, which shows how changes in one or more elements of a SAM generate changes elsewhere in the matrix. Here we will only ~onsidera very simple example to illustrate the approach. 78. The starting point is to assume a simple economy and a highly aggregate SAM which has accounts only for the private sector, government -1/ The exercise is described in Chapter 6 of Social Accountinp for Development Planning, Pyatt, Roe and associates (1977). production, the rest of the world and a combined capital account. To keep things even simpler, it is assumed that only the private sector buys goods from the rest of the world. Corresponding to this simple economy, we assume we have a SAM for some base period, and ask the question: what would happen if demands on production activities were increased by increasing government expenditure (by an amount i2); investment (by an amount i3); and exports (by an amount i5)? Without loss of generality we can put the sum of the i's equal to 1. 79. The first part of the answer to this question is that whatever the processes of consequential changes might be, the end result will be a new SAM for our simple economy. Moreover, those elements of the initial SAM which are zero by definition will remain zero. Because our model assumes that only the private sector buys goods from abroad, the purchases of such goods by government, for example, will remain zero.-1/ 80. Given the accounting rules and model assumptions, the difference between the new SAM and the original one will imply an incremental SAM in which many cells have zero entries. This incremental SAM is shown as Table 17. At this stage we know the items i2, i3 and i5, because these are the ' changes which we have exogenously postulated. We also know that the blank entries in the table are zeros, because these follow from our model and accounting conventions. The question then is, what can be said about the non-zero entries apart from i2, i3 and is? 81. Vot much can be said about these non-zero entries without making further assumptions, e.g., about what will happen to prices, monetary -11 Of course we could change the model and get a different result if we wanted to. The purpose here is to give a simple illstration. Table 17 Multiplier Effects in the Form of an Incremental SAM 2. Government 3. Capital Account 4. Production 5. Rest of World Notation: p2 = marginal propensity to taxation pg = marginal propensity to save = marginal propensity to consume (domestic) p4 p5 marginal propensity to import H = multiplier i2 = impulse from increased government expenditure i3 = impulse from increased investment 15 impulse from increased exports 5 policy, and haw people choose to spend any extra income. We w i l l assume, for simplicity, that private sector income goes up by an aaount M, and then explore what the incremental SAM i n Table 17 can say about the relationships between M and the i's- 82. Because i n this simple model the private sector gets a l l its income from production activities, and because these activities pay a l l value-added t o the private sector, Row 1 and Colunn 4 of the incremental SAM are very simple and contain zeros apart from the entry i n Row 1, Column 4, which is H. 83. The increase i n private income (Row 1) musr match the increase i n private expenditures i n Column 1. The l a t t e r must now be spread over the different components of private expenditures. This spread is assumed to take plate i n the proportions p2, pg, p4 and pg which can be referred to as the marginal expenditure propensities of the private sector. Since a l l the extra income M must be spent or saved, the accounting balance for Row/Column 1 implies that P2 + Pg + P4 + Pg = 1 W e might also be prepared to assume that these propensities are constant. But i f we do, then this is clearly a behavioral assumption, not an accounting rule. 84. Since Mpp is the only increase i n income for government, Row/Column 2 of Table 13 must have sums Mp2. From Column 2, this implies that the entry i n Row 3, Column 2, ice., the increase in government savings, must be Mp2 - i2. 85. For now, we skip over the details of accounts 3 and 4 and move to the rest of the world account, account 5. Rere, the only increase i n receipts is Mp5 because only the private sector buys imports i n this model. This then implies that the entry i n Row 3, column 5, must be Mp5 - This entry measures the extent t o which foreign savings or a reduction i5. i n domestic reserves of foreign exchange finances the increased investment, i3. 86. Returning now t o account 4, the fact that row and column sums must be equal implies that M = Mp4 + i2 i3 i5= Mp4 + + + 1 or M = l/(l-p4) In other words, the value-added M tcust be equal t o the aggregate increases i n government expenditure, investment and exports, inflated by a factor l/(l-p4). This factor is the familiar expenditure multiplier, i.e., the reciprocal of the complement of the marginal propensity t o consume domestic goods. Hence, while the SAM does not t e l l us what value t o give t o M or p4, it does show that once one is fixed, the other is also, and i n that sense it defines the relationship between the i n i t i a l increments i n expenditure (the i's) and the increase i n t o t a l value added, M. 87. A t this stage we have discussed the balancing of four of the five accounts df the incremental SAM. This is a l l that is necessary because it is always true that within a SAM the last account w i l l balance if a l l the others balance. To i l l u s t r a t e this point, the rule requires that, for our account 3, i2 i3 is= (p2 + p3 + + + p5) H Since the sum of the i's and the sum of the p's are each equal to 1, this can be written as: 1 = (l-p4) M , Hence the condition for account 3 to balance is the same as that for account 4, i.e., M = 1/(1-p4) This result simply repeats that obtained previously. If all but one account is balanced, then all accounts are balanced and the story of the incremental SAM shown as Table 17 is thus completed. 88. Application of multiplier analysis with a complete SAM is little different in priciple, though far more complex. It takes into account all the interactions within each step of the process of linkages between incomes, sxpenditures and production, e-g., the effects on other industries cf expansion within a particular industry. There is, howe-:=r, no longer a single multiplier, but an entire matrix of multipliers, which potentially shows the effect of expansion in one cell of the original SAM on any other cell. Row these effects are to be interpreted must always be approached with care since the effect of one variable on another ultimately depends on economic behavior and not just on accounting constraints. However, the approach has some value in distinguishing accounts or subaccounts which are likely to be affected from those which are likely to be bypassed. This may well have fmportance in considering the effect of exogenous changes on the distribution of income. The analysis may also serve to %dentify the important elements in resulting changes in government accounts or in the balance of payments. 89 Several different applications of this type of analysis have been made using the Botswana SAM. In such applications the SAM relationships can trace the conplex interactions inherent in the circular process- If initial changes in prices or wages are involved, the analysis can show--at least in terms of orders of magnitude--how the initial changzs affect prices in different industrial sectors and the consumption patterns of different household groups; if interindustry relations are at all complex and if, as is more than likely, household consumption patterns are very different, the resultant pattern may be difficult to predict a priori. Such analysis, however, is based in the first instance on the assumption that patterns of production and consumption are unaffected by price changes. Adaptation to take into account assumed responses can, however, be introduced. This, in fact, is of the essence in modelling an economy. 90. In a small open economy or a region, which is almost by definition a c m s l l open economy withi.- Tcountry, interindustry relations tend to be weak and leakages in the multiplier process large. SAMs have been constructe'd in each of these contexts: for Swaziland, for example, and for the Muda Valley in Malaysia.-I/ In both contexts, SANS have been used to explore the implications of large development projects whose "downstream" effects on the national or regional economy were of interest. Since interindustry relations tend to be weak in such cases, it is the pattern of household consumption that is most important in determining how large these downstream effects are and how they are distributed. 91. Construction of accounts for a region, as opposed to a country, are likely to reveal features of the regional economy that were little appreciated before. This is obviously because regions do not possess 11 national" accounts and other data normally associated with an economy as a -1/ See Bell and Hazel1(1980) for the Muda Valley SAM. whole. The construction of the SAM for the Muda Valley is a good illustration. I n this region a large irrigation project had been carried out, more or less doubling the output of rice, the main crop. Several features of the "downstream" effects of the resu?tant increase i n farmers' incomes are of interest. Perhaps the main one is the very large o u t f l w of capital from the region Lo the rest of Malaysia. This fact and other data i n the SAM are consistent with the theory that the principal downstream effect was t o increase the incomes of nonfarm households such as traders, who were i n effect "importers" from the rest of Malaysia. Leakages from the regional economy were thus substantial. It is perhaps significant that there was still a substantial number of poor landless laborers. Although the SAM may have been constructed too soon a f t e r the completion of the irrigation system t o allow for opportunities for reinvestment i n the region to be taken up, it is, nevertheless, a clear reminder that downstream effects can simply not be taken for granted. 92. CLearly no SBM can ever be constructed t o answer questions except i n the broad. Specialists i n any particular subJect may have a much better idea of specific consequences, based on accumulation of intimate know- ledge. But no one since Thomas Jefferson and his contemporaries can be a specialist i n everything. A SAM can be used t o bring out what is likely to be important i n any given context and, therefore, to order the consultation of specialist knowledge t o the occasion. A highly disaggregated SAM scch a s that reported i n Pyatt, Roe and associates (1973) for Sri kanka would, of course, be physically difficult t o reproduce on one sheet of paper. Even i f it were, it would be comprehensible only to a very limited group. The great advantage of a SAM is that one can select for any occasion those parts of it to be aggregated and those parts where detail is to be preserved. Learninp bv Doing 93. Explicit reference has already been made t o the fact that there are more detailed accounts of the SAMs used i n this paper, which the reader can consult, if he wishes to dig deeper. But, while further reading may give the reader a fuller taste of what a SAM is a l l about, there is probably no substitute for learning by doing. A do-it-yourself SAM does not have to be on the scale of the SAPZs described in this paper. National accounts, balance of payments and financial data (such as the central bank's balance sheets and consolidated statistics for coomoercial banks) are often readily accessible and are sufficient t o s t a r t the construction of a rudimentary SAM, or even a series of SAMs for different years. The data may appear Inconsistent or inadequate even t o do this; or other questions may crop up. Once one s t a r t s asking questions about the data the closer one comes to appreciating what a SAM is for. 94. Data of this kind can often be found i n the appendices of World Bank reports. The example shown i n Table 18 is taken from a recent report on the Yemen Arab Republic. The SAM itself is only is only a f i r s t cut using eight tables i n the appendix for basic information and two others to make very crude estimates of the allocation of indirect taxes (Line 2b) and imports (Line 4) t o the two production sectors (Columns 3a and 3b).- I/ Anyone with access t o the original data could easily improve on it. The - 11 The report i n question is entitled Yemen Arab Republic. Development of a Traditional Economy, a World Bank Country Study, January 1919. The tables used were Nos. 2.1, 2.3, 2.4, 3.1, 3.8, 5.1, 6.1, and 6.3 for the basic data and 3.4 and 3.5 to make the crude estimates. point here is to show how even a limited exercise can throw up question: both of substance and of consistency for further investigation. 95. There are 39 entries i n the core of the SAM and If totals (each appearing twice); of these, 26, including the rough estimates, could be directly entered i n the core of the SAM and 10 in the totals. The rest ! ! followed easily by simple addition or subtraction, making, it is true, some I quite arbitrary assumptions about the location of small residuals; there was one independent check on the outcome. 96. A special feature of the table is the Row/Column lb devoted t o remittances. A large part of the Yemeni labor force was working i n Saudi Arabia and the Gulf States a t that me. In the year in question, gross remittances (there was some reverse flow) were equal t o nearly 50% of factor income. The proportions were changing extremely rap2dly; the corresponding figure i n the previous year was less than 25X and i n thz following year over 752. The response of the private sector t o this rapid increase i n their resources was t o save more than 25% (Column 2a). Of . these savings, less than half was invested i n physical assets; the rest, plus a substantial amount of borrowing, was retained i n the form of currency and deposits with commercial banks (Column 5). A t the same time, the government was borrowing abroad, more than enough t o finance its investment ( i t too, accumulated funds i n the central bank, Row 8, Column 6a). 97. Two other features of the SAM are the consequence of the form in which the data were available. The public sector capital account has been s p l i t into budgetary and "other" transactions (Row/Columns 6a and 6b). Government expenditure (investment plus consumption) in the national accounts exceeded expenditure i n the budget. Similarly, official loans or grants i n the balance of payments exceeded borrowing recorded i n the government budget. There were evidently "government" transactions outside the budget, a large part of which must have been due t o investment i n public enterprises. The division is intended t o draw attention to t h i s fact; some gratification can be drawn from the fact that the adjustments required t o balance the line and column were trivial. 98. Similarly, there are two subdivisions of the Rest of the World capital account. One is for the official - balance of payments. However, again, there were additional transactions. This was evident from the fact that there was an increase i n foreign assets (Row 7b) over and above the officially recorded reserves i n the central bank (Row 7a). This has been balanced by a corresponding inflow of capital, here allocated mainly t o the private sector. One residual i n the table is the inter5ection of Row 8 and Column 9. This represents an increase i n deposits of the commercial banks with the central bank. Obviously, thete should be l i t t l e difficulty i n checking t h i s figure, but our purpose here has s i m ~ l ybeen to present what can be done with a particular s e t of information rather than going beyond it. 99. The time invested was not large, about half a day of uninterrupted timeA/ Since this included the preparation of SAM'S for six successive years, it seems a small price t o pay for an articulated s e t of accounts, which reveals, a t least i n order of magnitude, the salient features of the economy. It is at least arguable that six successive tableaux of this kind ,gThe coapiler, moreover, had no previous familiarity with the country concerned. give a better appreciation of change--in this case kaleidoscopic change- than do individual tables of the traditional kind. If this f i n a l exampl: strikes a mundane note on which t o finish, that may not be inappropriate. The SAM approach is a flexible tool which can be deployed with varying degrees of sophistication. While economists have long since understood that their various analyses can each be s e t within a framework of accounts, this aspect has not usually been developed. A s an all- purpose tool, SAMs can well be used for a variety of purposes, once an i n i t i a l investment has been made t o learn how. The general point is that an economist who understands SAMs w i l l probably be better equipped t o tackle a variety of problems than one who does not. Bell, C. and P.B.R. Raze11 (1980), *Measuring the Indirect Effects of an Agricultural Investment Project on its Surrounding Region," American Journal of Agricultural Economics, Vol. 62, No. 1, (February 1980). World Bank Reprint Series No. 154. Chander, R., S. Gnasegarh, G. Pyatt and J.I. Round (1980),"Social Accounts and the Distribution of Income: The Malaysian Economy i n 1970," Review of Income and 1-Iealth, Vol. 26, No.1 (March 1980). World Bank Reprint Series No. 143. de Melo, M. (1979), "Agricultural Policies and Development: A Socioeconomic Investigation Applied t o Sri Lanka,' Journal of Policy Modellin~,Vol. 1, No. 2. Dervis, K. and S. Robinson (1978), "The Foreign Exchange Gap, Growth and Industrial Strategy i n Turkey: 1973-83,'' World Bank Staff Working Paper No. 306, Washington, D. C., The World Bank , J. de Melo and S. Robinson (1981), General Equilibrium Models . for Developinv Policy. Cambridge University Press (forthcoming) Greenfield, C.C., 8. Fell and Associates (19771, Social Accounting Matrix: Botswana 1974/75, Ministry of Overseas Development, Statistics Division, Economic Planning Staff, London. Pyatt, G. and E. Thorbecke (1976), Planninp Techniques for a Better Future, I.L.O., Geneva Pyatt, G. and A. R. Roe with R. M. Lindley, J. 1. Round and cthers (1977), Social Accounting for Development Planning: With Special Reference t o Sri Lanka. Cambridge University Press. and J.I. Rand (19771, "Social Accounting Matrices for Development Planning," Review of Income and Wealth, Series 23, No. 4, World Bank Reprint Series No. 74. United Nations---(19681,-stem of National Accounts, Series F, No. 2, Rev. 3, New York. World Rank (1979), Yemen Afab Republic, Development of a Traditional Economy, a World Bank Country Study, Washington, D. C.