1. The concept of income

To understand the theoretical rationale for the full accrual method of recording interest income, it is first necessary to review the concept of income in economic statistics. In economic theory, there are considered to be four factors of production: land, labor, capital, and entrepreneurship. It is the use of these factors of production that is considered Income, not the return to the owners of the factors of production. (The returns to the owners of the factors of production are: rent, in the case of land; wages, salaries and other forms of employee compensation, in the case of labor; interest, in the case of capital; and profits, in the case of entrepreneurship. ^4/) However, the return to a factor of production is considered to be the most appropriate way in which to both value and describe the use of the factor of production. For example, the wages that an employer pays to an employee can be used to value the labor used by the employer. The use of labor is classified as compensation of employees as this is considered to be the most meaningful label to describe the activity.

It is this concept of the use of, rather than the return to, the factors of production that determines the sign convention associated with income entries in the balance of payments. The users of the factors are acquiring a real resource; hence income debits are recorded when a resident of the compiling country uses factors of production provided by residents of another economy. (In the balance of payments, along with other forms of accounting, debit entries denote the acquisition of resources, both real and financial.) The returns to the nonresident owners of the factors of production, which are typically provided by way of financial instruments, are recorded as credit entries elsewhere in the balance of payments. Thus, when one considers the interest income component in the balance of payments, it is the use of the capital on which interest is charged that is being measured, not the actual payment for the interest.

Having established that income refers to the use of, rather than the return to, the factors of production (although the latter can be used to value the former), the economic statistician is not constrained to recording income at the time payments are actually made. Rather, income should be recorded in the balance of payments during the period or periods in which the economic benefits arising from the use of a factor of production are enjoyed by the user. (An analogous case is one when goods are acquired by their new owner in one period but where payment is made in another period. The transaction in goods is recorded at the time the economic benefits of the goods are acquired by the new owner–by convention, when the goods change ownership—rather than when payment is actually made.)

For most reasonable economic analyses, factors of production are considered to be in more or less continuous use. If an enterprise borrows capital in order to acquire a machine, then the economic benefits—the machine’s output—enjoyed by that enterprise accrue whenever the machine is in use. Likewise, an employee working forty hours a week provides his or her labor during each of those forty hours, and not just in the hour in which wages are actually paid. Therefore, in order to reflect the continuous nature of the use of factors of production, income in economic statistics should be recorded on an accrued basis, and not at discrete points in time when payments are actually made (or due to be made). ^5/

It is interesting to compare the treatment of interest in the fifth edition of the BPM with the fourth edition treatment. In the fourth edition of the BPM. it was recommended that interest should be recorded on a due for payment basis. This basis for recording reflects a debtor’s use of capital even when the payments for the use of that capital are in arrears and, as such, has some of the features of recording interest on an accrual basis. In defense of the fourth edition, it should be noted that at the time (mid 1970’s) when it was written, interest on most forms of debt instruments was typically due for payment either in or very close to the period in which the capital was actually used, particularly when statistics are compiled on an annual basis. It was only in the 1980’s when international financial markets became more sophisticated—particularly with the growing use of zero coupon and other deep discounted bonds—that the distortions in economic analysis caused by the use of the due for payment basis became apparent.

2. Financial account offsets

The balance of payments, like the national accounts, has its roots in double entry accounting. This, of course, requires that every transaction is recorded in a transactor’s records twice, once as a credit entry and once as a debit entry. For the accrual of interest, the offset entries to investment income accrued will typically be recorded in the financial account. When Interest becomes due for payment, these financial account entries should be extinguished, offset either by a transaction in the instrument by which payment is actually effected, or by a new short-term liability (arrears) if payment does not occur when due.

The financial account entries, which represent a liability of the user of capital (the debtor) to the provider of capital (the creditor), that offset accrual of interest should be recorded against the same instrument as the instrument through which the capital was extended in the first place. Only interest in arrears (also recorded on a full accrual basis) should be reflected as investment into a separate instrument. In other words, for example, the offset to Interest accrued (before it is due for payment) on bonds should be recorded as a transaction in bonds and the offset to interest accrued (before it is due for payment) on loans should be recorded as a transaction in loans.

For tradable instruments, an argument for treating offsets to accrued interest in this manner is that creditors “reinvest” accrued interest back into the underlying instruments. The argument is based on the premise that creditors could have, if they wished, realized the accrued interest by selling their bonds; if the bonds are not sold then this is prima facie evidence that the creditor willingly extended additional financing, via the accrued interest, to the bond issuer.

Obviously, this argument does not carry through to non-tradable debt instruments (such as loans). With such instruments, the creditor, presumably, has no way of realizing any accrued interest until the interest becomes due for payment. However, for both tradable and non-tradable debt, the terms–either explicit or implicit—for the payment of interest are an inherent part of the contract between the debtor and the creditor and therefore can be considered an Integral part of the instrument by which capital was provided. Only when interest is in arrears is there a breaking of the contract and the subsequent need to create a new instrument to recognize these arrears.

This approach is also consistent with the principle of recording stocks of financial instruments at market value. In any particular period, the market value of a tradable instrument will change because of, among other things, interest accrued but not paid. If accruals of interest were to be shown as transactions in separate instruments, then there would be odd–looking reconciliation items required to reconcile transactions and changes in stocks. For example, for the tradable instruments themselves, the accrued interest would be reflected as part of the price changes element in the reconciliation statement. This would imply a capital gain on the part of the creditor and a capital loss on the part of the debtor, which is incongruous with the concept of treating the accrual of interest as a transaction. Also, the notional instrument created to record the accrued interest would have no market value, as the market value of accrued interest is inherent in the underlying Instrument. This would require the recording of a capital gain, on the part of the debtor, and a capital loss, on the part of the creditor, to reconcile the transaction with the opening and closing stock positions; both positions with a zero value. This peculiar result has no economic meaning.

3. Calculation of accrued interest

Having established that interest should be recorded on a full accrual basis, and having established that accrued interest should be reflected as a transaction in the underlying financial instrument, the next theoretical Issue for resolution is how accrued Interest should be calculated. Here, for reasons that should become apparent, it is necessary to distinguish between tradable and non-tradable debt.

For non-tradable debt, the terms of the contract between the debtor and the creditor establish the method by which Interest payments should be calculated. Typically, these terms cannot be changed. Therefore, the contractual terms, when expressed on a compound basis, should be used to determine the accrual of Interest, and these terms should be applied to principal outstanding plus any interest accrued but not yet due for payment. For example, if a loan contract requires that interest of 1,000 be paid at the end of each year on a loan of 10,000, then—in order to compile quarterly balance of payments statistics—the annual, nominal, rate of 10 percent should be converted to a quarterly rate of 2.411 percent. ^6/ The table below shows the accrued interest calculation:

		Interest accrued		Interest accrued
	Principle	but not paid (at		during period
	outstanding at	start of period)
	start of period		(A) + (B)	(C) *0.02411
		(B)
	(A)		(c)	(D)
First quarter	10,000	0	10,000	241
Second quarter	10,000	241	10.241	247
Third quarter	10,000	488	10,488	253
Fourth quarter	10,000	741	10,471	259
Year	10,000	0	10,000	1,000

View Table

		Interest accrued		Interest accrued
	Principle	but not paid (at		during period
	outstanding at	start of period)
	start of period		(A) + (B)	(C) *0.02411
		(B)
	(A)		(c)	(D)
First quarter	10,000	0	10,000	241
Second quarter	10,000	241	10.241	247
Third quarter	10,000	488	10,488	253
Fourth quarter	10,000	741	10,471	259
Year	10,000	0	10,000	1,000

The calculations are complicated somewhat if the debtor repays principal or draws down additional funds. In most cases, the contract between the debtor and the creditor require that accrued interest be calculated, using an appropriate compound rate, just prior to any repayment/drawdown, and added to the outstanding balance. Such an accrual of interest is in accordance with the full accrual method for recording interest in the balance of payments.

When non-tradable debt instruments are issued at floating, rather than fixed, rates of interest, then the exact accrual of interest can only be calculated retrospectively. For example, for a loan on which interest is calculated once a year on the basis of a rate prevailing on the date of calculation, the exact interest to be accrued over the year will only be known at the end of the year. However, by referring to relevant prevailing interest rates, an approximate accrual of interest can be calculated for each period, with adjustments made when the interest is finally known.

In theory, for tradable debt instruments, the prevailing rate of interest, rather than some other rate such as the rate of interest when the instrument was issued, should be applied to the market value of an instrument in order to determine the accrued interest. The reasons for this are subsequently discussed; however, it should be recognized that the BPM only explicitly recommends this method for securities subject to secondary trading. ^7/

The first reason why the prevailing rate is appropriate for accruing interest on tradable securities is that this method best matches transactions with stocks. As discussed above, accrued interest is a transaction, not a capital gain/loss. Apart from exchange rate changes, only two factors can lead to capital gains or losses on tradable debt: changes in general interest rates and changes in the credit-worthiness of the debtor. However, if interest were to be accrued using a rate other than the prevailing rate, a reconciliation item that could not be attributable to either of these two affects would be necessary. Such a reconciliation item would have no economic meaning.

The above proposition can be best demonstrated by an example. Suppose at the end of 1992 an Australian resident issued a 3-year zero-coupon bond with a face value of $100 that was acquired by a resident of Germany, and that the appropriate Interest rate for the issuer was 15 percent. (The issue price was $65.75.) In 1993 there was general decline in interest rates and the market reassessed the appropriate rate for the issuer at 10 percent. At the end of 1993, the market value of the bond was, therefore, $82.64. In 1994, there were no changes to either general Interest rates or the credit-worthiness of the borrower.

Clearly, at the end of 1994, the market value of the bond would be $90.91, an increase of 10 percent on the value of a year earlier. However, as there was no capital gains or losses on the bond in 1994, the 10 percent increase in value over 1994 must, therefore, represent a transaction. The only explanation for such a transaction is that it is the offset to accrued interest, which must, therefore, be calculated by applying the prevailing rate to current market values. Calculation of accrued interest using some other method, such as accruing interest using the rate at the time of issue (15 percent in the example), would almost certainly lead to a different result. ^8/

The second reason for calculating accrual of interest using prevailing rates is that this method ensures symmetrical recording by both debtor an creditor. Other methods of calculating accrued interest, such as using the interest rate prevailing at the time a security is issued, are susceptible to asymmetrical recording, particularly when debt is traded. When a security is traded, the new holder of the security is unlikely to know, or care, what the interest rate was at the time the security was issued.

The third reason for preferring the prevailing rate for accruing interest on tradable debt is that it is the prevailing rate that determines both the debtor’s cost of capital and the creditor’s opportunity cost of extending capital. The debtor is willing to accept financing at the cost implied by the prevailing rate. If this cost was too high—for example, if the debtor could obtain cheaper financing or if the return to the project being financed was less than the cost of capital—then the debtor would buy back the debt. The creditor, on the other hand, is willing to provide financing at the opportunity cost implied by the prevailing rate. If this rate was too low–for example, if the creditor could achieve a better rate of return elsewhere—then the creditor would sell the debt instrument. Therefore, as the prevailing rate is the only rate that is agreed to–albeit implicitly—by the debtor and the creditor, it must be the market price. ^9/

The calculation of accrued interest on securities will, when compared to coupon payments, often produce a different result. Nevertheless, it is the accrued value that should be reflected in the income account, not the coupon payments. Instead, coupon payments—like other interest payments—should be treated as a withdrawals of investment by creditors. In any period, coupon payments will either be equal to, less than, or greater than the interest accrued in the period. When coupon payments and accrued interest are equal, it implies that all interest accrued during the period is subsequently paid during that period. When coupon payments are less than accrued interest, it implies that only a part of accrued Interest Is actually paid; the remainder of the accrued interest is reinvested back into the debt instruments. The accrued interest will actually be paid either in the form of a discount, which is payable on redemption, or in future coupon payments. ^10/ When coupon payments are greater than accrued interest, it implies that part of the coupon payment represents either a withdrawal of investment attributable to interest accrued but not paid in an earlier period or to amortization of a premium. In these cases, the interest rates implied by the coupons are greater than the prevailing interest rate and the instruments must be trading at premiums to their face value.

This last result sometimes leads to the labelling of the amortization of premiums as “negative” income. However, this term is not strictly correct. Rather, a premium extended by the creditor to the debtor should be considered as capital additional to the face value. This capital is subsequently repaid through coupon payments that, when as expressed as a percentage of the face value of the bond, are greater than they would otherwise need to be. Therefore, the amortization of premiums should be seen as a downward adjustment to coupon payments for the purpose of calculating interest, offset by a withdrawal of capital.

1. Introduction

Regardless of the source or method used to measure interest income, the results will often not be consistent with the accrual method of recording discussed in the previous part of this paper. This part of the paper reviews the main sources and methods for measuring income, discusses the possible limitations of these sources, and suggests methods for overcoming these limitations.

In many cases, the conceptual requirements of the balance of payments are not satisfied by the source information available to the compiler. Faced with these cases, the compiler can attempt to change the nature of the source in order to better align with conceptual requirements. If this is not possible, the compiler should attempt to adjust source data to a basis that is more consistent with conceptual requirements. How much effort the compiler is willing to put into these adjustments should depend on the significance of discrepancies between the source data and the conceptual model. It is impossible to be categorical about the extent of this adjustment process. In one country, a particular data source may be a good—in terms of conceptual requirements—source of data; in another country a similar data source may be quite poor. Also, circumstances may change over time. The suitability of a particular data source may improve or deteriorate because of evolving conditions.

Therefore, when assessing the sources and methods for measuring interest income, one should not dismiss a particular data source simply because it is unable to meet completely the conceptual requirements of the full accrual method of recording interest income. Instead, one should focus on the extent and significance of any deficiencies, and what steps should be undertaken to minimize the impact of these. Put another way, it is impossible to universally recommend a particular data source for measuring Interest income in the balance of payments; each country needs to make judgments based on the particular circumstances facing that country.

2. International transaction reporting systems

The essential feature of an international transaction reporting system (ITRS) is that it records each balance of payments transaction which passes through a country’s banking system. ^11/ However, this in itself is not sufficient to compile good quality balance of payments statistics and a well-designed ITRS should measure other balance of payments transactions, including those involving domestic enterprises bank accounts with banks abroad.

Because an ITRS essentially records transactions when payments are made, interest income will be recorded in such systems when paid, rather than when accrued. Consequently, in order to measure interest income on a full accrual basis, supplementary information would be required.

However, in many cases, interest is paid in the period in which it is accrued and, in these cases, supplementary Information would not be required. For example, for a country compiling quarterly balance of payments statistics, an ITRS will record, in the correct reporting period, interest–other than interest in the form of discounts–on any instrument that pays interest on a quarterly basis. Furthermore, for instruments that pay interest on a less frequent basis, if there are many such instruments and if it is assumed that interest is paid evenly throughout a year, then while the interest recorded in an ITRS for each particular Instrument will not be on an accrual basis, the total Interest recorded will approximate the full accrual basis. For example, assume that for a particular country interest on loans from nonresidents is typically paid once a year. If It can be assumed that in each quarter, interest is due (and paid) on a quarter of the loans, then for any particular quarter the overstatement of interest recorded in an ITRS for the loans for which interest is paid will offset the lack of recording of interest accrued but not paid on the remainder of the loans. If stocks or interest rates change over time then this approach will cause errors. However, for these errors to be significant then the changes in stocks or rates, in percentage terms, have to be significant. Yield analysis (as described below) could be used to make adjustments in such circumstances.

For securities issued at discount (or at a premium), an ITRS will measure, on a basis that approximates the full accrual basis, interest in the form of discounts for securities that are issued with maturities less than the periodicity of the balance of payments that is being compiled from the ITRS. ^12/ For securities issued with longer maturities, any discrepancies (in aggregate) between discount interest measured on a full accrual basis and discount interest as measured in an ITRS will be relatively minor in cases where discounts, relative to face values, are themselves small. However, in the case of deep discounted and zero coupon bonds, significant discrepancies may arise. In these cases, supplementary information should be sought from creditors and debtors—or their agents—so that appropriate adjustments can be made.

3. Enterprise surveys

Enterprise surveys generally are approaches to enterprises, with information reported a summary, for a particular period, of information recorded in enterprise books. Therefore, to the extent that enterprises measure, in their own books, interest on a full accrual basis, enterprise surveys will provide the balance of payments compiler with data that are of high quality. If, on the other hand, enterprises do not record interest using this method, then the compiler has two choices. Enterprises can be provided with instructions on balance of payments requirements, with the view to their reporting on this basis. Alternatively, the compiler can accept the information provided in enterprise books and, when significant, adjust data to a basis that is consistent with balance of payments requirements. The information for these adjustments could come from supplementary approaches to enterprises.

For non-tradable debt, and for coupon Interest payments on tradable debt, enterprises are most likely to account for interest in a manner consistent with the full accrual method. ^13/ Enterprises that do not account in this manner will probably simply be recording interest on a due for payment basis. In such cases, the enterprises could be asked to recalculate, based on the methodology described in part II.3, Interest on a full accrual basis. Alternatively, the compiler could, in significant cases, make the adjustments. This either would require the compiler to obtain from enterprises supplementary Information on debt terms or accrued Interest could be calculated using yield analysis, as discussed in section 5 below.

For tradable debt, enterprises are likely to account for Interest Income In the form of discounts or premium in one of two ways. An enterprise may amortize a discount or premium using the interest rate prevailing at the time of issue of a security (in the case of debtors) or at the time of acquisition of a security (in the case of creditors)—the amortized discount or premium would be reflected as income in the enterprise’s books and used to adjust the balance sheet value of the asset or liability. Alternatively, enterprises may mark–to–market. Under this approach, at the end of each accounting period the market value of a security is determined, either by referring to the price of a similar security traded in the market place or by using a proxy method, such as net present value. Any gains or losses arising from a change in market value from one period to the next are taken to the profit and loss account as “Income”.

The amortization of discounts and premiums will produce similar results to full accrual accounting (based on prevailing rates) if interest rates remain stable over the life of a security. However, if interest rates rise, interest income will be understated (compared to interest calculated using prevailing rates) and if Interest rates fall, Interest income will be overstated. ^14/ If changes in rates are significant and if the compiler prefers to use prevailing rates, the compiler may wish to make appropriate adjustments. ^15/ Such adjustments could be accomplished by determining the market value of securities and applying the prevailing interest rate. Any coupon payments would need to be deducted to calculate the financial account entry reflecting interest accrued but not paid.

“Income” calculated using mark-to-market methods will consist of two elements: interest income that should be recorded in the balance of payments and—in periods of changing interest rates—unrealized capital gains and losses that should be excluded from the balance of payments. If Interest rates are relatively stable, the compiler could assume that the second component (the unrealized capital gains and losses) is relatively insignificant and can therefore be ignored. On the other hand, when interest rates do vary significantly, the compiler should calculate, or Instruct enterprises to calculate, accrued Interest by applying to the market value of securities (which, in the case of mark-to-market methods, will be the same as book values) the prevailing interest rate. Any coupon payments would need to be deducted to calculate the financial account entry reflecting interest accrued but not paid.

4. Official sources

In many countries, data on interest on official debt (and, in some cases, interest on non-official debt either guaranteed or otherwise monitored by the authorities) are obtained from official sources such as a centralized debt management office. In almost all instances, such records will be maintained on either a cash or due for payment basis. When records are maintained on a cash basis, it will be first necessary to establish the extent of any arrears, as these are accounted for separately in the balance of payments.

In the case of non-tradable debt, Interest due for payment in any period can be converted to a full accrual basis using the method described in part II.3. Compilers either could Instruct the data provider on how to make this calculation or, by obtaining information relating to debt terms, could calculate interest themselves. Alternatively, if the periodicity of interest payments is the same or less than the periodicity of balance of payments statistics, or if there are a large number of loans and interest payments fall due throughout the year, then the compiler may accept information on interest payments (provided there are no arrears) as an acceptable proxy for the full accrual recording of interest.

In the case of tradable debt, if the compiler is confident that he or she can establish the market values of debt outstanding, then Interest could be calculated using yield analysis as discussed in section 5 below. ^16/ On the other hand, if the compiler is not confident of his or her ability to establish market values, then interest could be calculated by adding coupon payments (accrued, when appropriate, over the period to which the payments relate) to amortized discounts. ^17/ (In the case of securities issued with premia, amortized premia would need to be subtracted from coupon payments.) The results obtained from this approach will be a good proxy for the full accrual of interest to the extent that prevailing interest rates are relatively similar to the interest rate at the time of issue.

Alternatively, if discounts (or premia) are relatively small and there is a more or less continual process of issuing and redeeming securities, or if the maturities of securities are relatively short, the compiler could assume that interest in the form of discounts (or premia) is accrued in the period In which the discounts (or premia) are actually paid—that is, when the securities are redeemed.

5. Yield analysis

For a particular period, the yield for any debt instrument can be determined by dividing the interest accrued during the period into the average value of the instrument during the period. From this, it can be seen that interest for any period can be calculated by multiplying the average value of an instrument during a period by an appropriate yield. ^18/ For tradable debt and floating-rate non-tradable debt, an appropriate yield would simply be the interest rate prevailing for the type of debt under consideration. However, for non–tradable debt with fixed interest payments, it would be necessary to use the interest rate at the time of disbursement. This information could be obtained from, for example, a sample of the non-tradable debt agreements between residents and nonresidents.

When yield analysis is used to calculate interest in the balance of payments, information on actual interest payments would also generally be necessary in order to measure financial account transactions attributable to interest accrued but not paid. In other words, the difference between interest as calculated using yield analysis and actual interest payments (assuming that there are no arrears) should be recorded as financial transactions.