THE USE OF GOVERNMENT ACCOUNTS FOR THE IMPLEMENTATION OF THE SGP: LESSONS FOR THE SNA REVIEW. Marco Buti

Transcription

1 Joint meeting of the Advisory Expert Group (AEG) on National Accounts and national accounts users Frankfurt, 9 February 2006 Session 1: The use of national and financial accounts for monetary and fiscal policy analysis THE USE OF GOVERNMENT ACCOUNTS FOR THE IMPLEMENTATION OF THE SGP: LESSONS FOR THE SNA REVIEW Marco Buti 1. The European Commission, in particular the Directorate General Economic and Financial Affairs (or DG ECFIN), is certainly one the largest consumers of macroeconomic statistics in the EU. In our task of monitoring economic developments in the EU, we make use of the full range of macroeconomic statistics compiled by Eurostat and the national statistical institutes. However, when Steven Keuning invited me to come to this meeting between the AEG and national accounts users, I immediately thought that we would have some comparative advantage if this intervention was circumscribed to government accounts. The Commission has a specific responsibility in the implementation of the Stability and Growth Pact (or SGP) and of the excessive deficit procedure (or EDP 1 ). For those less familiar with the institutional framework of the EU, the SGP is a piece of European Law on the conduct of fiscal policies, which is legally binding in the EU Member States. The national governments have a legal obligation of avoiding excessive deficits, of aiming at fiscal positions that ensure sustainability and of reducing their debts, with the aim of avoiding negative externalities among the EU members and conflicts between the centralised monetary policy of the ECB and decentralised fiscal policies The EDP, which is a component of the SGP, is a legally-based administrative procedure which is put in motion whenever any EU government has a deficit which is deemed to be excessive: in practice when the government deficit exceeds 3% of GDP. For the economic foundations and institutional procedure of fiscal surveillance in the EU, see Brunila, Buti and Franco (eds.) (2001) and Buti and Franco (2005) and references therein. 1

2 2. In 1992, when the Maastricht Treaty was signed, the EU Member States decided that the government deficit and debt would be compiled according to the rules of the European System of Accounts (ESA). As far as I know, there is no other example of legally-binding commitments among sovereign countries that refer to numerical indicators, such as deficit and debt, defined according to national accounting rules. The implementation of the SGP has met a number of difficulties, which culminated in the reform of the Pact in The political and economic difficulties are well known, and are well documented in the political and academic literature. 3 There have also been some difficulties of a statistical nature. The statistical difficulties concern the large revisions in data series in a number of countries, suspicions of illegitimate data manipulations, creative accounting and governance issues. 4 In the EU, the political incentives to manipulate the deficit and debt statistics entered into play when Maastricht Treaty was agreed in As a version of the Goodhart law states: when a measure becomes a target, it ceases to be a good measure. However, when countries are closer to the threshold of 3% of GDP as many of them are now, the incentives for data manipulation and exotic transactions are even stronger. 5 Anyhow, irrespective of the European context, one should recognise that ESA and SNA are relatively fragile to the political incentives to embellish the deficit and debt statistics. ESA and SNA were conceived to be tools for economic analysis, but eventually became the cornerstone of legal and political procedures, and are now much more than an instrument for professional economists For a discussion of the political and economic difficulties encountered in the implementation of the SGP and EDP in the EU see the yearly European Commission s report Public Finances in EMU, published in European Economy since The 2003 report contains a specific chapter (II-4) on the governance of budgetary statistics in the EU. The governance issues concern notably the tasks of Eurostat and of the national statistical institutes and the autonomy of the latter vis-à-vis the Governments. These issues were the subject of in-depth discussions within the EU from mid 2004 to end At the end of last year, the EU adopted a regulation, which strengthen and clarifies the Eurostat role as guarantor of the quality of statistics for the SGP, and a European Statistics Code of Practice. Strong evidence of the widespread use of one-off measures and creative accounting in the early years of EMU is found in Koen and van den Noord (2005). 2

3 3. A basic feature of a national accounting system, such as ESA or SNA, is the fact that transactions are recorded in an integrated system, whereby each transaction is recorded symmetrically in the accounts of the two counterparts. 6 National account compilers treat each unit and sector exactly the same way, with the same rules. However, this misses the fact that governments and other economic agents behave quite differently in relation to national accounts. Accounting history is full of bad examples of enterprises that made up their accounts, or that entered in some transactions just to paint their balance sheets in rosier tones. Of course, no enterprise would ever consider changing its behaviour because of national accounts. Yet governments do. The relationship between governments and national accounts, and between other economic agents and national accounts is, therefore, fundamentally different. The way accounting rules are conceived and interpreted, and the way data are compiled, should be differently for government and for other sectors. As users of national accounts, we would see favourably that the convention of conservatism or principle of prudence 7 a cornerstone of commercial accounting which does not exist in integrated systems such as ESA and SNA also became explicitly applicable in national accounts, in particular when compiling government accounts. This prudence could be applicable, in relation to contingent liabilities (e.g. loan guarantees) which for the time being are virtually ignored by national accounting; by treating differently the government expenditure pending of cash payment and revenue pending of effective collection; whenever there are doubts on the appropriate classification of transactions with public enterprises and whenever there are quality problems with basic data. 6 7 ESA and SNA are quadruple-entry systems, rather than the double-entry bookkeeping of commercial accounting. According to IFAC (2005): Prudence is the inclusion of a degree of caution in the exercise of the judgments needed in making the estimates required under conditions of uncertainty, such that assets or revenue are not overstated and liabilities or expenses are not understated. 3

4 4. A generic difficulty we have felt in the implementation of the SGP concerns the reliability of the government accounts and the revision of statistical series. I use here the term reliability to denote how statistical series are revised after their first publication; the more reliable series are those for which the first outcome is closer to final data. 8 In the implementation of the SGP, the fact that the first outcomes for a given year usually available in March of the following year are subject to revision after one, two or more semesters may be a quite significant nuisance. Data revisions have consequences that are quite distinct when data are used by economists for analytical purposes, and when some legal and political procedure is set in motion in case a specific policy target is achieved or missed. In legal and political procedures, the first outcomes are those that really matter and revisions in data often arrive too late to be relevant. Moreover, these revisions may weaken the credibility of the fiscal rules and raise issues of equity among countries. The reliability of data depends mainly on the compilation procedures, on the quality of basic data sources and on the professionalism of data compilers. However, the reliability of accounts depends also on the complexity of the accounting rules. In principle, more sophisticated and more complex rules contribute to improving the accuracy of data. However, there is a trade-off between accuracy and reliability. I use here the term accuracy to denote how statistics are close to the reality, that is, to the event we want effectively to measure. 9 Allow me to provoke you with a concrete example in this respect. The move towards accrual accounting is a major improvement in national accounting. None of us would conceive recording the government wage bill or interest on a purely cash basis, as we did some decades ago. The move towards accrual accounting has been a major advance in improving the economic meaning of national accounts, that is, in improving their accuracy. However, we should also recognise that in some cases e.g. the recording of 8 9 According to Carson and Laliberté (2002), reliability refers to the closeness of the initial estimated value(s) to the subsequent estimated values. Assessing reliability involves comparing estimates over time. In other words, assessing reliability refers to revisions. For Carson and Laliberté (2002) accuracy refers to the closeness between the estimated value and the (unknown) true value that the statistics were intended to measure. 4

5 taxes accrual accounting is considerably more complex than a simpler cash-based accounting, may be less prudent, is often a source of errors and internal inconsistencies, lacks transparency and, because of that, may be even more fragile in relation to illegitimate data manipulations. 10 The complexity of accrual accounting has been behind large revisions in data in several countries. Does it pay to move to accruals in all circumstances? Or should we allow that some transactions are recorded on a simpler cash basis? Is the ambition for higher accuracy always worth the potential loss in reliability? 11 There is an interesting parallel between this trade-off and the ideal quality of fiscal rules as they are presented in the academic literature (well-defined, transparent, simple, flexible, adequacy to final goal, enforceable, consistent, and their link with structural reform). 12 Fiscal rules based on national accounts will not be enforceable if the accounts are not timely and reliable; the transparency of the fiscal numerical rules is directly connected with the transparency of the available data; simple rules (e.g. those that refer to a single numerical indicator with a small or no margin for judgement) will be adequate to the final goal only if the numerical indicator is very accurate and relevant; more flexible fiscal rules which allow a wider margin of judgement will be less dependent on the accuracy of the numerical indicators. In the recent reform of the SGP, the flexibility of the fiscal rule and its adequacy to the final goal have been strengthened for example by giving more importance to cyclicallyadjusted data and longer-term considerations instead of nominal short-term figures, at the cost of transparency and simplicity. One should also note that the EU agreed to postpone the date of transmission of the first outcomes for the government deficit and debt from the end February to the end March, as an attempt to improve the reliability and consistency of the reported data. This move is all the more revealing of the importance of In a recent paper, Tanzi (2006) argues that the [accrual-based] concept that the statisticians have chosen [for measuring fiscal policy], though superior to the cash concept in some ways, suffers from the shortcoming that it does not provide to the policymakers a gauge of the fiscal situation at the time it is needed during and not after, or much after the fiscal year. Franco, Balassone and Zotteri (2004) also argue that the accrual-based government deficit is a misleading compass as it cannot be timely monitored, allows too much room for discretion and is subject to relevant revisions. Of course, we are fully aware that cash data are also subject to manipulation. For the ideal qualities of fiscal rules, see Kopits and Symansky (1998); for the way the EU fiscal rules (before and after the SGP reform) perform in relation to ideal rules see Buti, Eijffinger and Franco (2003 and 2005). 5

6 reliability, as it goes against the tide for improving the timeliness of macroeconomic statistics in the EU. 5. A specific area where we have often perceived difficulties relates to transactions between government and government-owned enterprises. There are at least two aspects to this problem. The first one concerns the delimitation of general government. In ESA, government-controlled units are classified in the corporate sector or in government depending whether sales cover more or less than 50% of costs. This criterion makes sense when the prime aim is the compilation of GDP, when there is a need to decide whether output should be valued according to prices or according to costs. But it is not immediately obvious that this is the ideal criterion when we are interested in fiscal surveillance. A public enterprise whose sales cover only 51% of its costs is not viable without continuous government support. Ultimately, its economic behaviour is not similar to a private firm, but to the government. I believe that most national account users, in particular those who are interested in fiscal surveillance, would appreciate that the rules on the sectoral classification of public enterprises became more stringent than they are now, for example if the 50% criterion became a 100% criterion, or if the distinction between general government and the government-owned enterprises classified in the corporate sector contained some reference to economic viability. A second set of difficulties in relation to public enterprises concerns the accounting of capital injections. The rules in SNA and ESA are, in this respect, clearly unsatisfactory. Eurostat has made important efforts in defining meaningful criteria for distinguishing the capital injections that are no more than disguised subsidisation and those that are eligible to be recorded as financial transactions. 13 However, the recording of these transactions remains the major source of disputes and conflicts between Eurostat and the national statistical institutes and of uncertainty for data users, 14 as the existing rules require a See Eurostat Manual of government deficit and debt, in particular section II.3.1 (Capital injections into public enterprises). For example, in September 2005, when Eurostat published and validated the deficit and debt data reported by Member States (see Eurostat News Release N 120/2005), it mentioned that there was a 6

7 considerable degree of judgement for their application. A solution to this problem could be found in stricter sectoral classification rules or in a principle of prudence to which I alluded to before. Yet, another possibility which to our mind, would better reflect economic behaviour is the recording of reinvested earnings of public enterprises. This consists in recording government revenue (expenditure) at the time public enterprises accumulate profits (losses), irrespective of the effective payment of dividends (payments from governments to offset losses) In our work of fiscal surveillance, we have progressively given more importance to fiscal sustainability and longer-term considerations. The reform of the SGP explicitly recognises the need to integrate implicit liabilities in our analysis, and more specifically in the definition of the medium-term budgetary objective (MTO) of each Member State. To assess sustainability of public finances of the EU member states, we now project pensions to be paid by governments (and other ageing-related spending) until Our work is therefore closely connected with the accounting of pension-related liabilities: one of the more interesting topics in your agenda. In the accounting field, there are several competing options. One possibility is keeping the accounting rule basically unchanged. This is not to say that statisticians should keep closing their eyes to government commitments vis-à-vis pensioners and future pensioners. We certainly hope that statisticians will help us in measuring pension-related commitments as rigorously as possible. However, this could be done through separate measurements, for example in satellite accounts or supplementary tables, without disturbing the core data set. I understand that this is the option that most European statisticians prefer. 15 need to clarify the accounting treatment of capital injections for a number of countries of, in same cases, four years earlier. The recording of reinvested earnings involves several technical difficulties, requires a number of arbitrary conventions (should the reinvested earnings apply to enterprises which are controlled by government or cover also minority stakes?) and could also be somewhat detrimental to the reliability and transparency of accounts. Incidentally, the recognition of reinvested earning would also solve another problem, which is also a source of frequent conflicts in the EU, namely the recording of the socalled super dividends. 7

8 Another possibility is an explicit recognition of the pension commitments as government liabilities. Theoretically, this option would be superior. But, we are afraid that the cost in terms of reliability, transparency and timeliness could be excessively high and make the accounts unfit for the political and legal purposes that we have in the EU. Moreover, we believe that the government commitment vis-à-vis future pensioners is fundamentally different from the private companies obligation vis-à-vis their employees, or the private pension funds vis-à-vis their subscribers. As the distinction between social security and employer schemes is also on your agenda, allow me also to add that, from the economic viewpoint, we do not see any significant difference between the government commitments in relation to the pensions for civil servants and the pensions to be paid to the general population by social security. Whatever the methodological option that will be adopt, the measurement of pension commitments will be a major challenge for statisticians. As I said, we, at the European Commission, in co-operation with the EU Members States and the Ageing Working of the Economic Policy Committee (EPC/AWG), already have some expertise in this area, and would be ready to share it with you. 7. To conclude: I believe that some useful lessons can be drawn from the cooperation between economists and statisticians in the implementation of the SGP. Both statisticians and economists should acknowledge and keep in mind the inevitable political incentives which develop each time a statistical indicator becomes a target or a high-profile reference value. When designing accounting rules as you are doing now as well as when applying them, statisticians should carefully consider the trade-offs among the different dimensions of quality in statistics: accuracy, reliability, timeliness and transparency; and the implications that fiscal statistics have on fiscal rules. Given the use of accounts for political and legal purposes, the statisticians do have an important and difficult task on their shoulders. 8

9 Economists often consider statistics a trivial issue; each figure or series is taken as given without further considerations. 16 The implementation of the Pact has shown that fiscal rules cannot be based on one single simple indicator; economists need to crosscheck the existing data and base their analysis in a widened data set. They should be more involved in the design of the accounting rules, be more interested in the quality of each series and acknowledge how decisive high-quality statistics are indispensable for the success of fiscal rules. Thank you very much for your attention. References: Balassone, Fabrizio, Daniele Franco and Stefania Zotteri (2004), EMU Fiscal Indicators: A Misleading Compass?, mimeo. Brunila, Anne, Marco Buti and Daniele Franco (eds) (2001), The Stability and Growth Pact The Architecture of Fiscal Policy in EMU, Palgrave. Buti, Marco and Daniele Franco (2005), Fiscal Policy in EMU - Theory, Evidence and Institutions, Edward Elgar Buti, Marco, Sylvester Eijffinger and Daniele Franco (2003), Revising the Stability and Growth Pact: Grand Design or Internal Adjustment, CEPR Discussion Paper N (January). Buti, Marco, Sylvester Eijffinger and Daniele Franco (2005), The Stability Pact Pains: a Forward-Looking Assessment of the Reform Debate, CEPR Discussion Paper N (September). Carson, Carol and Lucie Laliberté (2002), Assessing accuracy and reliability: a note based on approaches used in national accounts and balance of payment statistics, IMF Working Paper 02/24. European Commission, Public Finances in EMU, yearly report published in European Economy since IFAC (2005), Handbook of International Public Sector Accounting Standards. Koen, Vincent and Paul van den Noord (2005), Fiscal Gimmickry in Europe: One-off measures and creative accounting, OECD Economics Department Working Papers, As Tanzi (2006) says, when economists, or even fiscal economists, write papers about fiscal policy and discuss the fiscal deficit of a country, they rarely bother to specify which concept they are using. They assume that this issue is trivial. 9

10 Kopits, G. and S. Symansky (1998), Fiscal Policy Rules, IMF Occasional Paper,162. Tanzi, Vito (2006), Fiscal policy: when theory collides with reality, mimeo. 10