(p.199) Appendix: Other Eurostat Stage II Decisions, 1995–7
(p.199) Appendix: Other Eurostat Stage II Decisions, 1995–7
A.1 Three Categories of Lesser Rulings
In addition to their major national accounts rulings, Eurostat and the Committee for Monetary, Financial, and Balance of Payment Statistics (CMFB) considered a number of lesser cases that also influenced the deficit and debt calculation of the European Union (EU) member states that directly influenced the Stage II convergence process. Many of these lesser rulings stemmed from questions raised by the National Statistical Institutes (NSIs) themselves during the Eurostat mission visits of 1996 and 1997, and both the more prominent and the less well-known decisions may be divided into three categories. First, there are decisions related to the classification of transactions of government, where they are regarded as financial transactions with no effect on the deficit or debt calculation or nonfinancial transactions that do influence the calculation. The Belgium privatization, France Telecom, the various gold cases, and central bank payments to the state, as well as those indicated in the table, fall under this category. Perhaps most significant of these lesser decisions concerned state guarantees on public enterprise debt, which lowered Italy's deficit by 0.2 of gross domestic product (GDP) for the critical 1997 fiscal year. Another Italian case that attracted significant public attention concerned currency and interest rate swaps, and is described here in some detail. Second, there are decisions related to the time of recording transactions, as in the cases of Ireland and retroactive accounting, the treatment of interest, and the timing of payment dates for taxes, including the eurotax. Last, there are decisions related to the classification of units inside or outside of general government, as, for instance the German hospitals case, where a distinction must be made between those activities falling within the governmental sphere, which therefore affect the deficit calculation.
(p.200) A.1.1 Decisions Related to the Classification of Transactions
1. Interest Rate and Currency Swaps: In the case of interest rate swaps, only the net payments of interest between the two parties to the swap should be recorded. The influence on a government deficit would be limited to the difference between the interest flows agreed to in the exchange. In the case of currency swaps, any outside foreign currency debt will be valued according to the market exchange rate and not the exchange rate in the swap contract. The existence of a swap agreement has no effect on the valuation of outstanding debt in a foreign currency. This interpretation raised Ireland's 1997 deficit by. 4 percent and lowered its 1997 debt by 1 percent of GDP; it increased Finland's 1997 deficit by 0.1 percent, and lowered Italy's deficit by 0.02 percent in 1997.
2. The Italian Interest Rate and Currency Swaps Case: In November 2001, the Financial Times reported on allegations made by the International Securities Market Association that Italy improperly counted the gains from an exchange and interest rate swaps deal as revenues for calculating its 1997 deficit and debt figures.1 In such a swap, a government may issue a security in a foreign currency as a way to manage the public debt and the government's foreign exchange reserves, and later elect to swap that security for its own national currency in order to realize the gains from trade. In 1995, the Italian government issued a three-year bond for 200 billion yen, with the yen valued at 19.3 lira. Although the bond matured in September 1998, these liabilities would be counted at the end of 1997 and applied towards that year's deficit and debt because, as Article 1 of Council Regulation 3605/93 requires, “liabilities denominated in foreign currencies shall be converted into the national currency at the representative market exchange rate prevailing on the last working day of each year.” Consequently, any unexpected appreciation in the yen would contribute to an expansion in Italy's deficit and debt above its 1996 level, as the yen were converted into lira. Rather than risk this possibility, the Italians incorporated into the terms of these securities anticipated rates of exchange and interest over the lifetime of the bond.
Critics of this financial structuring claimed it to be “creative accounting,” no better than that perpetrated in the Enron scandal in the United States.
Although Eurostat was not mentioned by name, the author inferentially referred to the agency by asking, “who shall guard the guardians?”2
The contract required Italy to make negative interest payments to the bank of Lira Libor minus an astounding 1,677 basis points in 1997 (meaning that Italy received funds) and then in effect to reverse the payments in September 1998. This reduced Italy's official deficit in 1997 only by raising it in 1998. … The damning part of this explanation is (p.201) the admission that Italy was taking a cash advance against an expected foreign exchange profit in 1998. … It is clear that the existence of the yen bond was merely a convenient pretext for Italy to borrow money that it could then misclassify as a hedge, just as Enron's commodities trades were a pretext for its own disguised borrowing.
The problem for Eurostat in evaluating this aspect of Italy's financial activities was that ESA 79, the outdated national accounting standard set by the Maastricht Treaty, made no mention of swaps transactions, thus making the transaction fully allowable. Moreover, the replacement set of rules, ESA 95, did consider swaps, but had yet to be formally adopted. Eurostat subsequently consulted with the CMFB, whose advice led Eurostat to conclude that nothing had been specified in either set of national accounting rules that prohibited the Italian Treasury from structuring the bond as it did. In a manner explicitly consistent with Regulation 3065/93, the valuation of the liability had to converted into the domestic currency, with its effect on the deficit calculation limited to the difference between the interest flows specified in the bond. In any event, the effect of the transaction improved Italy's deficit by 0.02 percent, not an amount that would influence Italy's entry into the EMU.3
3. Financial Leasing: All leasing transactions shall be treated as operating leasing. So, if a government sells fixed assets and then leases these assets with the intent of reacquiring them, this constitutes an operating lease. The sale of these fixed assets constitutes the direct sale of nonfinancial assets, and receipts from this sale may be applied to the government's deficit calculation. The obligation to repurchase these assets at the end of the lease is a contingent liability and is not included in the government's debt calculation.
4. Financing and Exploiting of Public Infrastructures by the Enterprise Sector: Three cases fell under this category. In the first case associated with the United Kingdom, an enterprise constructs the infrastructure and then operates it while receiving annual payments from the state for services produced with the infra-structure. After the agreed upon period expires, the infrastructure is turned over to the state. Supported unanimously by the CMFB, Eurostat ruled that only the annual payments shall be counted towards the deficit calculation. The actual acquisition of the infrastructure has no effect on the government's debt calculation. The second case applies particularly to the construction and prefinancing of at least twelve roads in Germany, which in 1997 total some 4–5 billion DM. In these instances, the enterprise is required to build and prefinance a fixed asset (p.202) for the state, where the state becomes the owner of the asset as it is constructed. This investment in gross fixed capital formation should be recorded in the general government sector S60, and included in the public deficit calculation, but excluded from the debt calculation. The third case addresses similar situations occurring in the United Kingdom, Denmark, and Sweden, where an enterprise is required to build a fixed asset, operate it, and retain ownership of the asset. In the case of Denmark and Sweden, for example, a bridge was constructed in 1996 by a public corporation where the finances were backed by state guarantees. In each case, gross fixed capital formation is recorded in the enterprise sector with no effect on the public deficit.
5. State Guarantees on Public Enterprise Debt: Supported by a large majority vote of the CMFB, Eurostat ruled that public enterprise debt guaranteed by the state shall be included in the calculation of the government's debt when, first, the law authorizing the issuance of this debt indicates the government's repayment obligation, and, second, when the government's budget specifies the annual amount of repayment. This rule applies to the case of the Italian railroad, the Ferrovie dello Stato, whose debt issues since 1981 have been authorized by law by the Ministry of Finance. Consequently, the government's debt was reclassified to reflect an increase for each year the railroad's debt increased, while the government's deficit was reduced to reflect the amortization of old debt. For the period 1993–7, this reclassification increased Italy's collective debt by 13.51 billion lire, and reduced its collective deficits by 2.4 billion lire. For the 1997 fiscal year specifically, there was no effect on the debt and an improvement in the deficit of 3.687 billion lire, or 0.2 percent of GDP.
6. Export Insurance Guaranteed by the State: In each EU member state there are export insurance enterprises that benefit from state guarantees, which in ESA are classified in either the insurance sector S50 or general government sector S60. All flows from these firms, premiums and indemnities, will be recorded as nonfinancial transactions. Consequently, losses incurred by the state on guaranteed payments will also be recorded as nonfinancial transactions, and must be included in the calculation of the public deficit. This ruling reduced Spain's 1997 deficit by 0.1 percent, and increased Austria's deficit by 0.02 percent of GDP.
A.1.2 Decisions Related to the Time Recording of Transactions
1. Capitalized Interest on Deposits and Other Financial Instruments Covered by ESA79: In the case of deposits or similar financial instruments that act as government liabilities, the capitalized interest will recorded as government expenditure, which contributes to the budget deficit, when interest is paid to holders of these instruments. Interest will be recorded separately from the (p.203) principal, and will be recorded when the capitalized amount falls due for payment, rather than distribute it to different periods. This ruling increased Portugal's 1997 deficit by 0.15 percent and decreased Italy's 1997 deficit by 0.26 percent of GDP.
2. Treatment of Fungible Bonds Issued in Several Tranches: In the case of fungible bonds, that is, bonds that are issued in tranches at different points in time without change in the coupon payment date, the accrued coupon will be recorded as a short-term liability under the heading “accounts receivable and payable,” which therefore does not enter into the calculation of government debt. This reduced the French 1996 deficit by 0.18 percent of GDP.
3. Treatment of Interest for Short-, Medium-, and Long-Term Bonds: To harmonize bills and short-term bonds across the EU member states, they are defined as those with a maturity up to and including twelve months. For these short-term bonds, the difference between the issue price and the nominal value will be regarded as interest recorded at the issuance of the bonds. This difference will influence the calculation of a government's deficit. For medium- and long-term bonds, the difference between the issue price and nominal price will not be regarded as interest, but as holding gains and losses. This difference will not affect the calculation of a government's deficits. This ruling increased Finland's 1997 deficit by 0.04 percent of GDP.
4. Treatment of Interest for Linear Bonds: Like many other unconventional bonds, linear bonds are not addressed in ESA 79. Consequently, Eurostat ruled on how to classify the difference between the nominal value and price at issue for linear bonds, those fungible bonds issued in several tranches from the same lineage with the same nominal interest rate as well as identical dates for coupon payments. All the difference between the nominal value and issue price must be considered as interest recorded at the time of the coupon payments, such that this difference must be included in the calculation of the public deficit. Linear bonds are used particularly in Belgium, Denmark, Finland, Portugal, and Sweden. As a result of this interpretation, Belgium's deficit for 1997 was increased by 0.04 percent, Denmark's by 0.02 percent, Finland's by 0.17 percent, Portugal's by 0.05 percent, while Sweden's deficit was reduced by 0.28 percent.
5. Treatment of Interest In the Case of Zero Coupon Bonds: The difference between the issue price and redemption price of a zero coupon bond is to be treated as interest, to be recorded as interest paid at the bond's maturity. Interest is defined as a payment at predetermined dates of a fixed percentage of the nominal value of the asset, so the price difference is interest. This interpretation raised Denmark's 1997 deficit by 0.01 percent of GDP.
(p.204) 6. Changes in the Due for Payment Dates for Taxes, Salaries, Social Contributions, and Benefits: Transactions can be recorded at three different points in time: When amounts are cashed, as in a cash payment; when amounts are due, as in due for payment date; and when activities giving rise to transactions occur, on an accrual basis. In 1996, Sweden reduced by one month the time-lag for its value-added tax (VAT) payments granted to large enterprises. This measure led in the Swedish public accounts to cashed amounts corresponding to thirteen months of VAT receipts in 1996. During the same year under the same set of circumstances, Portugal reduced the time-lag for VAT payments by ten days. Both cases raised the matter of whether the government's deficit position is improved by this additional period of VAT payments. In the same context, Eurostat considered the national accounts effect of the time of certain government expenses, particularly public employee salaries and social benefits, and the imposition of certain taxes and social contributions. In making its ruling, supported by the near unanimous opinion of the CMFB, Eurostat distinguished between taxes tied to production and imports, and those taxes on income, wealth, salaries, social contributions, and benefits. In the first case, ESA 79 specifies that they be recorded on an accrual basis, in other words at the time goods and services are produced, sold, or imported. Consequently, any change in the payment due date of taxes in this case, though affecting the cash balances in public budgets, should not be included in the calculation of taxes recorded in national accounts and the determination of the government deficit. This situation applies to Sweden and Portugal, which permitted Sweden to reclassify its national accounts to show an increase of 0.6 percent of its deficit for 1996, and by 0.36 for Portugal during the same year, as these VAT taxes could not be included to reduce the government's deficit. In the second case, of taxes on salaries, etc. these should be recorded at the payment due date established by law, and should be included in the deficit calculation. Temporary administrative adjustments in tax payment dates, though improving the government's cash balances, should not affect the government's deficit position.
A.1.3 Decisions Related to the Classification of Units
1. Classification of National Bodies Acting on Behalf of the European Commission: In the case of those national institutional units that are engaged in market regulation and the distribution of subsidies, such as the Irish Intervention Agency, if their activities cannot be divided along these lines, then these units will be classified in the sector General Government, which will influence the government's deficit and debt calculation, when their costs incurred in (p.205) market regulation compared to total costs are less than 80 percent. This decision increased Ireland's 1997 debt by 1 percent of GDP.
2. Pension Funds: Pension funds that finance benefits primarily on a pay-as-you-go basis and to a minor extent on a capital funding basis have to be classified in the subsector S63 “Social Security Funds” of General Government sector S60. The classification criteria are that these funds are institutional units, as they have a complete set of accounts, have autonomy of decision, and pay benefits to the insured without reference to the individual exposure of risk, where these employment-based pension schemes are built on a collective financial balance principle.
Sources: Eurostat news releases and CMFB minutes.
(1) James Blitz (2001). “Italy Rebuts Swap Contract Claims,” Financial Times, November 6, p. 6.
(2) Benn Steil (2002). “Enron and Italy,” Financial Times. July 21, p. 13.
(3) Peter Norman (2001). “Rome ‘Did Not Cheat Over Deficit,’” Financial Times. November 6, p. 6.