2016년 2월 19일 금요일

[발췌: J. Assa's] Financial Output as Economic Input: Resolving the Inconsistent Treatment of Financial Services in the National Accounts (2015)

자료: ... ...


※ 발췌 (excerpts): 


1. Introduction

The three major types of financial services are treated in the standard national accounts in three different ways. 
  • Capital gains are excluded a priori from the production accounts;
  • interest flows generated by financial intermediation are treated as an input to other industries and deducted from total value-added to arrive Gross Domestic Product (GDP); and
  • fee-based financial services are considered productive and are imputed a value-added based on net revenue.

While there is consensus regarding the exclusion of capital gains since there is no productive activity associated with them, the other two treatments─of interest-based financial intermediation and fee-based financial services─are more controversial. Furthermore, there is an inconsistency in how the standard accounting framework treats these two sources of financial income. While the netting out of interest-based income does not affect overall GDP, the value-added imputation for the fee-based income inflates GDP by its amount, leading to a divergence of this measure of output from its historic correlation with other variables such as employment and median income.

  After reviewing some of the technical issues involved in the debate over the assumed productiveness (of lack thereof) of interest-yielding financial intermediation services, the discussion turns to the fee-based financial revenues which have received less attention. Unlike the interest-based part of financial incomes, the fees generated from financial services are not netted out of GDP and show up as value added on the production (output) side of the accounts. We assess the differences between finance and the other sectors for which value added is imputed, both conceptually and empirically, and argue that it is always an input (or a cost) for other industries and the economy as a whole, having no final use value. We then construct an alternative measure of overall economic activity which treats finance as a cost to be netted out from total value added, consistent with the way financial intermediation services indirectly measured (FISIM) income is treated in the standard accounts. The resulting adjusted measure of output─Final Gross Domestic Product or FGDP─includes only ^final^ goods and services, and is further reconciles with the expenditure and income side of the accounting framework. ( ... ... )


2. Accounting for Finance

GDP "is the primary indicator of economic activity and...can be estimated in three ways, which are theoretically equal" (Lee 2012) and must identically agree, thus also providing mutual control for each other. In compiling national income statistics from various sources using three methods, however, the results inevitably disagree, and are reconciled by including a "statistical discrepancy" in the accounts. The expenditure approach to GDP is the sum of all final expenditures, and is denoted by the familiar equation below:
(1) GDP=C+I+G+X-M
     That is, GDP equals the sume of consumption of final expenditure by household, investment by firms, final consumption by government, and net exports. The income approach, by contrast, is as follows:
(2) GDP=CE+NT+GOS
     Here GDP equals the sume of compensation of employees (wages plus benefits), net taxes (taxes on production and imports less subsidies), and gross operating surplus (profits). Finally, the production (output) approach to GDP sums up all activities deemed productive across industries:
(3) GDP=Σ(Yi-ICi)+NT
     where Y stands for output, IC for intermediate consumption, and the term in the summation expression represents value added for each industry i. GDP is thus equal to value added plus taxes minus subsidies. As Lee explains, "[o]utput is all the goods and services produced, whilst intermediate consumption comprises all the goods and services consumed or transformed in a production process. The net taxes are included in order to put all three approaches on a consistent valuation basis" (ibid). In other words, value added itself is not directly comparable to GDP by expenditure or income approaches, and we must use GDP by the output approach (i.e. value added+taxes-subsidies) to make comparisons with other two GDP measure more consistent.

  Next, when discussing value added in 'finance', the official System of National Accounts (SNA) includes financial intermediation, insurance and pension funds, and other activities such as administration of financial markets. Some authors (e.g. Basu and Foley 2013) add real-estate, resulting in the FIRE acronym. It is important to note, however, that even the narrow definition of finance which refers only to financial services (and does not include real-estate, insurance or other business services) itself comprises three types of activities performed by financial industry (including non-bank financial institutions):
  • 1. Services for which banks explicitly charge a fee, and are thus relatively straightforward to record in the national accounts. These services include overdraft fees, foreign exchange commissions, consulting on mergers and acquisitions, underwriting securities, as well as market-making activities (Akritidis, 2007, Haldane 2010).
  • 2. Financial intermediation resulting in net interest income─this part of banks' business is not as easily captured. "Finance─and commercial banking in particular─relies heavily on interest flows as a means of payment for the services they provide. Banks charge an interest rate margin to capture these intermediation services" (Haldane 91), which gives rise to the FISIM income mentioned above.
  • 3. Net Spread Earnings (NSE), e.g. capital gains or dealing profits from spot trading in the foreign exchange market.
  These three types of financial services are treated in the national accounts in three different ways. 

  • Capital gains are excluded a priori from the production accounts; 
  • interest flows generated by financial intermediation are treated as an input to other industries and deducted from total value-added to arrive at GDP; and
  • fee-based financial services are considered productive and are imputed a value-added based on net revenue.

While there is consensus regarding the exclusion of capital gains since there is no productive activity associated with them, the other two treatments─of interest-based financial intermediation and of fee-based financial services─are more controversial. Furthermore, there is an inconsistency in how the standard accounting framework treats these two sources of financial income.

  Since fees paid for financial services are easily captured by national accountants, most of the debate has recently focused on the (non-fee based) net interest part─financial intermediation and the imputation of its output through FISIM. Financial intermediation has long been problematic to measure. Christophers (2011) describes the history of the so-called 'banking problem'─the fact that, without imputations, the value-added of the financial sector (that is, output minus intermediate consumption) would be negligible or even negative (since if its costs are deducted from fee-base revenues alone, the former would often exceed the latter).

  • At a first stage in the history of this question (SNA 53 and before), all financial intermediation activities were excluded from calculations of national output based on the value-added approach, since they were considered to be mere transfers of funds (similar to social security payments) and hence unproductive.
  • An intermediate approach followed with the SNA 68, where the output of the financial sector was considered to be an input to a notional (i.e. imaginary) industry which has no output. In spite of the bizarre nature of this approach, "ascribing a negative income to an imaginary industry sector...has probably been the most used for financial intermediation services in the entire history of Western national accounting" (Christophers 130).

  A useful example of the Gross Value Added (GVA) of the UK financial sector in 2003, which would be £39.8 billion under SNA 68 (4.1% of total GVA). The imputed banking service charge (IBSC), however, was a negative £45.9 billion. Under SNA 1953 the financial sector would have thus shown a negative £6.1 billion value added. "Adopting SNA 1968 had, in effect, made UK finance productive" (Christophers 130, emphasis in original). Table 1 below illustrates this point.


Finally, with the 1993 SNA, financial intermediation became an explicitly productive activity, for which value added is imputed based on the net interest received by financial institutions (the FISIM approach)[n.1]. As can be seen from Table 2 below[n.2], the FIRE sector is bigger under SNA 68 (in absolute terms) since it includes FISIM income (line 14), which is then deducted from value added to derive GDP (thus affecting only the relative size of the financial sector rather than the total GDP). In SNA 93, however, FISIM is not deducted as it is already distributed to various uses (i.e. deducted individually from the value added of the industries purchasing financial intermediation services). As a consequence, several industries show a lower value added (in absolute terms) under SNA 93 than when one uses the SNA 68 approach (Agriculture, hunting, forestry and fishing; Mining and quarrying; Manufacturing; Electricity, gas an water supply; and Public administration and defense; compulsory social security).

[n.1] The latest revision, SNA 2008, extends the boundaries of SNA 1993 to include ever more exotic financial 'products'.
[n.2] The data in Table 2 are for 1996, since that was the last year for which data were reported accoring to both SNA 68 and SNA 93 for the United States.

Table 2: ...

  Even within the 1993 FISIM framework, however, there are two possible estimation methodologies.
  • The approach recommended by SNA 93 is to allocate FISIM across the sectors and industries that use these services, in order to "identity the purchase of these services explicitly and to classify them as intermediate consumption, final consumption expenditure or exports according to which sector incurs the expenditure" (Akritidis, 2007, 30). Use of this approach gives rise to Oulton's (2013) argument that "if banking output has been overstated, then the output of some other industry or industries must have been understate" (p. 3). 
  • However, given practical difficulties with allocating FISIM to users (because of conceptual and data availability problems), SNA 93 allows for a simpler approach, which treats all of FISIM as the intermediate consumption of a 'nominal sector' (simpler to the SNA 68 method). Using this more simplified approach GDP is not affected by the size of the FISIM output” (ibid.) Oulton further argues that even using the first approach, the FISIM component has a negligible effect on GDP growth estimation (Oulton, 6). This observation notwithstanding, the FISIM approach still distorts the relative size of the financial and other sectors within the economy.
Two comments are approriate at this stage. First, different countries use different methodologies based on their preferenes, data availability and conceptual choices. The UK and several other European countries use the "nominal" sector approach, while the US (since 1996) distributes FISIM to uses as shown in Table 2 above. Second, the relative shares of FISIM vs. fee-based services in overall financial incomes have changed over time. As Akritidis observes, the share of FISIM income in total banking income declined from 72% in 1992 to 66% in 2004, while "the share taken by explicit charges, such as fees and commissions rose" (Akritidis, 30).

  This fact, as well as the existence of a simpler FISIM approach which does not affect overall GDP, raises the following question: why are fee-based financial services treated as value added, while interest-based financial intermediation is netted out of GDP as intermediate consumption (of either a nominal sector or the total economy?) This inconsistency is understandable from a measurement point of view, since fee-base financial services are easy to capture and therefore present less of an empirical problem than the FISIM issue. From a ^theoretical^ point of view, however, the non-FISIM part of financial services, that is, the fee-based income in the GDP-by-output approach, is as problematic as interest-based income. Finance, in its various manifestations, ultimately involves the transfer of money. Unlike other commodities, money has no use value, only an exchange value. In fact it ^is^ exchange value ^par excellence^. Gold and silver still had some practical uses when they were the common means of payment, but fiat money is merely symbolic. As the textbooks tell us, money serves as a unit of account, means of exchange, and store of value. Neither consumers nor firms can directly consume money, but rather purchase goods and services with it, either for final consumption or for intermediate consumption in a production process.[n.3]

( ... ... )


댓글 쓰기