2016년 2월 23일 화요일

[발췌: Peter Hill's] The services of financial intermediaries or FISIM revisited (1996)


출처: Peter Hill. “The services of financial intermediaries or FISIM revisited.” Paper presented to the Joint UNECE/Eurostat/OECD Meeting on National Accounts, Geneva, 30 April-3 May 1996. Submitted by the Australian Bureau of Statistics.


※ 발췌 (excerpt):

The Author's Note:

This paper has been stimulated by three other recent papers on the subject of calculating an allocating FISIM by Eurostat(1996) and Australian Bureau of Statistics (Carl Obst 1995 and 1996). These papers have moved the subject forward to the point at which it should be possible to reach consensus on these issues. There is much common ground between them, despite disagreement on one final point. The present paper seeks to take many of the arguments already advanced in the Eurostat and ABS papers to their logical conclusion and is intended as a further contribution and clarification. In order finally to settle the treatment of FISIM, it is necessary to go back and undertake a critical review of the basic principles underlying the measurement of FISIM in the SNA to expose a basic misconception about the productive activities of financial intermediaries that has dogged the subject for the last few decades.


Introduction

The measurement of the production and consumption of the services of banks, or financial intermediaries, has a long and troubled history in national accounts. There are two main reasons for revisiting FISIM at the present time and undertaking a critical review of the basic principles underlying its measurement. The first is that the treatment of FISIM was not in fact settled in the 1993 SNA. To quote the first two paragraphs of page xliii of the 1993 SNA:

  “The 1993 SNA ... represents a stage in the evolution of national accounting. To continue that evolution, further research will need be carried out. Consensus must be reached on certain topics before the can be incorporated into international guidelines and standards. ... In 1993, the Statistical Commission agreed that the highest immediate priority was to develop guidelines for the explicit allocation of financial intermediation services charges indirectly measured (FISIM) to specific users.” (emphasis added)

  Inter-Secretariat Working Group was charged with the responsibility of pursuing this and a few other topics in order to incorporate some improved guidelines and standards into the SNA without waiting 25 years for another comprehensive revision. In the meantime, and until an international guideline could be agreed on FISIM, the Statistical Commission recommended that countries should be permitted some "flexibility" with regard to the treatment of FISIM. This paper should not be construed as re-opening issues that were decided in the 1993 SNA. On the contrary, the treatment of FISIM is still very much on the agenda of unfinished business. There is no international standard as yer for FiSIM. 

  The expedient adopted in the 1968 SNA of treating the charge as the intermediate consumption of a "nominal industry" is no solution. It is debatable whether inventing an imaginary industy with negative value added is much better than recording negative value added for financial intermediaries. It is certainly much less transparent to users. Recording negative value added for financial intermediaries would, of course, understate GDP, but the 1968 expedient continues to understate GDP, although to a lesser extent. The major concern of many countries and the European Union, however, is that the 1968 expedient understates GDP to an extent that varies from country to country.

  If financial intermediaries are actually producing services, some other real institutional units must also be consuming those services. They cannot disappear into a black hole. If output is to be recorded in a comprehensive, integrated system of economic accounts the use of that output must also be recorded, a view felt strongly by many experts involved on the 1993 revision.

  The second reason why this unfinished business of the 1993 SNA needs to be finished quickly is the urgent need not merely to arrive at an agreed solution, but an economically defensible solution, within the context of the 1995 ESA. The calculation and allocation of FISIM affects GDP and has important policy and financial implications therefore.


The nature of the services produced by financial intermediaries

  There are presumably two points on which there is a consensus and which are taken for granted throughout this paper.

(1) Financial intermediaries are actually engage in production as defined in the SNA. They incur costs of production in the form of intermediate inputs, compensation of employees, and capital consumption in the same way as other producers. They are also profitable. However, the exact nature of some of their outputs is somewhat elusive and their value is difficult to estimate.

(2) Financial intermediaries perform a unique and highly important function in the economy by facilitating the flow of funds from lenders to borrowers. For this reason, financial corporations are distinguished from non-financial corporations at the first level of sectoring in the System, as explained in para. 4.77 of the 1993 SNA.

  However, lending out of own funds, as distinct from intermediated funds, is deemed in the SNA not to require any productive activity. It will be argued here that this restriction on the productive activity of financial institutions is based on a conceptual error that results from failing to distinguish the act of lending ^per se^ (which is not production whatever the source of the funds) from the activities that precede, accompany and facilitate it. The restriction is unjustified and unnecessary, and seems to be the main reason why it may be difficult, or even impossible, to reconcile the estimated total value of services provided to individual units or sectors with global FISIM calculating using the standard SNA definition.


Calculating FISIM at the level of individual institutions or sub-sectors

The 1993 SNA provides explicit guidance not merely on the nature of the services financial institutions produce and how they should be valued but also to whom they are delivered. Para. 6.124 makes the familiar point that financial institutions do not always "charge their customers individually for services provided" but charge higher, or pay lower, rates of interest "than would otherwise be the case"; that is, than if some 'pure' market rate of interest were used. Financial institutions choose to substitute higher or lower interest charges, or payments, for the direct payment of fees for many of the services they render to their customers. In order to charge indirectly in this way, they must be able to fix the interest rates they charge or pay, discriminating not only between lenders and borrowers but also between individual customers.

  The significant point in paragraph 6.124 is that it addresses explicitly not merely the question of how the services should be estimated and valued but ^to whom^ the services for which no explicit charges are made are delivered. If the guidelines given in paragraph 6.124 are followed, there are no great conceptual problems about estimating the production and consumption of the services, although there may be data problems, including the choice of an appropriate 'pure' rate of interest, generally called the 'reference' rate of interest.

  The description just given of the way in which financial intermediaries are considered to go about charging indirectly for many of their services is meant to be realistic. The fact that many of activities of financial intermediaries are finance by manipulating the interest rates they charge, or pay, their customers is widely accepted. For example, on page 9 of the Eurostat paper it is stated that "FISIM output is generated by the ^management^ by financial intermediaries of loans and deposits whose ^rates they control" ... (emphasis added).

  The customers to whom financial intermediaries are providing services consist of those institutional units who have made deposits with them or taken loans or advances from them. These units have ^continuing relationships^ with the intermediaries which go beyond pure creditor/debtor relationships.

  In practice, financial institutions engage in three broad groups of productive activities delivering different kinds of services to their customers:

  (1) taking, managing and transferring deposits;
  (2) making loans or other investments;
  (3) offering financial advice or other business services.

They often charge specific fees for some of these services, especially the third kind, but their customers are likely to be charged indirectly for the first two by reducing interest payments on deposits or increasing interest charges on loans. When they take deposits, financial intermediaries do not simply borrow funds. They provide security and convenience to their depositors and also perform the important function of managing the transferable deposits which are used as means of payment. Similarly, when they make loans or advances, they do not simply lend. They have to mobilize funds in order to be in a position to offer immediate and flexible creit facilities to their customers, most of whom would otherwise find it difficult, costly and time consuming to raise finance.

  Like many other producers, especially large producers, financial intermediaries are thus not engaged in just one type of production but in two or more different activities at the same time. In practice, financial intermediaries deliver two different kinds of services in parallel to their depositors and borrowers for which they charge separately. Of course, the activities complement each other and there are good economic and financial reasons to explain why intermediaries carry out both together. But each activity nevertheless remains productive in its right, delivering different services to different customers.


( ... ... )

Lending out of own funds

The definition of 'global' FISIM as given in para. 6.125 of the 1993 SNA is that "total value of FISIM is measured in the System as the total property income receivable by financial intermediaries minus their total interest payable, excluding the value of any property income receivable from the investment of their own funds, as such income does not arise from financial intermediation." The definition is carried over from the 1968 SNA (para. 6.33).

  This definition implies that the services that financial intermediaries actually provide to their customers are only to be recorded  and valued when the activities which produce the services are not only undertaken jointly but specifically linked to each other. When a financial institution lends funds, it is deemed to be engaged in production only if the funds it lends have been borrowed for the purpose, even though the institution itself may be incapable of identifying the origin of the particular funds in question and even though the activities involved may be the same whatever the origin of the funds. By implication, taking deposits is also deemed not to be productive unless the funds deposited are lent to others.

  The justification for excluding lending from own funds seems to be that lending is not itself a process of production in an SNA sense. If one institutional units lends to another the transaction is recorded in the financial accounts of both parties, and no entry is needed in the lender's production account. The newly created financial asset in the form of a loan is certainly not output. For example, when a household makes a deposit with, i.e., lends to, a financial institution, it is not producing anything. Nor does the household have to engage in production to make the deposit. It seems to follow from this that when a financial institution simply lends its own funds it cannot be engaged in production either. This is a fallacy, however, which results from failing to distinguish the lending from the productive activity in which an institutional unit is obliged to engage when it makes a business of lending to many customers.

  ( ... ... )

  The service charges demanded by lenders in the business of lending is a reflection of the fact tht they are engaged in production and providing services for which their customer have to pay if they want to obtain loans. These services are the same whether the lenders use their own funds or borrowed funds. In either case, the lending itself continues to be recorded in the financial accounts of both parties and does not involve the production account. It may be concluded that the fundamental distinction drawn in the SNA between lending out of own funds and intermediated funds is irrelevant and misconceived. It cause the production boundary to be drawn too narrowly and imposes an unnecessary restriction on the productive activities of financial institutions.


Financial intermediation as a process of production

  Because lending itself is not production, and if lending is not differentiated from the activities which precede, accompany and facilitate it, a need may be felt to invent another form of productive activity, namely financial intermediation, to justify the inclusion of the activities of the major financial institutions concerned within the production boundary, as they obviously must be. Financial intermediation may be an extremely important function performed by many financial institutions but it is not a process of production. Moreover, as just noted, by insisting that financial institutions are only productive when they engage in intermediation, the SNA understates the contribution that financial institutions make to the production of the economy.

  Financial intermediation is defined in para. 4.78 of the 1993 SNA as "a productive activity in which an institutional unit incurs liabilities on its own account for the purpose of acquiring financial assets by engaging in financial transactions on the market." This definition is given in the chapter concerned with the definition and classification of institutional units and not in the chapter concerned with the concept and definition of production. It is intended not as a definition of a process of production in an SNA sense, but as a description, in an ISIC sense, of the distinctive nature, or characteristics, of the activities undertaken by financial intermediaries which enables the institutions concerned to be distinguished and classified separately from other institutional units at the first level of sectoring in the SNA. The description is needed and actually used to identify those corporations that should be classified as financial corporations and include in Sector S.12 (see para. 4.77 of the 1993 SNA).

  A process of production in an SNA, or general economic, sense has to have clearly specified inputs and outputs of goods and services (see para. 6.15 of the 1993 SNA). Financial intermediation cannot be interpreted as a production process in this sense.  Intermediation may perhaps be interpreted as the transformation of liabilities into assets, but this kind of transformation makes no sense as a process of production. It is totally different from the physical transformation of inputs into outputs. Financial assets and liabilities are not goods and services which are consumed as intermediate inputs or produced as outputs. Instead of being consumed by the intermediation process, the liabilities are actually created in the process and continues to exist alongside the assets after the production has taken place. The new assets, the loans, are no more 'outputs' than the newly created liabilities.

  In fact, of course, the outputs produced by financial intermediaries are neither assets nor liabilities but the services they provide to their customers─the borrowers and lenders involved in the intermediation process.


Property income from securities

  Why financial institutions buy and sell long or short term securities on the open market they are price takers who cannot manipulate, or control, the prices at which they buy or sell and hence the interest rates they receive or pay. They do not provide services to the units from whom they buy or to whom they sell. Their relationships with the latter are no different from those of any other units such as households, non-financial corporations or government units trading on financial markets.

  The same position is taken on page 9 of the 1996 Eurostat paper: "there is no intermediation service for securities other than shares since the relationship between a financial intermediary and another unit is the same as the relationship between two non-financial intermediary units. ... Thus, when financial intermediaries issue bonds, they do not render FISIM to the units acquiring those bonds. Likewise, when the financial intermediaries acquire bonds, they do not render FISIM to the issuers of those bonds either." The fact that financial institutions are unable to levy service charges on the units from whom they buy, or to whom they sell, securities on the open market has also been noted by other writers on FISIM.

  If a financial institution issues securities in oder to make loans, it may be engaged in intermediation but the only services produced are those provided to its borrowers. The interest paid on the securities is irrelevant and should not enter into the calculation of FISIM, either at the level of the individual institution or on aggregate for the economy as a whole. Conversely, if financial institutions purchase securities, the property income receivable should not enter into the calculation of FISIM, whatever the source of the funds. In any case, the interest payable or receivable on marketable securities may be so close to the reference rate that it may make little difference in practice whether they are included or excluded, but nevertheless it is conceptually wrong to include them. The same conclusion is reached in the Eurostat paper where it is stated on page 9: "Consequently, only loans and deposits (including loans and deposits internal to the FIs) are to be considered as generating a financial intermediation service."

  It may be concluded that the global definition of FISIM is conceptually wrong on two counts: (1) by including property income and interest on marketable securities in the calculation of FISIM and (2) by restricting the property income receivable to income from loans or other assets acquired out of borrowed (intermediated) funds. It is now necessary to turn to the actual calculation of the production and consumption of the services for which financial institutions charge indirectly, a task which is not the same as trying to find the the least bad way of allocating a global charge that is incorrectly defined in the first place.


Total service charges payable by depositors and borrowers versus global FISIM

  It is necessary to compare the sum of the service charges payable to financial institutions by their depositors and borrowers with global FISIM as defined in the SNA. Call the Sum of the Service Charges SSC to avoid confusion with global FISIM. SSC needs to be calculated for all financial corporations in sub-sectors S.122 and S.123 of the 1993 SNA. The treatment of the Central Bank may be deferred for separate consideration and the FISIM considered here should therefore be assumed to exclue the Central Bank.

SSC

  As explained in paras 6.124 to 6.128 of the 1993 SNA, [:]
  • the indirect or implicit service charge payable by a borrower from a financial institution is equal to the excess of the 'interest' payable over what would be payable if some 'pure' rate of interest, or reference rate, were to be charged
  • Similarly, the service charge payable by a depositor is the excess of the interest that would be receivable if the reference rate were used over the 'interest' actually receivable
  • Inverted commas are place about 'interest' in this context as a remainder that, on the assumptions made, the actual amounts of 'interest' receivable or payable are not true interest, in an SNA sense, but true interest plus or minus a service charge.
  As already emphasised, two facts (not assumptions) underlie this method of estimation. The first is that the indirect charges received by financial institutions must be paid by real institutional units in respect of services delivered to these units, who must be their customers. The second is that financial institutions can only levy charges indirectly when they are in a position to control the interest rates they pay or charge their customers. It follows that indirect service charges are only levied on depositors and borrowers in respect of their deposits and loans, as also argued by Eurostat.

  It is convenient to introduce some notation to explain the calculation of SSC and to see how it differs from global FISIM.


Let L=Total Loans, D=Total Deposits, OF=Own Funds, 
    R(L)='Interest' rate on loans, R(D)='Interest' rate on deposits, r=reference rate of interest.

The total value of the indirect service charges receivable by a financial institution from its customers, or its SSC, is given by the following expression:
SSC=[R(L)L-rL] + [rD-R(D)D]      (1)
     =[R(L)L-R(D)D] -  r[L-D]      (2)
The term in the first bracket in (1) is the total indirect service charge receivable from borrowers, while the second bracket is the total indirect service charge receivable from depositors. The charges are recorded as intermediate or final consumption expenditure of domestic customer, or as exports. Expression (1) appears to be identical with the formula proposed by Eurostat on page 17 of their paper for calculating the service charges for individual sectors. The formula follows inexorably from the way the charges are made.

  rL, loans multiplied by the reference rate, is true interest payable by borrowers, while rD is the true interest receivable by depositors. These are the interest flows to be recorded in the SNA's primary income accounts. It is also worth noting that when it is stated in the texts of the SNA and ESA that the global FISIM is equal to total property income receivable minus total interest payable, the terms 'property income' and 'interest' should, in principle, also be in inverted commas to draw attention to the fact that these flows are not actually the property incomes or interest recorded in the primary income accounts but the true property incomes or interest ^plus^ or ^minus^ the implicit service charges.

  Expression (2) is merely a rearrangement of (1) to which it is algebraically equivalent, but it is often convenient to be able to refer to this version of the SSC formula. It can be seen that SSC is independent of the reference rate only in the special case in which the total values of loans and deposits happen to be equal to each other, in which case SSC simplifies to total 'interest' receivable ^minus^ total 'interest' payable. (Assuming there is also no other property income on other assets or liabilities, SSC would also equal global FISIM as defined in the SNA in this simple special case, because all loans could be treated as being made out of the deposits, no deduction being needed for the interest received on own funds.) When loans and deposits are not equal, SSC depends not only on the reference rate selected but also on the size of the difference between loans and deposits. If loans are greater (less) than deposits SSC is less (greater) than total 'interest' receivable ^minus^ 'interest' payable. If loans exceed deposits, SSC varies inversely with the reference rate chosen, while if deposits exceed loans SSC increases as the reference rate increases.

  The value of the total indirect service charges receivable by all financial institutions in sub-sectors S.122 and S.123 must be equal to the total of the SSCs receivable by those institutions. Of course, as SSC is calculated as the sum of the indirect charges payable by all the customers of financial institutions, the consumption of the services is estimated simultaneously with the production.

  In general, loans need not equal deposits. In particular, a financial institution may make loans out of its own funds as well out of borrowed funds. It is necessary to bring in the rest of the balance sheet in order to analyse this case. In order to focus on the main issue, suppose initially that the institution does not have securities as assets or liabilities so that there are only four items in the balance sheet: fixed asset, denoted by FA, loans, deposits and own funds. The balance sheet can be written as an equation as follows:
FA+L=OF+D    (3)
Following the kind of reasoning used by Eurostat, assume that fixed assets are financed out of own funds, with only the residual own funds being available, and used, to make loans. In this case, lending out of own funds, denoted by L(OF) is given by:
L(OF)=OF-FA=L-D      (4)
Now, SSC is unchanged and continues to be calculated solely with respect to total loans and deposits using expressions (1) and (2) above. However, global FISIM as defined in the SNA requires the 'interest' received on own funds to be deducted from the difference between total 'interest' receivable and total 'interest' payable, because lending own funds is not intermediation. Thus, if follows from (4) that
FISIM=[R(L)L-R(D)D] - r(L)L(OF)
     =[R(L)L-R(D)D] - r(L)[L-D]     (5)
By comparing (2) with (5), it can be seen that the only difference between SSC and FISIM in this case is the term [L-D] is multiplied by r, the reference rate, in (2), whereas it is multiplied by R(L), the actual interest rate on loans, in (5). Or, to put the same point differently, if the deduction for the interest earned on own funds were to be estimated using the reference rate instead of the actual loan rate, there would be no difference between SSC and FISIM. However, the SNA definition of global FISIM clearly requires the use of the actual rate as the intention is to deduct "any property income receivable from the investment of ... own funds" (para. 6.125) and not the interest rate that would have been earned if own funds had been invested at the (lower) reference rate. It follows, incidentally, in this particular case that FISIM should be less than SSC, which is interpreted here as implying that FISIM would underestimate the total indirect service charges.

  This example can be taken one stage further. The difference between SSC and FISIM must get progressively larger as the proportion of the loans financed out of own funds increases. Consider the limiting case in which all loans are made out of own funds. In this case,
D=zero  and L=L(OF)
SSC=R(L)L-rL=[R(L)-r]L     (6)
whereas FISIM=R(L)L-R(L)L(OF)=zero     (7)
In principle, the first term in (7) denotes the total interest actually receivable while the second term denotes the deduction for interest earned on own funds, the two cancelling each other out. The SSC is equal to the total service charges payable by the borrowers, while FISIM is zero.

  Conceptually, the SNA seems to go even further by implying not merely that the estimated value of the output of the financial institution happens to be zero in this case, but that the institution is not even engaged in production. In the present example, the SNA position implies that if a financial intermediary progressively reduces the deposits it takes while maintaining the same level of lending by increasing use of own funds, it winds down its production of services to its borrowers at the same rate as it winds down its production of services to its depositors, eventually ceasing to provide any services to its borrowers when it stops taking deposits. This cannot be true if in fact it continues to provide exactly the same services to them and maintains the same client relationships with them as before. Using SSC, the total output of the institution also decreases as the services to depositors decline, , but it does not al to zero, the services provided to the borrowers continuing to be measured and valued as in (6), in effect as proposed in para 6.124 of the SNA.

  The next step is to introduce bills and bonds, i.e., securities other than shares, into the balance sheet. The calculation of SSC again remains unchanged in this case, as it equals the total value of the services provided to depositors and borrowers, no services being provided by the financial institutions to the units from whom they buy, or to whom they sell, securities (for reasons already explained). However, the introduction of bills and bonds begin to make the calculation of the global SNA FISIM more problematic as it becomes more difficult to pin down the precise way in which the own funds available for investment were actually invested and hence the precise amount of interest receivable from their investment. The technical problems are well spelled out on pages 10 to 13 in the Eurostat paper and need not be repeated, especially since they are interpreted here as not being real problems but pseudo problems created by adopting a defective definition of global FISIM. As the Eurostat paper shows, it may not be possible to derive a general formula which is satisfactory at a micro level without implicitly departing from the global SNA formula.

  In general, the total SSC for all financial institutions cannot equal global FISIM because of their different treatments of lending from own funds and interest on securities (and other property income, if any). If the service charges are calculated for individual institutions or sectors using the formula proposed here and by Eurostat (expression (2) above), they cannot add up to global FISIM. The conclusion that the global formula is flawed is inescapable. It should be ignored as irrelevant. The total services charge for the economy as a whole is given by the sum of its components, as with other aggregates in national accounts. By recognising this, the problem of allocating FISIM is solved, or rather disappears. The whole issue is to close to a satisfactory resolution that it ought not to founder by insisting on adhering to a global formula that has ^never actually be derived explicitly and rigorously form first principles^ and is seen to be exposed to serious conceptual deficiencies when examined in detail.

  There is, of course, no problem about 'allocating' SSC as the total SSC is the sume of its components. Eurostat suggests disposing of the discrepancy between global FISIM and total SSC by treating it as intermediate consumption of financial intermediaries. This might be the least bad way out of the impasse if the global SNA definition of FISIM were to be treated as a binding constraint, but the constraint is both unnecessary and unwarranted.

  It might be argued that the global definition of FISIM proposed in the SNA is not intended to yield a precise estimate of the aggregate indirect service charges receivable by financial institutions, but is meant to be a simple formula which provides a robust way of approximating to the total value of a set of services which is not easy to calculate in practice. However, lack of precision and rigour are not characteristic of the SNA and are not acceptable in other areas. The SNA seeks to provide precise concepts and definitions which are valid and consistent at both a micro and a macro level, putting practical estimation problems on one side until the concepts and definitions are settled.


References

Eurostat (1996). "Proposal for a Council regulation on the allocation of FISIM": the Working Party on National Accounts, Luxembourg, 28-29 March 1996.

Obst, C. (1995). Australian Bureau of Statistics, "Financial Intermediation Services Indirectly Measured (FISIM)", Paper presented to Eurostat, 6 Nov., 1995

Obst, C. (1996). Australian Bureau of Statistics, "Comments on the Proposal for a Method of Calculating and Allocating FISIM": Joint UN ECE/Eurostat/OECD Meeting on National Accounts, Geneva, 30 April-3 May 1996.

댓글 쓰기