Effects of the European Economic and Monetary Union (EMU) on Taxation and Interest Spending of National Governments

Francesco P Mongelli

doi:10.5089/9781451951271.001.A001

Jump to Content Jump to Main Navigation

Effects of the European Economic and Monetary Union (EMU) on Taxation and Interest Spending of National Governments

Author: Mr. Francesco P Mongelli

Contributor Notes

Author’s E-Mail Address: fmongelli@imf.org

Publication Date:: 01 Jul 1997

eISBN:: 9781451951271

Language:: English

Keywords:: WP; financial market; risk premium; discount rate; interest rate differential; single currency; devaluation risk premium; reference interest rate; case interest rate differential; medium-interest rate

Download PDF (1.7 MB)

This paper examines the interest spending and taxation channels through which EMU could affect the public finances. It provides a framework for examining different views on a further narrowing of interest rate differentials. A model of Blanchard and Fischer is amended to analyze the two channels, and empirical evidence on the tax harmonization process is presented. The paper argues that “high-debt” and “high-tax” countries pursuing prudent fiscal policies could benefit the most from EMU: if monetary and widespread fiscal discipline are jointly established, interest rates could decline rapidly, while tax harmonization is likely to be gradual.

Abstract

I. Introduction

The current members of the European Union (EU) are becoming increasingly integrated from an economic and financial standpoint. In the coming years, some EU members may also establish the European Economic and Monetary Union (EMU) and adopt a common single currency, the “euro.” While integration was deepening, the fiscal performance of most EU members gradually deteriorated leading to higher, albeit uneven, public indebtedness. How could the EMU affect the public finances of the participating countries?

The above question is very broad and this paper focusses on two channels, interest spending and taxation. ¹ Specifically, the contribution of the paper is in examining the possible balance between the interest spending and taxation channels in the wake of the EMU. A framework to examine different views concerning a further narrowing of interest rate differentials is provided. In addition, a model of Blanchard and Fischer (1989) is amended here to analyze the two channels. Finally, empirical evidence on the EU process of tax harmonization is presented and the main forces behind this ongoing process of harmonization are examined. Some potential effects of the EMU on the ability to raise additional taxes are also discussed.

In recent years, there have been significant improvements in the government finances of several EU members despite the economic slowdown in 1995-96. Structural deficits have declined by almost 3 percent of GDP during 1991-96, with a further percentage point or so projected for 1997. There is also a decline, or incipient decline, in debt ratios in high debt countries. If the current economic recovery stays on track and fiscal consolidation continues, the EU members will find it easier to reduce unemployment compensation and other social expenditures. Interest rates and interest expenditures should continue declining. The need for further increases in tax rates will fall, and debt ratios may decline. On the other hand, extended uncertainty about the start of the EMU and EU members’ participation may reduce market confidence, weaken the economic recovery, and slow down any further fiscal adjustment. As a result, the EU financial markets may be subjected to some risks of turbulence and higher interest rates. Tension may arise in countries that have more difficulty raising taxes due to political reasons or due to the constraints inherent in the deepening of economic and financial integration. Hence, there are upward potentials and downside risks as EU members establish the EMU.

There is largely a consensus that interest rate differentials are likely to narrow, with respect to the current levels, in the wake of the EMU. There is less consensus, though, on the possible extent of such a decline in each prospective member, i.e., interest rate differentials may not completely vanish in the wake of the EMU. Section II provides a framework to examine different views on this issue. We start by looking at the current interest rate differentials of each EU member with Germany and at their credit rating. Next we discuss the factors underlying interest rate differentials, i.e., the inflation/devaluation and the outright default risk premia. Last we ask what may be the likely trade-off between the vanishing inflation/devaluation risk premium and the outright default risk premium. We illustrate how the response to this question--and the extent of further convergence in interest rate--hinges on the combined strength of monetary and fiscal discipline. Appendix I discusses some of the main conditions for a decline in the EMU level of interest rates benefitting each participating country.

Section III provides an analytical framework to examine the interest spending and the taxation channels. This framework is based on a model that is a specialization of Blanchard and Fischer’s (1989) model. Their model is amended here in two ways. First, the model used here differentiates between a generalized decline in the level of interest rates that will benefit each EMU member, and the potential decline in the country-specific risk premium that will benefit a high debt-to-GDP country undertaking a fiscal consolidation. Second, the level of taxation is postulated to be positively related, but at declining rates, to the level of public indebtedness. Hence, the model establishes an analytical bridge between the two channels.

Section IV presents selected indicators of tax harmonization across the EU and discusses the main forces behind the ongoing process of harmonization of EU tax systems. This section also examines the prospects for a deepening of tax harmonization in several, but not all, areas of taxation, and the potential effects of the EMU on the ability to raise additional taxes. A relevant question for this paper is: under what circumstances could tax harmonization bite the most? A country with a relatively high level of taxation could be more affected than the others particularly if a shock lessens its chances of participating in the EMU, or if a shock increases the risk that EMU itself would be postponed for a significant period. Section V, ties together the previous analysis and presents three scenarios. Section VI provides the paper’s concluding remarks.

The new EU members--Austria, Finland, and Sweden--are excluded from the analysis in this paper, because, having joined at a late stage some of the statistical relationships presented in the paper may be less defined. The discussion of the interest spending and taxation channels is not meant to provide a specific forecast of narrowing of interest differentials or further tax harmonization across EU members. Instead, the discussion describes the possible balance of the influences of the two channels. The list of notation is in Appendix II.

II. The Likely Decline and Convergence in Interest Rates

1. The factors underlying interest rate differentials

In recent years, the gradual, albeit uneven, fiscal adjustment--in most, but not all, EU members--and the continuing decline in inflation have fostered a general, but erratic, decline in nominal interest rates and interest rate differentials with Germany (Table 1 and Figure 1). Despite this progress, there are still broad differences in fiscal standing and international financial markets still discriminate among EU sovereign borrowers. In particular, nominal interest rate differential with Germany, and the credit rating on national public debt set by Standard and Poor’s and Moody’s permit to classify EU members as follows:

“low-interest rate” countries. This group exhibits a very low nominal interest rate differential with Germany (currently within 30 basis points) and a high credit rating. In addition to Germany, this group includes Belgium, France, Luxembourg, the Netherlands, and the United Kingdom. The last country has a relatively low interest rate in the perspective of the past few years (1993-96) and a credit rating equal to that of Germany;
“medium-interest rate” countries. This group exhibits a modest interest rate differential with Germany (currently between 50 and 100 basis points) and in some cases a lower credit rating than the countries in the first group. It includes Denmark, Ireland, and Spain; and
“high-interest rate” countries. This group exhibits more significant nominal interest rate differentials with Germany and lower credit ratings particularly on foreign currency denominated debt. It includes Greece, Italy, and Portugal. The last country has a lower credit rating on its foreign-currency debt than Italy.

Table 1.

Selected Financial and Budgetary Indicators, and Current and Projected Credit Ratings, 1993–97. ^1/

Source: National Sources; IMF International Financial Statistics and World Economic Outlook; and Standard and Poor’s and Moody’s.

^{^1/}1997 data are April 1997 IMF World Economic Outlook Projections.

^{^2/}Ten–year government bond yield or nearest maturity. See ^{footnote 1} Figure 2 for description of interest rates.

^{^3/}End 1996-beginning 1997 credit rating by Standard and Poor’s and Moody’s respectively. The last column reflects the projected unified rating under the hypothesis of a participation in the EMU. nr denotes no rating

^{^4/}Weighted by national GDP.

Table 1.

Selected Financial and Budgetary Indicators, and Current and Projected Credit Ratings, 1993–97. ^1/

	Consumer Price Inflation		Budget Deficit /GDP		Public Debt /GDP		Long Term Nominal Interest Rates ^2/		Credit Rating on Public Debt
	93–96	97	93–96	97	93–96	97	93–96	Jan. 97	Foreign Currency S & P’s/Moody’s ^3/	Domestic Currency S & P’s/Moody’s ^3/	Euros after EMU start S & P’s/Moody’s ^3/
Belgium	2.4	1.8	−4.9	−2.9	134.0	127.1	7.3	5.9	AA+/Aa1	AAA/Aa1	AA+/Aa1
Denmark	1.8	2.5	−3.0	0.1	74.5	67.1	7.0	6.5	AA+/Aa1	AAA/Aa1	AA+/Aaa
France	2.0	1.5	−5.2	−3.3	50.9	57.5	7.1	5.8	AAA/Aaa	AAA/Aaa	AAA/Aaa
Germany	2.5	1.7	−3.4	−3.3	54.3	61.7	7.0	5.8	AAA/Aaa	AAA/Aaa	AAA/Aaa
Greece	10.6	6.9	−10.6	−5.1	114.9	107.7	15.4	10.9	BBB-/Baa1	nr/A2	BBB-/A2
Ireland	2.1	2.2	−2.0	−1.6	85.5	72.3	7.9	6.3	AA/Aa1	AAA/Aaa	AA/Aaa
Italy	4.6	2.6	−8.3	−3.3	123.8	123.2	10.7	7.3	AA/Aa3	AAA/Aa3	AA/Aa3
Luxembourg	2.4	2.0	0.1	−0.1	5.7	6.8	6.1	5.7	AAA/Aaa	AAA/nr	AAA/Aaa
Netherlands	2.3	2.6	−3.1	−2.2	78.9	75.7	6.6	6.1	AAA/Aaa	AAA/Aaa	AAA/Aaa
Portugal	4.7	2.7	−3.1	−2.9	70.3	69.6	10.2	6.8	AA-/Aa3	AAA/Aa2	AA-/Aa2
Spain	4.4	2.8	−6.3	−3.2	64.2	67.4	10.1	6.8	AA/Aa2	AAA/Aa2	AA/Aa2
United Kingdom	2.5	2.6	−6.0	−3.1	45.7	49.4	7.8	7.6	AAA/Aaa	AAA/Aaa	AAA/Aaa
EU Average	3.5	2.7	−4.6	−2.6	75.2	73.8	8.6	6.8
EU Weighted Average ^4/	3.0	2.2	−5.5	−3.1	71.6	74.7	8.2	6.5

Source: National Sources; IMF International Financial Statistics and World Economic Outlook; and Standard and Poor’s and Moody’s.

^{^1/}1997 data are April 1997 IMF World Economic Outlook Projections.

^{^2/}Ten–year government bond yield or nearest maturity. See ^{footnote 1} Figure 2 for description of interest rates.

^{^4/}Weighted by national GDP.

Table 1.

Selected Financial and Budgetary Indicators, and Current and Projected Credit Ratings, 1993–97. ^1/

	Consumer Price Inflation		Budget Deficit /GDP		Public Debt /GDP		Long Term Nominal Interest Rates ^2/		Credit Rating on Public Debt
	93–96	97	93–96	97	93–96	97	93–96	Jan. 97	Foreign Currency S & P’s/Moody’s ^3/	Domestic Currency S & P’s/Moody’s ^3/	Euros after EMU start S & P’s/Moody’s ^3/
Belgium	2.4	1.8	−4.9	−2.9	134.0	127.1	7.3	5.9	AA+/Aa1	AAA/Aa1	AA+/Aa1
Denmark	1.8	2.5	−3.0	0.1	74.5	67.1	7.0	6.5	AA+/Aa1	AAA/Aa1	AA+/Aaa
France	2.0	1.5	−5.2	−3.3	50.9	57.5	7.1	5.8	AAA/Aaa	AAA/Aaa	AAA/Aaa
Germany	2.5	1.7	−3.4	−3.3	54.3	61.7	7.0	5.8	AAA/Aaa	AAA/Aaa	AAA/Aaa
Greece	10.6	6.9	−10.6	−5.1	114.9	107.7	15.4	10.9	BBB-/Baa1	nr/A2	BBB-/A2
Ireland	2.1	2.2	−2.0	−1.6	85.5	72.3	7.9	6.3	AA/Aa1	AAA/Aaa	AA/Aaa
Italy	4.6	2.6	−8.3	−3.3	123.8	123.2	10.7	7.3	AA/Aa3	AAA/Aa3	AA/Aa3
Luxembourg	2.4	2.0	0.1	−0.1	5.7	6.8	6.1	5.7	AAA/Aaa	AAA/nr	AAA/Aaa
Netherlands	2.3	2.6	−3.1	−2.2	78.9	75.7	6.6	6.1	AAA/Aaa	AAA/Aaa	AAA/Aaa
Portugal	4.7	2.7	−3.1	−2.9	70.3	69.6	10.2	6.8	AA-/Aa3	AAA/Aa2	AA-/Aa2
Spain	4.4	2.8	−6.3	−3.2	64.2	67.4	10.1	6.8	AA/Aa2	AAA/Aa2	AA/Aa2
United Kingdom	2.5	2.6	−6.0	−3.1	45.7	49.4	7.8	7.6	AAA/Aaa	AAA/Aaa	AAA/Aaa
EU Average	3.5	2.7	−4.6	−2.6	75.2	73.8	8.6	6.8
EU Weighted Average ^4/	3.0	2.2	−5.5	−3.1	71.6	74.7	8.2	6.5

Source: National Sources; IMF International Financial Statistics and World Economic Outlook; and Standard and Poor’s and Moody’s.

^{^1/}1997 data are April 1997 IMF World Economic Outlook Projections.

^{^2/}Ten–year government bond yield or nearest maturity. See ^{footnote 1} Figure 2 for description of interest rates.

^{^4/}Weighted by national GDP.

Before discussing the prospect for a further narrowing of interest rate differentials, it is necessary to gauge the extent to which the financial market penalizes relatively higher debt and current or expected inflation/devaluation. Alesina, De Broek, Prati and Tabellini (1992) have found that in highly indebted OECD countries the differential between domestic public and private nominal rates of return is positively related to the stock of debt. However, the magnitude of the outright default risk premium, although statistically significant, is quantitatively small. Masson and Symansky (1995) estimate default risk premia from the long term interest rates prevailing on the EU members’ foreign currency debt. In their procedure, German long term bonds are used as the reference “risk free” bonds for the calculation of the interest differentials. These differentials provide estimates of country-specific outright default risk premia. The premia are relatively small and never exceed 40-50 basis points even in highly indebted members.

Favero, Giavazzi, and Spaventa (1996) focus on total interest rate differentials on government bonds between high yield countries and Germany. In the long-run they find evidence of cointegration between total interest rate differentials and exchange rate factors and of uni-directional causality going from the exchange rate factors to the total yield differential. Hence, total interest rate differentials are largely determined by exchange rate factors. Furthermore, membership in the ERM was not sufficient to stabilize exchange rate expectations. Last, the default risk premium (that is calculated as a residual) is low, highly variable, consistently positive, and increases with maturity.

The above studies, as well as Cottarelli and Mecagni (1990) and Giovannini and Piga (1992) concur that current EU outright default risk premia, although somewhat elusive, are small: hence, the interest rate differentials in Table 1 reflect for the most part differences in current inflation/devaluation risk premia. This finding has important implications because intra-EMU exchange rates will be irrevocably fixed and the European Central Bank (ECB) is mandated to maintain inflation low and stable. As a result, differences in inflation/devaluation risk premia across EMU members can be expected to vanish. However, differences in fiscal standings will likely not vanish (Section IV). Hence, could higher outright default risk premia replace the vanishing inflation/devaluation risk premia in the wake of the EMU?

2. The critical factors to establish monetary and fiscal discipline

The answer to the above questions is not simple. The EMU will in fact be “guided” by the European Central Bank (ECB) that will set the common monetary policy; the national governments of the participating countries that will maintain a considerable degree of fiscal autonomy; and to a certain extent also the bodies of the EU that manage very limited but not insignificant resources. Coordinating these “actors” may be complex because their goals and instruments may overlap and even conflict at times. This paper does not attempt to examine the implications of coordinating these actors. Instead, it focusses on how monetary and fiscal discipline can jointly reduce the likelihood of any explicit or implicit bailout--or “financial solidarity” as some authors prefer to call it (where the two terms are used interchangeably here below)--and foster lower interest rates.

Establishing monetary discipline and pursuing low and stable inflation will be the primary responsibility of the European Central Bank (Figure 2). The Bank will also independently manage the instruments to conduct monetary policy and the common foreign exchange reserves of the EMU. Any form of financial solidarity, or bailout, will be barred (Art. 105 of the Maastricht Treaty).² Two dilemmas are likely to arise in the operation of the European Central Bank. The first dilemma is that some flexibility in monetary policy may be desirable from time to time to offset the adverse effects of tight budgetary policies or a downturn in the economy. To what extent does the Bank need to stick to “unwavering” tight monetary precepts--even in the face of increasing unemployment across Europe--to ensure monetary discipline? The second dilemma is that the resources of the Bank could be used to rescue a member in financial distress through some form of discreet or open financial solidarity, or bailout.³ Financial solidarity could be dictated, for example, by the desire to avoid any systemic ill-effect--such as a potential disruption of the financial market or trade--and could be interpreted as a form of risk sharing. Any actual or expected loosening of the monetary stance or any form of implicit or explicit financial solidarity, or bailout--may hamper financial market-based discipline and even reduce the incentive of national governments to pursue prudent fiscal policies. Hence, monetary discipline will be established when there will be no doubts concerning the Bank’s commitment to pursue low and stable inflation and both dilemmas are resolved. In this case the Bank will be able to loosen the monetary stance in the wake of a recession, or alter its portfolio, without jeopardizing the public’s perception about the Bank’s commitment to low inflation.

Fiscal discipline will be crucial in the EMU to allow national governments to pursue counter-cyclical fiscal policies and reduce the negative spill overs of high debt.⁴ When fiscal discipline is low there are three main remedies. The first remedy is provided by financial market-based discipline that would leave each country largely in control of its fiscal policy: i.e., market forces would exert a restraining role on national governments by rendering continuous and excessive borrowing progressively more expensive.⁵ The second remedy calls for rigid fiscal requirements and/or budget rules that reduce the room for national governments’ arbitrariness.⁶ The third remedy entails a system of mutual budgetary surveillance that could allow better consideration of national circumstances.⁷ The rest of the paper assumes that fiscal discipline will rely on a combination of all three remedies: i.e., a strengthening of financial market-based discipline; some common fiscal requirements; and a “stability and growth pact” that will institutionalize the fiscal criteria of the Maastricht Treaty, strengthen mutual surveillance, and enforce automatic fines against excessive budget deficits.

3. The effects of monetary and fiscal discipline, or the lack thereof, on interests

We now tie together the previous discussion. For simplicity and clarity, we focus on three illustrative cases. Mongelli (1996c) discusses more cases in greater detail.⁸ In Case A, tight monetary discipline complements and ensures tight fiscal discipline. Accordingly, the task of both the European Central Bank and the national governments should be eased because countries would not be allowed to run sizable and continuing budget deficits or to accumulate public debts in excess of the Maastricht debt criteria. The decline and convergence in interest rate is likely to be significant in every EMU member, albeit by a different extent. In particular, who gains the most from the standpoint of interest rates? The medium- and high-debt-to-GDP countries with relatively higher interest rates and a track record of relatively higher inflation may benefit the most by borrowing from the anti-inflationary reputation of the European Central Bank. In addition, all EMU members are likely to benefit from a general decline in interest rate levels along the lines discussed in Appendix I.

In Case B, we relax for a moment the assumption of fiscal discipline but there is tight monetary discipline. Specifically, one (and only one) member country is “caught” circumventing the fiscal regulations and running continuous and excessive budget deficits. Investors will hold the liabilities of the “free-rider” only until its risk-adjusted rate of return is comparable with those of other similar assets issued by other countries. The risk (and the costs) of a partial or complete outright default are borne entirely by the bondholders and the taxpayers of the free-rider. As a result, this country’s specific interest rate will rise, but the reference interest rate of the EMU (e.g., the German interest rate) should be relatively unaffected as long as this undisciplined member is not expected to be bailed out by the European Central Bank. Therefore, even in this adverse scenario the fall in interest rate can be significant in fiscally disciplined countries while any undisciplined member will be penalized by the financial market. In this case interest rate differentials may not disappear and could even widen depending on the extent of the fiscal unbalance, the strength of financial market-based discipline, and the loss in fiscal autonomy.

In Case C, one country circumvents the fiscal regulations (and free-rides the others’ fiscal discipline) while the European Central Bank has not yet fully established monetary discipline. For example, a coalition of member countries could “persuade” the Bank to undertake some form of financial solidarity, or bailout, to rescue any member in financial distress. Let’s consider the following hypothetical initial situation: one profligate national government starts running continuous and excessive budget deficits. Under normal conditions, additional borrowing should take place at progressively higher interest rates: i.e., financial market-based discipline should “kick-in.” The country-specific outright default risk premium should in turn signal to the national government that it needs to reverse its budgetary stance and reduce deficits and borrowing. However, if the financial market does not rule out some form of financial solidarity, or bailout, the country-specific risk premium may not rise as fast. If the financial market is convinced that somebody else will “foot the bill,” it may not cautiously screen and appropriately penalize the free rider.⁹ At the same time, a negative interest rate externality could occur and the “reference” interest rate of the whole monetary union (e.g., the German interest rate) would go up because it would no longer be risk free. In the mean time, the fiscal problem of the profligate government is hidden, or delayed, by the lack of a strong financial market signal. Hence, there is an element of cross-subsidy at work, and every EMU member will lose, albeit to different extent. Paradoxically, the free-rider’s interest rate may not raise as fast, and even fall, if financial markets expects the European Central Bank to undertake some form of financial solidarity directly or indirectly. If, at this stage, the free-riding member is unable, or not “required” (for example through the stability pact) to reverse its fiscal stance and thwart the accumulation of debt, the financial distress may become more severe over time. Eventually, the critical debt threshold at which some form of financial solidarity, or bailout, will indeed be needed may be higher, and the resource transfer process may become more severe. At this point, the pressure on the European Central Bank to rescue this government by reducing the real value of the country’s debt may increase even further since more resources are at stake (see also Restoy (1996) and Barber (1997)).

We are now ready to respond to the earlier question: what is the trade-off between the vanishing inflation/devaluation risk premium and the outright default risk premium in the wake of the EMU? Some authors, such as McKinnon (1991) and Bishop (1991), fear that the vanishing inflation/devaluation risk premium could be offset by a one-to-one increase in the outright default risk premium, at least in the “high-debt” EMU members that previously had a relatively higher inflation rate. Therefore, the expected decline in interest rates may be negligible in those countries. These authors implicitly assume that monetary discipline and financial market-based discipline will be firmly established, and that some EMU members will not be able, or willing, to reduce their public indebtedness after being admitted in the EMU: i.e., binding legal requirements and mutual surveillance will be ineffective and the financial market will punish the free riders. Others, such as Masson and Symansky (1995), maintain that interest rates will converge even further in the EMU. Eventually, interest rate differentials may fall below 40-50 basis points (which is the highest outright default differential estimated in their paper). These authors implicitly assume that both monetary and fiscal discipline will be established in the EMU and that the “high-debt” EMU members will continue their fiscal consolidation. Restoy (1996), and several other authors, take an intermediate view but assumes a “subsidy element” between “high-debt” and “low-debt” countries.

Standard and Poor’s unifies its credit rating in euro on the credit rating in foreign currency before the EMU because national governments will lose their privileged position (i.e., to monetize the public debt if needed) after adopting the single currency. Moody’s, instead, unifies its credit ratings in euro on domestic credit rating before the EMU because it assumes that national governments have already lost the ability to monetize the public debt by granting independence to their central banks. Hence, Standard and Poor’s is taking a more conservative approach to the effects of EMU on the “high-debt” members. Incidentally, Moody’s domestic currency ratings are lower than those of Standard and Poor’s.

What is the possible order of magnitude of the decline in interest rate levels in the high- and medium-interest countries? Early in 1997 interest rate differentials with Germany were still in the 150-200 basis points range in some EU members (excluding Greece). However, outright default risk premia are estimated at about 40-50 basis points, or less, in these countries (see Masson and Symansky (1995)). The gap between 150-200 basis points and 40-50 basis points could decline if both monetary and fiscal discipline are established (i.e., if Cases A or B prevail).

The rest of the paper assumes that monetary discipline will be firmly established from the onset of the EMU and that financial market-based discipline will rapidly deepen. In addition the EMU will also foster a decline in the overall level of interest rates along the lines discussed in Appendix I. The conditions discussed in the appendix will benefit each participating country and will be strengthened by a tight monetary discipline and widespread fiscal discipline. Hence, benefit each country).¹⁰ The next section examines the gains that a country with a relatively high debt-to-GDP ratio could obtain by pursuing a fiscal consolidation after being admitted in the EMU.

III. A Model to Analyze the Interest Spending and Taxation Channels

This section is based on a model that is a specialization of Blanchard and Fischer (1989). Their objective was to examine whether an increase in the government budget deficit and public debt increases the real interest rate. With this purpose in mind, Blanchard and Fischer adopted a version of the overlapping generation model in an exchange economy in which fiscal policy is not neutral. Their model is amended here in two ways to examine the operation of the two channels.

1. The Relation Between Real Interest Rates, Debt, and Deficits

Blanchard and Fischer (1989) show that real interest rates depend on the anticipated sequence of debt, or, equivalently, on the current level of debt and the anticipated sequence of deficits. The main features of this model are also described by Yaari (1965), Blanchard (1994 and 1985), Calvo (1988), Weil (1989), and others. This paper amends the Blanchard and Fischer model in two ways. The first amendment is that the model used here differentiates between: (a) the potential decline in country-specific risk premium, that will benefit more the high debt-to-GDP EU members that are willing to undertake a fiscal consolidation; and (b) a generalized fall in the level of interest rates reflecting, inter alia, the reduction in transaction and information costs following the adoption of a single currency (Appendix I).

The second amendment is in the assumptions concerning the tax function T(B_t, x). In the Blanchard and Fischer model, taxes are lump sum, and are positively related, but at declining rates, to the level of public indebtedness B_t: i.e., as public expenditure increases the authorities respond by increasing taxes. A feedback function from debt to taxes that is similar to the above specification is discussed in Sims (1993) and Leeper and Sims (1994). The signs of the first and second partial derivatives are respectively T_B ≥ 0 and T_BB < 0. T has an upper limit which is given by the asymptote in Figure 3 (b), i.e., national governments cannot increase taxation undefinedly. However, this limit cannot be reached and T_BB < 0 even for high levels of debt. Taxes also rise in the event of an increase of a fiscal policy shift parameter x (T_x > 0). In the EMU the ability to increase taxation as the debt rises (i.e., T_B) is likely to decline as follows.

The EMU will expedite the process of economic and financial integration of each country with the other members. Integration will in turn foster tax harmonization, as is discussed in Section IV, and reduce the ability of each country to increase their tax rates beyond those of the trading partners. Such a decline in fiscal autonomy is likely to be more significant in “high tax” countries.

To illustrate the foundation of the model, the behavior of economic agents is now briefly described. Time is continuous, and economic agents have a finite horizon. Output in the model is assumed to be exogenous and constant. Each agent is uncertain about his or her life span and faces an instantaneous probability of death p (where 0 < p < 1). Therefore, life expectancy is p⁻¹ These assumptions imply that, although agents are of different ages, their expected remaining life and marginal propensity to consume out of wealth are identical. Agents have a subjective discount rate θ, share labor income equally (independently of age), and save or dissave by acquiring or selling actuarial bonds b_t that pay an interest rate r(z) at time z plus a premium p until the bondholder dies. The agents choose to sell the claims on their wealth in the event they die and obtain a premium.

The premium derives from life insurance in the presence of an uncertain lifetime; bondholders cannot go bankrupt, and upon their death the debt of the government is canceled. This assumption eases the following analysis but is not the source of the non-neutrality of fiscal policies. As a result of this assumption, future utility is discounted at a rate [θ + p], which differs from the effective interest rate faced by individuals [r(z)+ p]. Finally, the agents of this economy maximize their expected lifetime utility under uncertainty about their lifespan. Specifically, at each point in time t they maximize,

\begin{matrix} 1) & E | \int_{t}^{\infty} {u (c}_{z} {) e}^{- θ (z - t)} d z | t \end{matrix}

Assuming a logarithmic utility that is a function of consumption c_t and that the probability of being alive at time z is e^-p(z-t), that in turn implies an increase of the rate of time preference, the objective function can be rewritten as,

\begin{matrix} 1') & \int_{t}^{\infty} \log (c_{z}) e^{- (θ + p) (z - t)} d z \end{matrix}

Individual’s wealth consists of human and nonhuman wealth. The first is the discounted stream of future disposable labor income, that is obtained by subtracting lump-sum taxes τ_t from gross income from labor y_t. Nonhuman wealth is held in the form of actuarial bonds b_t yielding a return of [r(z) + p] per period until the bondholder dies. The budget constraint faced by individuals, while alive is,

\begin{matrix} (2) & \frac{d b_{z}}{d z} = [r_{z} + p] b_{z} + y_{z} - c_{z} \end{matrix}

Economic agents cannot accumulate debt forever at a rate higher than the effective rate of interest [r(z) + p] which they face. Therefore, the following no-Ponzi game condition must be satisfied,

\underset{z \to \infty}{Lim} b_{t} e^{- \int_{t}^{z} (r_{μ} + p) d μ} = 0

By integrating (2) forward and applying the no-Ponzi game condition, the following intertemporal budget constraint is obtained,

\begin{matrix} 2') & \int_{t}^{\infty} c_{z} e^{- \int_{t}^{z} [r_{μ} + p] d μ} d z = b_{t} + h_{t} \end{matrix}

It postulates that lifetime consumption of an individual is equal in present value to its total wealth because he or she leave no bequests. Applying the maximum principle to (1′) subject to (2′) yields the following first order condition $\frac{d c_{z}}{d z} = [r_{z} - θ] c_{z}$ which when integrated to express c(z) as a function of c(t) and after substitution in (2′) yields c_t = (θ+p) [b_t + h_t]. Hence, [θ + p] is the propensity to consume out of total wealth.

The aggregate consumption of the population involves the consumption of all the agents who are still benefitting from the insurance scheme (i.e., agents who are alive at time t) $C_{t} = \int_{- \infty}^{t} C (s, t) p e^{- p (t - s)} ds$ . The double time index indicates all those agents born at time s (i.e., those from the generation “s”), that is prior to the current time t, who may still be alive today. Using the previous definitions and assuming that the propensity to consume is independent of age, aggregate consumption can be expressed as, c_t = (θ + p) [B_t + H_t]. The aggregate budget constraint of all the individuals alive is,

\begin{matrix} (3) & \frac{d B_{t}}{d t} = r_{t} B_{t} + Y_{t} - C_{t}, \end{matrix}

In a state of equilibrium, individuals save by purchasing the government’s liabilities. The aggregate non-human wealth accumulates at a rate (r), and not (r+p), because a portion pB_t of debt is extinguished with the death of individuals. Aggregate human wealth is given by the discounted stream of future disposable incomes,

\begin{matrix} (4) & H_{t} = \int_{t}^{\infty} [Y_{z} - T_{z}] e^{- \int_{t}^{z} [x_{μ} + p] d μ} dz \end{matrix}

The discount rate for aggregate human wealth (r+p) is higher than the discount rate for aggregate non-human wealth (r) due to the positive probability of death faced by individuals. This is the cause of the non-neutrality of debt and deficits.

In order to focus on the process of adjustment ensuing a change in fiscal policy, it is now assumed that output is exogenous and constant, and that the existing capital endowment is constant and cannot be traded by individuals. Therefore, the aggregate outstanding stock of wealth over which economic agents have control is H_t+B_t, where the latter term embodies the stock of public debt of the government. This assumption implies that the interest rate is not linked to the stock of capital but must adjust to secure equilibrium between aggregate demand and supply. The budget constraint of the government is,

\begin{matrix} (5) & \frac{d B_{t}}{d t} = r_{t} B_{t} + G_{t} - T (B_{t}, x) \end{matrix}

It is assumed that government expenditure does not affect marginal utility of private consumption, and that owing to finite lifetime and the absence of bequest in this model, taxes are shifted upon future generations. The public debt cannot be accumulated forever and the no-Ponzi game condition must be satisfied by the government as well: therefore public indebtedness must grows asymptotically at a rate lower than the real interest rate at which it is discounted. It must be reiterated that the government faces a lower interest rate than economic agents. Integrating (5) forward subject to the no-Ponzi game condition yields $B_{t} = {\int_{t}^{\infty} [G_{z} - T_{z}] e}^{- \int_{t}^{z} r_{μ} d μ} d z$ , which is the intertemporal government budget constraint. It implies that the current level of debt must be equal to the discounted stream of future primary surpluses. The equilibrium at the aggregate level is described by the budget constraint (5), which illustrates the dynamic of public debt, and

\begin{matrix} (6) & Y_{t} = (p + θ) (B_{t} + H_{t}) + G_{t}; \end{matrix}

\begin{matrix} (7) & \frac{d H_{t}}{d t} = (r_{t} + p) H_{t} - Y_{t} + T (B_{t}, x) . \end{matrix}

(6) illustrates the equilibrium in the goods market. The first term on the right hand side represents consumption of economic agents, which is a share of the stock of total wealth. This wealth is constituted by nonhuman wealth B_t and aggregate human wealth H_t (this is the discounted stream of all future disposable incomes in the economy). (7) is obtained by taking the time derivative of (6), and it describes the dynamic of aggregate human wealth. Whenever the condition $\frac{d B_{t}}{d t} + \frac{d H_{t}}{d t} =$ is satisfied, and after some substitutions, the following expression for the short-term interest rate is obtained,

\begin{matrix} (8) & r_{t} = θ + (p + θ) p | \frac{B_{t}}{Y - G} | . \end{matrix}

The short-term interest rate increases with the level of public debt and government expenditure (note that throughout this discussion that Y_t ≠ G_t). Budget deficits will affect the short-run interest rate over time, as they lead to a progressive increase in debt. Given a level of output and government expenditure, the interest rate must adjust in order to maintain aggregate consumption. Therefore, the sum of public debt and human wealth must remain constant as well, which implies that $\frac{d B_{t}}{d t} = - \frac{d H_{t}}{d t}$ (such that (H_t,+B_t) remains unchanged). Substituting in (6) and (7) the expression rB_t + (r + p) H_t = Y_t − G_t is obtained. Using (5) to simplify this expression again yields (8), showing that this relationship holds at any point in time.

To analyze the effects of fiscal policies on the term structure of interest rates, a long-term rate R is introduced. R is the yield of a perpetuity paying a constant coupon flow of unity. Therefore, 1/R_t is the price of these consoles, whose instantaneous rate of return is

\frac{1 + d (1 / R_{t}) / d t}{1 / R_{t}} = R_{t} - \frac{d R_{t} / d t}{R_{t}} .

Blanchard defines the long-term rate as a “shadow rate,” which would apply if the government actually introduced such a console. Due to the arbitrage condition, the relation between the short-term and the long-term rate is

\begin{matrix} (9) & r_{t} = R_{t} - \frac{d R_{t} / d t}{R} . \end{matrix}

Eliminating the short-term interest rate, r yields the following dynamic system in B and R:

\begin{matrix} (10) & \frac{d B_{t}}{d t} = (θ + p (p + θ) \frac{B_{t}}{Y - G}) B_{t} + G_{t} - T (B_{t}, x), \end{matrix}

\begin{matrix} (11) & \frac{d R_{t} / d t}{R} = R_{t} - θ - (p + θ) p | \frac{B_{t}}{Y - G} | . \end{matrix}

This system is not linear, and its properties can be investigated around the steady state by performing a linearization around this point. An implicit expression for the steady state can be found by setting the two time derivatives in (10) and (11) equal to zero, and solving for (B^*, R^*). However, the steady state may not be unique, and before proceeding further, it is useful to observe some features of the $\frac{d B_{t}}{d t} = 0$ schedule.

The first two terms on the right hand side of (10) represent the total expenditure of the government such as interest payments and the other non-interest government expenditure. These terms, which are labeled EXP, are shown in Figure 3(a). The last term represents total revenue, and is labeled REV in Figure 3(b).

Due to the properties of the REV function T(B_t, x), the EXP and REV schedules never meet, are tangent to each other, or intersect twice. The schedules will never meet if total expenditure always exceeds total revenue: this scenario will cause public debt to continuously increase. Because it violates the transversality condition it can be excluded from the investigation. The case in which total revenue exceeds total expenditure can also be excluded because it would not lead to any debt issue by the government. Instead, it would lead to an accumulation of assets and, as a result, the two schedules also would never meet.

When total expenditure exceeds total revenue but displays a lower slope than the latter, the EXP schedule could be tangent to the REV schedule or intersect the REV schedule twice, as in Figure 3 (c). If the schedules were tangent there would be an unstable equilibrium as the debt would increase before and after the tangency point. The economy could be at equilibrium only if it never departed from the tangency point; an excessively strict and unrealistic condition. In the second case, the system obtained from setting (10) and (11) equal to zero has two solutions, designated E and F, which are the focus of the investigation below. The difference between these solutions lies in the slopes of the EXP and REV schedules.

The slope of the EXP schedule is equal to θ + 2 (p(p + θ) B_t)/(Y_t −G_t), and is always positive and positively related to the stock of debt. However, this slope will fall, if the subjective discount rate θ declines. Changes in this rate are discussed by Blanchard and Fischer (1989), and others. Blanchard and Fischer observe that if a lower subjective discount rate θ is present elsewhere in an economic area, all θs will equalize at that level. Otherwise, a country (or region) may endure continuous capital accumulation or contraction. For example, a country with a relatively higher interest rate will reduce its consumption, continuously accumulate capital, and expand. A different country with a relatively lower interest rate will have an incentive to increase its consumption, continuously reduce its capital stock, and endure a slow down in economic growth. The slope of the REV schedule is determined by T_B, reflecting the ability of the government to raise additional taxes as the stock of debt increases. At low levels of the debt (e.g., at the level of debt corresponding to the solution E in Figure 4), T_B is high, while at high levels of the debt (say at F), T_B is low.

The system is linearized by using the Taylor series expansion over the two candidate steady state solutions. At the first solution E, the linearized system has two real and distinct roots of opposite signs. This establishes the steady state as a saddle point along which the economy can reach the stable arm AA in Figure 4a. The positive root pulls the system away from E, while the negative root pulls the system toward it along the stable arm AA. It can be shown that the slope of AA is positively related to T_B and θ, and negatively related to r and the level of G^*. The negative root is subject to an additional interpretation; it is the speed of adjustment of the dynamic process. This speed increases with T_B and θ, and decreases with r. The dynamic of the points in the (B_t, R_t) space (i.e., the arrows of the phase diagram) are obtained by taking partial differences from (10) and (11): to the left of the $\frac{d B_{t}}{d t} = 0$ schedule, B_t is increasing, while to the right it is decreasing. Above the $\frac{d R_{t}}{d t} = 0$ schedule, R_t is increasing; below it, it is decreasing.

Solution F cannot be a stable steady state because if the economy falls anywhere near this point, it will tend to move away from it as both roots associated to the Jacobean pull the economy away from it. This result can be further verified by the new direction of the arrows in the phase diagram: specifically, now both partial derivatives have a positive sign, indicating that the system moves away from F. A tangency point would yield an asymmetric equilibrium: if the economy found itself at this point no new debt would be accumulated.

The intuition behind the above result is as follows: if the level of taxation is high in a country with a high debt-to-GDP ratio, the government will find it increasingly difficult to raise additional taxes. Therefore, the national government will be constrained in its ability to respond to further increases in the level of indebtedness and secure fiscal sustainability. The debt will thus start to grow. At F, this restriction is so severe that the country joining the EMU may not be able to benefit from the fall in the level of interest rates. Of course, the higher the initial debt-to-GDP ratio and the level of taxation, the more “likely” the country is to lean toward F in the EMU. This point is discussed further in Section V below.

IV. The Possible Effects of the EMU on the Ability to Raise Additional Taxes

This section discusses the taxation channel, i.e., the extent by which the EMU may contribute to expedite the ongoing tax harmonization process reducing the ability of individual governments to raise additional revenue and curb budget deficits and public debt. Tax harmonization--that is bolstered by market forces and is mandated by institutional factors--is slowly reducing the differentiation in several areas of EU tax systems as is examined below. ¹¹

Before proceeding it is useful to classify the EU members according to their current “tax” burden (Table 2). This ranking will then be referred to in the ensuing discussion:

the “low-tax” countries such as Ireland (whose tax revenue-to-GDP ratio was 35.2 percent), Portugal (33 percent), Spain (34.2 percent), and the United Kingdom (35.2 percent);
the “medium-tax” countries such as Germany (39.1 percent) after the unification with East Germany which exhibited lower tax rates during the communist days, Greece (42.5 percent), and Italy (41.8 percent); and
the “high-tax” countries, such as Belgium (45.8 percent), Denmark (51.6 percent), France (44.5 percent), Luxembourg (43.7), and the Netherlands (44.4 percent).

Table 2.

Tax Ratios and Selected Statistical Indicators for the EU Members ^1/

Sources: OECD, Revenue Statistics of OECD Member Countries 1996 edition.

^{^1/}The selected statistical indicators are the mean weighted by GDP, the simple mean, the variance, the standard deviation, and the coefficient of variation. The latter is the ratio between the standard deviation and the mean.

^{^2/}1995 data are the latest available. 1994 data are the latest available for selected tax ratios of Greece and Portugal.

^{^3/}Social security contributions paid by employees, employers, and self-employed.

^{^4/}Includes taxes on immovable property of households, corporation, and government.

^{^5/}Includes general sales tax, value–added tax, excises, and all other consumption taxes.

^{^6/}Is a component of Taxes on Goods and Services. It includes general taxes, general sales taxes, and value added taxes.

^{^7/}Is a component of Taxes on Goods and Services. It includes excises, custom and import duties, taxes on export, taxes on investment goods, and other taxes on international trades and transactions.

^{^8/}Includes all members of the European union before January 1, 1995.

^{^9/}Includes the six founding members of the European union: Belgium, France, Germany, Italy, Luxembourg, and the Netherlands.

Table 2.

Tax Ratios and Selected Statistical Indicators for the EU Members ^1/

		Total Tax Revenue ^2/			Total Taxes on Income Profits and Capital Gains			Social Security Contributions ^3/			Taxes on Property ^4/			Taxes on Goods and Services ^5/			Taxes on General Consumption ^6/			Taxes on Specific Goods and services ^7/
	1965	1980	1995	1965	1980	1995	1965	1980	1995	1965	1980	1995	1965	1980	1995	1965	1980	1995	1965	1980	1995
	(in percent of GDP)
Belgium	31.2	44.4	45.8	8.6	18.2	17.5	9.8	13.5	15.5	1.2	1.0	1.1	11.6	11.6	11.7	6.6	7.5	7.3	4.0	3.4	4.2
Denmark	29.9	45.5	51.6	13.7	25.0	31.2	1.6	0.8	1.6	2.4	2.6	1.8	12.1	17.0	16.6	2.7	10.1	9.9	8.7	6.1	5.9
France	34.5	41.7	44.5	5.5	7.6	7.8	11.8	17.8	19.3	1.5	1.5	2.3	13.2	12.7	12.2	8.0	8.8	7.8	4.9	3.5	4.0
Germany	31.6	38.2	39.1	10.7	13.4	11.8	8.5	13.1	15.4	1.8	1.3	1.1	10.4	10.3	10.9	5.2	6.3	6.8	4.6	3.5	3.7
Greece	22.0	29.2	42.5	2.0	5.7	8.7	6.9	9.7	14.7	2.1	1.3	1.4	10.7	12.1	17.4	2.3	3.9	9.4	7.4	7.4	7.7
Ireland	24.7	32.4	35.2	6.4	11.8	13.7	1.6	4.6	5.2	3.7	1.7	1.6	13.0	14.1	14.2	1.4	4.8	7.5	10.6	9.1	6.2
Italy	25.5	30.2	41.8	4.5	9.4	15.2	8.7	11.5	13.0	1.8	1.1	2.3	10.1	8.0	11.2	3.3	4.7	5.8	6.1	2.9	4.5
Luxembourg	30.6	46.0	43.7	11.0	19.9	17.0	9.9	13.5	11.8	1.9	2.6	3.2	7.6	9.7	11.7	3.8	4.9	6.0	3.4	4.5	5.5
Netherlands	32.7	45.0	44.4	11.7	14.8	11.6	10.1	17.1	18.8	1.4	1.6	1.8	9.4	11.4	12.0	4.0	7.1	6.7	4.8	3.3	4.0
Portugal	18.7	28.7	33.0	4.5	5.7	8.6	4.0	8.5	8.7	0.9	0.4	0.8	8.1	16.7	14.7		4.6	7.5	7.6	7.8	7.0
Spain	14.7	24.1	34.2	3.6	6.3	10.0	4.2	11.7	12.4	0.9	1.1	1.8	6.0	5.0	9.8	3.3	2.5	5.5	2.7	2.5	3.5
United Kingdom	30.4	35.3	35.2	11.3	13.5	13.0	4.7	5.8	6.3	4.4	4.2	3.7	10.0	10.3	12.2	1.8	5.1	6.7	7.6	4.6	5.0
European Union^8/
Weighted Mean	29.1	36.2	40.4	7.8	11.2	12.2	8.1	12.3	13.8	2.0	1.8	2.1	10.5	10.3	11.7	4.5	6.1	6.8	5.4	3.6	4.3
Mean	27.2	36.7	40.9	7.8	12.6	13.8	6.8	10.6	11.9	2.0	1.7	1.9	10.2	11.6	12.9	3.9	5.9	7.2	6.0	4.9	5.1
Variance	34.4	55.1	29.2	13.5	34.6	36.8	11.2	23.3	27.5	1.0	0.9	0.7	4.3	10.5	5.0	3.7	4.3	1.6	5.1	4.4	1.7
Standard deviation	5.86	7.42	5.40	3.67	5.88	6.07	3.34	4.83	5.24	1.02	0.96	0.82	2.09	3.25	2.24	1.93	2.08	1.28	2.27	2.09	1.31
Coeffic. of Variation	0.22	0.20	0.13	0.47	0.47	0.44	0.49	0.45	0.44	0.51	0.57	0.43	0.20	0.28	0.17	0.50	0.35	0.18	0.38	0.43	0.26
European Union 6^9/
Mean	31.02	40.92	44.83	8.67	13.88	13.98	9.80	14.42	16.62	1.60	1.52	2.05	10.38	10.62	11.67	5.15	6.55	6.77	4.63	3.52	4.17
Variance	7.66	29.58	9.09	7.70	19.25	10.44	1.17	5.09	5.11	0.06	0.28	0.66	3.03	2.28	0.19	2.80	2.08	0.34	0.70	0.23	0.29
Standard deviation	2.77	5.44	3.01	2.77	4.39	3.23	1.08	2.26	2.26	0.25	0.53	0.81	1.74	1.51	0.43	1.67	1.44	0.58	0.83	0.48	0.54
Coeffic. of Variation	0.09	0.13	0.07	0.32	0.32	0.23	0.11	0.16	0.14	0.16	0.35	0.40	0.17	0.14	0.04	0.32	0.22	0.09	0.18	0.14	0.13