1 Session Number : 3 Session Title : Measurement Under Inflation Session Organizer : Jean-Etiennne CHAPRON Paper prepared for the 25th General Conference of The International Association for Research in Income and Wealth Cambridge, England 23-29 August 1998 Interest and Inflation Accounting (A Reply to the Paper by André VANOLI) Peter HILL (ECE, Geneva 1 ) For additional information please contact: Peter Hill, Statistical Division, UNECE, Palais des Nations, 1211 Geneva. Email: Peter.Hill at UNOGECE Fax: (41) 22 917 0044 Tel: (41) 22 917 1145 1 The views in this paper are those of the author and do not necessarily represent those of the United Nations Economic Commission for Europe,
2 Introduction This paper has been written in response to the paper by André Vanoli for the session on Measurement Under Inflation. As most of André Vanoli s lengthy paper is devoted to criticising the Manual which I drafted for the OECD on Inflation Accounting, I am obliged to reply, especially since the criticisms do not appear to be valid. In order to understand the substantive issues involved it is necessary to go back to Annex B to Chapter XIX of the 1993 SNA, also drafted by André Vanoli, on A parallel treatment of interest under significant inflation. This Annex contains a number of conceptual errors which make it inconsistent with the economic and accounting principles on which the SNA is based. André Vanoli s paper to the present Conference provides a convenient opportunity for highlighting the differences between Annex B and Inflation Accounting which need to be more widely understood and addressed by the international national accounting community. The timing is opportune since the mechanism is being put in place to update and correct the 1993 SNA. For brevity, the three papers under discussion, which are listed under References at the end of this paper, are referred to simply as Annex B, Inflation Accounting and the IARIW Paper (by André Vanoli) This paper tries to focus on the main issues of principle on which Annex B and Inflation Accounting disagree. Responding to every single detailed point made in the lengthy IARIW Paper, some of which are merely tendentious and rhetorical, would scarcely interest the general reader. Such a paper would tax the patience of even the most dedicated national accountant. Readers are presumably interested in substance and essentials. At the core of the disagreement are few fundamental and basic points of principle from which everything else follows. These points of principle are important not only in the present context but also because countries suffering from high inflation, now or in the future, need to be clear about what are the appropriate accounting procedures. Inflation is likely to be around in some parts of the world or other indefinitely. The subject of inflation accounting has moved on since the 1980 s and some of the practices followed in the past by some countries can now be seen not to have been optimal. The main reason why the situation has improved is that the 1993 SNA now provides a much more comprehensive and precise framework for accounting under high inflation and the benefits should not be squandered through incorrect implementation. Perhaps the single most important improvement of the 1993 SNA over the 1968 SNA is the elaboration
3 of the Other Changes in Assets Account, which replaced the old Revaluation Account, and which has made it possible to have an exhaustive and detailed accounting for changes in the values of assets between the opening and closing balance sheets. Parts A and B of Chapter XII of the 1993 SNA both break new ground by distinguishing much more carefully changes in the value of assets which are due to changes in their quantities from those due to changes in prices. Changes in the values of asset due to changes in their prices constitute nominal holding gains and a clear distinction must be maintained between price and quantity changes if nominal holding gains are to be measured correctly. The core of the disagreement between Annex B and Inflation Accounting is that they measure nominal holding gains on financial assets such as deposits and loans differently. In view of the importance of the Other Changes in Assets Account to the issues under discussion in this paper a brief recapitulation of the role of this account in the SNA as a whole may be useful as a starting point. The Other Changes in Assets Account and Holding Gains The Other Changes in Assets Account records changes in the values of assets between the opening and closing balance sheets which are not attributable to the transactions recorded in the Capital and Financial Accounts. The account is sub-divided into two separate accounts: the Other Changes in Volume of Assets Account which, as its name suggests, records value changes which are due to quantity changes, and the Revaluation Account which records value changes due to changes in the prices of the assets. Changes in the values of assets due to changes in their own prices constitute nominal holding gains, so that the Revaluation Account actually records nominal holding gains. It is, in fact, the SNA s Holding Gains Account and it would have been more apt and informative if it had been so named instead of the rather ambiguous and misleading Revaluation Account 2. As is well known, it is a basic valuation rule in the SNA that at any, and every, moment of time assets should be valued at their market, or estimated, market prices at that time (see, for example, the section on general principles of asset valuation in paras. 13.25 to 13.28 of the 1993 SNA). Provided that assets are valued in this way, a fundamental accounting identity holds whereby the 2 The term Revaluation Account is an unfortunate hangover from the 1968 SNA. It suggests that assets may have to be revalued in order to reconcile their opening and closing balance sheet values or to correct inappropriate valuations (as, for example, in business accounts when an asset may be revalued from historic cost to current cost). Such revaluations are superfluous in the SNA where the prices of assets at any moment of time are determined by the market and cannot be adjusted at the discretion of their owners, creditors or debtors, or their accountants.
4 difference between the values of an asset recorded in the opening and closing balance sheets is fully accounted for by transactions in the asset plus other volume changes plus nominal holding gains. Nominal holding gains can be decomposed into neutral and real holding gains, if desired, but the basic identity holds in terms of nominal holding gains. The identity is explained in various places in the 1993 SNA such as paras. 10.15-10.19 and 12.85-12.86 and the Annex to Chapter XII where an algebraic proof is given. The main difference between Inflation Accounting and Annex B is that Annex B records nominal holding gains on financial assets such as deposits and loans whereas Inflation Accounting does not. As the nominal holding gain on an asset is equal to the quantity of the asset multiplied by the change in its price over the period in question (see paras.12.68 to 12.73 of the 1993 SNA) it is essential to establish exactly what are the prices and what are the quantities for the assets in question. Although this may seem trivial, apparently it is not so. In general, the price of any good or asset is defined as the number of units of national currency for which it can be exchanged. (In principle, something else could serve as the numeraire, but national accounts use monetary prices denominated in the national currency.) The price of a unit of the national currency itself is unity and cannot change. A deposit or loan of $100, for example, consists 100 units of currency --- these are the quantities --- each with a price of unity. A deposit or loan of 200 is twice as large as one of 100: it is not twice the price. As the price of a unit of currency is always unity, the price of these 100 units of currency cannot change over time so that the nominal holding gain for the depositor or lender must be zero (actually, this is already explained in detail in paras. 12.68 and 12.78 of the 1993 SNA). Every economist knows this and it should not be necessary to have to labour the point if it were not for the fact that Annex B does record nominal holding gains on loans. To avoid possible misunderstanding, it is worth digressing briefly to note that the price of a unit of foreign currency in terms of a unit of national currency is not fixed because it changes whenever the exchange rate changes. (The treatment of assets denominated in foreign currency is dealt with later.) Similarly, the price today, or forward price, of a unit of national currency receivable n years from now can also change because the forward price is the discounted price of a unit of currency n years from now and therefore varies with the rate of interest. Nominal holding gains can occur on deposits denominated in foreign currency or sums of money due for payment on a fixed date in the future (bonds).
5 Index linked loans Against this background, it is convenient to start by examining the treatment of index linked loans under high inflation in Annex B. By way of general explanation para. 17 of Annex B says: The amount entered as nominal holding gains/losses is the following.... For index-linked financial instruments denominated in national currency it is straightforwardly the amount which results from the application of the index-linking mechanism. Annex B does not explain why this is a straightforward nominal holding gain. It would be difficult. The fact that some countries may have been misled into adopting this treatment in the past is no justification. Para. 12 (b) of Annex B takes the numerical example of a non-financial corporation which is assumed to borrow 2,000 from a resident bank. The loan is index linked to the internal wholesale price index which trebles from 100 to 300 over the period. Annex B argues: The revaluation account records 4000 (2 times 2000) as a nominal holding loss for the non-financial corporation (revaluation of a liability) and a nominal holding gain for the bank (revaluation of an asset). As the price of units of currency has not increased over the year and obviously remains equal to unity, there can be no nominal holding gain or loss on the 2000 units of currency borrowed at the beginning of the year. It is necessary not to be deceived by appearances and terminology but to examine the underlying economic realities and to reflect them in the accounts accordingly. This is a procedure which the SNA adopts repeatedly and indeed provides the basic rationale for the SNA as a system of economic accounts 3. The economic facts here are: first, the price of units of currency cannot change, and second, the creditor and debtor have agreed on a mechanism for adjusting the amount of the debtor s liability under certain circumstances. The way in which to record this is quite simple and explained in Inflation Accounting. The debtor is simply treated as paying a capital transfer to the creditor which the latter then lends back again to the debtor. It is obviously the size of the loan (the number of units of currency lent and borrowed) which is increased and not its price. To treat it as a price increase would be as absurd a 3 A good example of a situation where the treatment in the SNA may not coincide with appearances or the perceptions of the units involved is provided by financial leasing, although there are many others, notably in respect of social contributions.
6 saying that a loan of 2,000 is twice the price of a loan 1,000. 4 The imputed payment which the debtor makes to the creditor fits the concept of capital transfer in the SNA perfectly (a point taken up again later in view of the assertion in the IARIW Paper that it is not a transfer) because the creditor is required to use the transfer in order to acquire an asset; i.e., to make an additional loan (see paras.10.131 and 10.132). The agreement between the creditor and the debtor is best described as an agreement to write up or write down a debt in proportion to the changes in the index selected for index-linking. The 1993 SNA already explains how to treat the writing down or writing off a debt by mutual agreement between the creditor and debtor (see paras.10.139 and 11.23). In the case of debt forgiveness, for example, the creditor is recorded as paying a capital transfer equal to the value of the loan which the debtor uses to pay off the loan. The treatment of index-linking is exactly the same except that index linking will usually (but not necessarily always) require the debtor to make the transfer to the creditor, rather than vice versa, as it may be presumed that the debt is likely to have to be written up when there is high inflation. Another reason why the procedure proposed in Annex B contravenes SNA accounting principles is that the size of the supposed nominal holding gain depends on the choice of some other index or price to which the loan is index linked whereas a nominal holding gain depends on the change in the price of the asset itself. Changes in other prices or indexes are irrelevant (although not, of course, for real holding gains, which are not under discussion at the moment). As an asset has only one price, there can be one, and only one, holding gain depending on how that price moves. It is not possible to have a whole range of different nominal holding gains on the same asset which depend on the kind of agreement reached by the creditor and debtor. On the other hand, the amount by which a loan is written up or down by mutual agreement between them can be varied by as much as they please. The next example given in para. 22 (c) of Annex B is flawed for the same reasons. This example is the same as in para. 22 (b) except that it is assumed that the whole of the increase in the loan due to index linking is repayable at the end of the year in question. Thus, the 4,000 which Annex B erroneously records in the Revaluation Account is assumed to be repaid in addition to any of the original principal due to be repaid. The correct procedure is to record the 4 Only when liabilities are specified to be repaid at different dates in the future can their prices differ. For example, the market price of a zero coupon bond for 1000 to be redeemed 15 years in the future will be less than that of a bond for 1000 due to be redeemed in 5 years. As already noted, the forward prices of units of currency not payable until some future date are less than unity because of discounting and they obviously vary between assets with different maturities, just as they may change over time if interest rates change.
7 additional 4,000 that the creditor lends to the debtor out of the capital transfer of 4,000 as being immediately repaid to the creditor, the two transactions effectively cancelling each other out in the financial accounts of both parties. The creditor is left with a capital transfer of 4,000 5. Nominal Interest Under High Inflation The next example given in para. 22 (d) of Annex B deals with the central issue of how to record payments of high nominal interest under high inflation by treating it as if it were a simple extension of the previous example in which the whole of the increase in the loan due to index linking is repayable at the end of the year. This method of presentation is very misleading as it prejudges the issue by assuming that the two cases are parallel when they have nothing in common. The basic assumptions for purposes of the example in para. 22 (d) are that 2,000 is lent at a nominal rate of interest of 215% and that the rate of inflation over the year is 200%. In other words, nominal interest payments amount to 4,300 while the price level trebles. No assumption is made about repaying the principal (in contrast to 22 (c)). According to Annex B, the creditor receives protection against inflation equal to 4,000 as a result of the implicit price adjustment mechanism (see para. 13 (b)) contained in the high nominal interest rate. This is treated as giving rise to a nominal holding gain for the creditor and holding loss for the debtor. The main features of the accounting may be summarised as follows: 4,000 is recorded in the Revaluation Account as a nominal holding gain/loss for the creditor/debtor; 4,000 is also recorded in the Financial Account as repayment of principal by the debtor: this is interpreted as some kind of accelerated repayment of the loan; 300 is recorded as real interest, or interest prime (see below). Three comments may be made. 5 Although recording the payment and repayment of the additional loan may appear cumbersome, it reflects the economic realities and is certainly no more complex than the treatment of employers imputed social contributions (see paras. 7.45 to 7.47) for example where the same kind of payment and repayment occurs.
8 The creditor is treated as automatically receiving a nominal holding gain, despite the fact that this is impossible when the price of the currency in which the loan is denominated cannot change. On the other hand, as a result of this purported nominal holding gain, the creditor is treated as incurring no real holding loss, despite the fact that it is the real holding loss which is responsible for the high nominal rate of interest in the first place. The debtor is treated as making an accelerated of repayment of principal, despite the fact that there is actually no repayment of principal, the debtor s liability at any moment of time remaining fixed in terms of a certain number of currency units, here 2,000. On all three points, the method of accounting in Annex B is diametrically opposed to the economic realities of the situation. It reflects obsolete ideas on inflation accounting which have been superseded by the 1993 SNA. The appropriate method, which is given Inflation Accounting and summarised in the following paragraphs, is both simple and economically realistic. Economists have long recognised that high nominal interest is the mechanism whereby the creditor obtains compensation for the real holding loss expected by the creditor on the principal as a result of high inflation. The object therefore is to capture this reality in the accounts. The amount of the creditor s actual, or ex post, real holding loss is measured at the end of the year on the basis of actual rate of inflation experienced. It is 4,000 in the example used in Annex B. The payment of nominal interest of 4,300 should therefore be partitioned into two components, one of which represents a capital transfer while the remainder denotes real interest. The main features of the accounting may be summarised as follows: one component of nominal interest represents the payment of a capital transfer of 4,000 from the debtor to the creditor to compensate the creditor for the real holding loss: it is recorded in the Capital Accounts of both parties; the remaining component, 300, constitutes real interest which is recorded in the Primary Distribution of Income Accounts. The accounting is thus fairly simple. The debtor s liability in the balance sheet remains 2,000 throughout, as it must do. The only entries in the Financial Account are the changes in cash or other short term assets associated with the
9 actual payments of the capital transfer and real interest. There are no entries in the Other Changes in Assets Accounts for either party. To be more precise, there are no nominal holding gains or losses to be recorded in the Revaluation Accounts. Even zero or non-existent nominal gains on monetary assets and liabilities, can be decomposed into neutral gains/losses and real holding losses/gains of equal value (see para. 12.79 of the SNA). In the present example, the creditor s real holding loss can be shown explicitly in the Revaluation Account, if desired, in which case an equal neutral gain of 4,000 must also be recorded to cancel it out so that the nominal gain remains identically equal to zero. It should be noted that the real holding loss for which the creditor is compensated by the high nominal interest remains in the creditor s Revaluation Account. Recording real interest instead of nominal interest does not require real holding gains and losses on monetary assets and liabilities to be included in the income accounts of the SNA (which is not feasible). This is a common fallacy which was prevalent in discussions of inflation accounting in the 1980 s and may have blocked progress towards the recording of real interest in the income accounts of the SNA under conditions of high inflation. Compensating the creditor by a capital transfer The IARIW Paper argues that the payment to the creditor in compensation for the the latter s real holding loss is not a transfer. It quotes the SNA definition of transfer: namely, A transfer is defined as a transaction in which one institutional unit provides, a good, service or asset to another unit without receiving in return from the latter any counterpart in the form of a good, asset or service (para. 10.131). The IARIW Paper then argues that there is a defined counterpart, the holding gain (real according to Inflation Accounting) accruing to the debtor. The debtor makes a payment to the creditor. On the other hand, it is obvious that no good, service or asset is received in return from the creditor. A holding gain is not some form of good or asset whose ownership can be transferred from one unit to another or some form of service that one unit can provide to another in a transaction. Holding gains cannot be counterparts. As already stated, one of the main advances in the 1993 SNA was to emphasise that certain changes in the values of assets are not attributable to transactions and have to be recorded in the Other Changes in Assets Account, the most important of which in practice are nominal holding gains. To argue that a holding gain can form part of transaction is mistaken. The argument that the
10 payment made to the creditor cannot be a transfer because there is a counterpart in the form of a holding gain is quite spurious. Of course, every transfer has a reason or purpose. Transfers are not made gratuituously or capriciously. In the present case, the reason the debtor makes a payment to the creditor is to compensate the creditor for the real holding loss incurred on the loan. Notice that the real holding loss is not due to any action on the part of the debtor but to the general rate of inflation over which neither party has any control, even though the creditor presumably expected some inflation when rate of interest on the loan was fixed. In general, transfers in the SNA typically consist of payments of compensation to units suffering some kind of loss or payments of social benefits to units with some special kind of need. Payments of insurance claims are treated as current and not capital transfers because from the point of view of the insurance enterprises receipts of net premiums and payments of claims must be treated as current transactions. From the point of view of the policy holders, however, it might be preferable to treat claims as capital transfers, but the same transaction cannot be classified in two different ways in the same system of macro economic accounts. Para. 10.141 of the SNA specifically states that payments of compensation for damage or losses not covered by insurance policies are to be classified under Other Capital Transfers (D.99). When real interest is recorded rather than nominal interest in the income accounts of the SNA, it would be better to create a new separate category of capital transfer under the general heading Capital Transfers (D9). As already noted, the payment of compensation for a real holding loss fits the concept of a capital transfer in the SNA perfectly since it is expected that the payment received will be used to acquire an asset to maintain the creditor s real net worth and not be regarded as available for consumption (see para. 10.132). In the case of index linked loans, the transfer is automatically reinvested in the loan. Real interest and interest prime One idiosyncracy of Annex B is its insistence that real interest must be non-negative. Such interest is called interest prime. Para. 3 of Annex B states: Real interest is derived from nominal interest by taking account of real holding gains/losses on the underlying assets/liabilities. For this reason, however, real interest may not be included in the central framework of the System since it is a basic principle that holding gains or losses should not be recorded in the current accounts of the central framework, but only in the
11 revaluation account. This is true for nominal holding gains /losses (and a fortiori real holding gains/losses) on all types of assets and liabilities. Thus real interest as well as other adjusments of current incomes for real holding gains/losses may only be introduced in a satellite construct.... There is a logical fallacy here. The objection to real interest given above, and repeated in the IARIW Paper, is that real interest cannot be recorded in the central framework because holding gains/losses cannot be recorded in the current accounts of the central framework. It is, of course, perfectly true that holding gains losses cannot be recorded in these accounts -- a point equally stressed in Inflation Accounting -- but it is false to assume that it is necessary to do so in order to record real interest. As just shown above, recording real interest does not require holding gains/losses to be introduced into the current accounts of the SNA and the reason given in Annex B and the IARIW Paper for dismissing real interest is spurious. It can even be seen from Table 2 of the IARIW Paper itself (p. 24) that the real holding loss of the creditor remains in the Revaluation Account when real interest is recorded in the Income Account. Notwithstanding the fact that Annex B argues that real interest cannot be introduced into the current accounts of the SNA it then proceeds to propose its own method of doing just that, subject to the constraint that real interest must be non-negative. Real interest subject to this constraint is called interest prime. First, it should be noted that if it were indeed to be the case the central accounts of the SNA were incapable of recording negative real interest the SNA would be fairly useless to economic analysts and policy makers dealing with high inflation for whom the phenomen of negative real interest is extremely important and should be signalled in the accounts when it occurs. Fortunately, the SNA is capable of recording real rather than nominal interest if that is what analysts and policy makers prefer and need 6. When the ex post rate of interest is negative, the amount of nominal interest actually payable is not sufficient to cover the whole of the capital transfer needed to compensate for the creditor s real holding loss, so that the deficit is recorded as being paid back by the creditor to the debtor as the negative real interest. This seems to be what economists understand by negative real interest. The argument in Annex B and the IARIW Paper that when real interest is negative part of the creditor s real holding loss has somehow crept into the income account is not correct. Interest prime is an unnecessary distraction from the central issue of inflation accounting which is to determine the appropriate way in which to record real interest, whether positive or negative. 6 When inflation is low, users seem to prefer nominal interest to real interest and for this reason the SNA proposes that nominal rather than real interest should normally be recorded in the Primary Distribution of Income Account.
12 Financial instruments denominated in foreign currency The first example in para. 22 (a) of Annex B refers to a situation in which a resident unit borrows foreign currency from a foreign bank. It shows that when there is a fall in the exchange rate the borrower incurs a nominal holding loss (an increase in the price of a liability) because the price in terms of national currency of each unit of foreign currency borrowed increases as the exchange rate falls. As noted at the beginning of the present paper and also in Inflation Accounting, the price of a unit of foreign currency must change in terms of the national currency whenever the exchange rate changes (the exchange rate is, in effect, no more than the price of a unit of foreign currency) so that nominal gains/losses do indeed occur on loans denominated in foreign currency. The example in para. 22 (a) is valid for this reason. Annex B and the IARIW Paper then proceed to argue that this case sets the precedent for arguing that nominal holding gains/losses can similarly occur on loans denominated in national currency, whereas it provides no precedent whatsoever. A loan denominated in foreign currency is quite different from one denominated in the national currency because the price of a unit of foreign currency can change whereas the price of a unit of national currency in terms of itself is obviously always unity. The examples in 22 (b), (c) and (d) referred to above have nothing in common with 22 (a) and it is misleading to present them as if they were simple extensions of 22 (a). The IARIW Paper actually draws comfort from the fact that the examples in 22 (b), (c) and (d) are consistent with the way of recording financial instruments denominated in foreign currency (p.25). By now, it is hoped that readers will draw their own conclusions. Conclusions The main purpose of this paper has been to show that Annex B to Chapter XIX of the 1993 SNA is seriously flawed conceptually and inconsistent with the basic economic and accounting principles on which the 1993 SNA is based. A second purpose, which is closely linked to the first, has been to explain why the criticisms of Inflation Accounting made in the IARIW Paper are invalid. As André Vanoli has chosen to force these issues out into the public domain they need to be addressed by the wider community of scholars interested in the subject of national accounting and not only by those responsible for updating and correcting the SNA. As is well known, several parts of the 1993 SNA, and not only Annex B, are already in urgent need of updating, improvement and correction, and action should not be delayed.
13 References Annex B (A parallel treatment of inflation under significant inflation) to Chapter XIX of System of National Accounts, 1993: EU, IMF, OECD, UN, and World Bank, 1993, pp. 455-460. Inflation Accounting : A Manual on National Accounting Under Conditions of High Inflation, prepared by Peter Hill, Consultant to OECD, OECD, Paris,1996. Interest and Inflation Accounting, by André Vanoli, (INSEE), a paper prepared for the 25th IARIW Conference, Cambridge, August 1998.